Citation
Household factors associated with environmental tobacco smoke exposure of young children

Material Information

Title:
Household factors associated with environmental tobacco smoke exposure of young children
Creator:
Yousey, Yvonne K
Place of Publication:
Denver, CO
Publisher:
University of Colorado Denver
Publication Date:
Language:
English
Physical Description:
268 leaves : ; 28 cm

Subjects

Subjects / Keywords:
Passive smoking in children ( lcsh )
Home ( lcsh )
Households ( lcsh )
Home ( fast )
Households ( fast )
Passive smoking in children ( fast )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Bibliography:
Includes bibliographical references (leaves 253-268).
Thesis:
Health and behavioral sciences
General Note:
Department of Health and Behavioral Sciences
Statement of Responsibility:
by Yvonne K. Yousey.

Record Information

Source Institution:
|University of Colorado Denver
Holding Location:
|Auraria Library
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
55626249 ( OCLC )
ocm55626249
Classification:
LD1190.L566 2003d Y68 ( lcc )

Downloads

This item has the following downloads:


Full Text
HOUSEHOLD FACTORS ASSOCIATED WITH ENVIRONMENTAL
TOBACCO SMOKE EXPOSURE OF YOUNG CHILDREN
B.S.N., Eastern Mennonite University, 1970
M.S., University of Colorado Health Sciences Center, 1977
A thesis submitted to the
University of Colorado at Denver
in partial fulfillment
of the requirements for the degree of
Doctor of Philosophy
Health and Behavioral Sciences
2003
by
Yvonne K. Yousey


This thesis for the Doctor of Philosophy
degree by
Yvonne K. Yousey
has been approved
by
Deborah S. Main


Yousey, Yvonne K., (Ph.D., Health and Behavioral Sciences)
Household Factors Associated With Environmental Tobacco Smoke Exposure
Of Young Children
Thesis directed by Professor Kitty Corbett
ABSTRACT
Reducing secondhand smoke exposure is one of the four priorities for global
tobacco prevention and control identified by the World Health Organiziation. Young
children are especially vulnerable to health effects related to tobacco smoke exposure,
most often exposed as a result of parental smoking practices in their home. Little
information is available about the behavioral factors which influence secondhand
tobacco smoke exposure in households where children reside.
This descriptive study examined household characteristics associated with
smoking policies in households using a household production of health framework
within a social ecology perspective. Several hypotheses examined characteristics
associated with smoking policies in households. They included: 1) no differences in
socio-demographic characteristics exist between household with complete smoking
bans and those without, 2) children in households with complete smoking bans have
less reported health effects related to smoke exposure than children in households
with no bans, 3) households with smoking bans have greater knowledge of
environmental tobacco smoke exposure, report more negative attitudes toward
environmental tobacco smoke exposure, and have less smoke exposure as measured
by cotinine levels than households with no smoking bans.
A sequential mixed-methods approach was used. Semi-structured interviews
were conducted with 20 households, some with smoke exposure and some without.
These interviews were used to increase content relevance and construct validity of the
survey developed for use in the quantitative phase. Surveys were administered to a
cross sectional sample of English and Spanish-speaking subjects, 18 years and older,
with children, age newborn through pre-school age residing in the household. Socio-
demographic factors, knowledge and attitudes of parents, health status of children in
the household, smoking behaviors and smoking bans in households were investigated.
Reports of smoking bans were validated through measurement of cotinine levels of
urine samples from children, newborn to five years in the household.
Two hundred twenty-five households completed surveys; 203 households
provided a urine sample from an age-appropriate child that was tested for cotinine.
Complete smoking bans were reported by 164 (73%) households, and no or partial
m


bans were reported by 61 households (27%); 97 children(48%) tested negative for
smoke exposure through cotinine levels and 107 (52%) tested positive, indicating
smoke exposure. Ethnicity of households when indicated by language in which the
survey was completed, and attitudes toward smoke exposure were household
characteristics significantly associated with the presence of smoking bans. No
smoking bans correlated positively with cotinine measures of smoke exposure
(r=.486). Smoke exposure in households was significantly better explained by the
inclusion of smoking bans in a model of smoking behaviors than by other measures
This abstract accurately represents the content of the candidates thesis. I
recommend its publication.
Signed
IV


DEDICATION
This dissertation is dedicated to my husband, Jim Dunn, and to our children,
Collin, Katie, and Kelsie with love and my deepest gratitude to all of you.


ACKNOWLEDGEMENTS
My sincere thanks go to the many people who made this dissertation possible.
I would like to thank the members of my committee, Kitty Corbett, Debbi Main,
Lauren Clark, Steve Koester, and Michael Zinser. Kitty, who supported me from the
beginning, and made the dissertation possible; Debbi Main who provided assistance
from afar and always answered my emails; Lauren Clark who always had a positive
comment and challenged me to strive forward; Steve Koester and Michael Zinser who
encouraged and challenged me as I completed the process. I also thank Carol Vojir
for her patience and never ending assistance in completing the statistical portions of
this work. And finally, thanks to Chris Pon and Norman Chandler for managing my
grant and keeping me abreast of the next deadline.
I also wish to thank the many people and organizations that participated in this
study: the staff and administration at Sanville Preschool, Adams County Head Start,
Adams District 50 Preschool, Sheridan Preschool/Head Start and the staff at the
School-Based Health Centers for their assistance in recruitment of subjects and in
data collection. Without their enthusiasm and loyal support, this project would not
have been possible. I thank the staff at Community Health Services of The Childrens
Hospital, and my colleagues from the School of Nursing at the University of
Colorado Health Sciences Center who provided support and encouragement along the
way. Special thanks go to Bonnie Gance-Cleveland, Lynn Gilbert, Geneva Jarvis,
Andra Opalinski, and Julie Degenstein for their assistance in data collection and
support of this project.
Many, many loving thanks go to my family. To my husband, Jim, who
walked by my side with unfaltering love, support, and patience, and took on countless
extra responsibilities for our family through the years so that I could complete this
project. I thank my children, Collin, Katie, and Kelsie, who always had an
encouraging word and understood when I couldnt be there all of the time. I thank
the members of my class, Cohort 4, for all of your encouragement and timely advice.
Finally, to my parents, family and friends from near and far who didnt always
understand what I was doing, but provided support and interest anyway. Thanks to
all of you for being there through the years and helping me to reach this goal.
This dissertation was supported by grant ID-068 from The Colorado Tobacco
Research Program, Boulder, Colorado.


CONTENTS
Figures....................................................xii
Tables.....................................................xiv
CHAPTER
1. INTRODUCTION................................................1
Overview.................................................1
Hypothesis and Specific Aims.............................4
Overview of Research Methods.............................6
Setting of the Study.....................................9
Description of the Study Population.....................11
Structure of Dissertation...............................12
2. BACKGROUND, THEORY AND LITERATURE REVIEW....................14
Background..............................................14
Theoretical Framework...................................18
Social Ecology....................................18
Household Production of Health....................19
Households and Smoking............................21
The Household as a Unit of Analysis...............23
Environmental Tobacco Smoke Exposure....................25
vii


Cotinine and Smoke Exposure............................25
Definition of Terms......................................... 28
Health Behaviors, Indicators, and Outcomes.............28
Smoking Policies in Households.........................29
Smoking Bans...........................................30
Review of Literature.........................................34
Epidemiology of Smoke Exposure.........................35
Passive Smoke Exposure in Homes: Intra-household
Factors................................................37
Smoking Policies in Households.........................40
Demographic Characteristics............................41
Health Care Utilization and Economic Costs.............43
Knowledge of Health Effects of Smoking.................45
Household Behaviors, Attitudes, and Smoke Exposure.....46
Social Support and Smoking.............................49
Self-efficacy..........................................50
Interventions for Smoke Exposure.......................51
Summary................................................52
3. METHODOLOGY....................................................54
Overview.....................................................54
Qualitative Methods
vm
56


Issues of Validity........................................58
Interview Instrument......................................59
Sample Selection for Interviews...........................61
Structure of Interview....................................62
Role of the Researcher....................................63
Data Collection...........................................65
Quantitative Methods: Survey Research............................66
Instrument Development....................................67
Measures..................................................73
Knowledge of Smoke Exposure........................76
Attitudes/Beliefs..................................77
Reliability and Validity of Instrument....................80
Content Validity...................................80
Reliability Analysis...............................81
Factor Analysis of Attitude Measure................83
Study Population and Setting..............................86
Participant Recruitment...................................87
Data Collection Procedures................................88
Subject Payment...........................................93
Cotinine Testing........................................ 94
Human Subjects Review............................................97
ix


Methods for Qualitative Analysis.............................97
Methods for Quantitative Analysis............................99
4. FINDINGS FROM QUALITATIVE DATA ANALYSIS.......................105
Description of the Sample...................................106
Observation of Smoking Behaviors............................108
Results of Qualitative Analysis.............................109
Description of Themes.......................................110
Health Promotion and Health Protection...............113
Implementation of Smoking Rules......................115
Knowledge of Effects of Smoking......................120
Attitudes and Beliefs................................122
Community Factors....................................125
Conclusions.................................................128
5. FINDINGS FROM QUANTITATIVE DATA ANALYSIS.....................130
Overview....................................................130
Demographic Characteristics.................................132
Ethnicity............................................135
Demographics by Recruitment Site.....................140
Household Characteristics and Smoking Bans..................144
Smoking Behaviors and Ethnicity......................152
Discussion of Smoking Measures.......................155
x


Health of Children in the Household..................155
Knowledge and Attitudes/Beliefs and Smoking Bans.....158
Smoking Bans and Cotinine Exposure...................163
Ethnicity and Cotinine Measurements.........................166
Smoke Exposure in Daycare...................................166
Mulitvariate Analysis-Logistic Regression...................167
Selection of Variables for Logistic Regression Model.168
Factors Associated With Smoking Bans.................173
Hypothesis Testing..........................................179
6. DISCUSSION AND CONCLUSIONS...................................182
Overview....................................................182
Discussion of Findings: Hypotheses..........................182
Findings Compared to Other Studies..........................187
Strengths and Limitations of the Study......................196
Future Research Questions...................................203
Conclusions.................................................207
APPENDIX
A. CONSENT FORM AND INSTRUMENT FOR
SEMI-STRUCTURED INTERVIEWS.................210
B. CONSENT FORMS FOR SURVEY QUESTIONNAIRE
SURVEY QUESTIONNAIRE.......................217
C. RECRUITMENT FLYERS FOR QUANTITATIVE PHASE...248
xi


BIBLIOGRAPHY


FIGURES
Figure
1.1 Household Factors and Tobacco Smoke Exposure................
2.1 Social Ecology, Household Production of Health and Environmental
Tobacco Smoke Exposure in Households.......................
xm


TABLES
Table
1.1 Study Sites..............................................................10
1.2 Ethnicity of Study Site Population.......................................12
3.1 Predictor and Outcome Variables in Survey................................71
3.2 Survey Questions For Situations in Households When
Smoking Is Allowed......................................................75
3.3 Knowledge Assessment.....................................................76
3.4 Attitude/Beliefs Assessment..............................................78
3.5 Reliability Analysis of Knowledge Scale..................................82
3.6 Factor Analysis: Rotated Component Matrix................................83
3.7 Comparison of Item Grouping in Factor Analysis...........................84
3.8 Reliability Analysis of Attitude/Belief Scale (13 Items).................85
3.9 Ethnicity/Racial Characteristics of Population...........................87
3.10 Independent Categorical Variables.......................................101
4.1 Interview Characteristics of Households by Smoking Status...............107
4.2 Interview Characteristics of Households by Ethnicity....................108
4.3 Summary of Themes and Codes Identified in
Household Interviews on Smoking........................................112
xiv


4.4 Health Promotion Activities and Health Habits of Families.............114
4.5 Reasons For and Against Smoking in the Household......................118
4.6 Knowledge of Effects of Smoking.......................................120
5.1 Characteristics of the Sample by Household............................134
5.2 Ethnicity of Subjects.................................................136
5.3 Demographic Characteristics and Ethnicity............................138
5.4 Sites From Which Subj ects Recruited..................................140
5.5 Demographic Characteristics of Preschool/Head Start Sites and SBHCs..142
5.6 Demographic Characteristics and Household Smoking Bans................144
5.7 Smoking Behaviors and Smoking Bans in Households......................149
5.8 Correlation of Smoking Measures With Presence of Smoking Bans.........151
5.9 Smoking Behaviors and Ethnicity.......................................153
5.10 Childhood Illness and Smoking Bans in Households.......................156
5.11 Correlations of Illness and Smoking Bans in Households...............157
5.12 Attitudes and Beliefs About Smoke Exposure............................159
5.13 Constructs of Attitudes and Beliefs...................................162
5.14 Cotinine Measurement of Passive Smoke Exposure........................164
5.15 Correlations of Reported Smoking Measures and Cotinine Measurements.. 165
5.16 Smoke Exposure in Daycare.............................................167
5.17 Categorical Variables For Regression Models...........................169
xv


5.18 Logistic Regression: Demographic Variables and Smoking Bans........172
5.19 All Variables and Smoking Bans in Households........................175
5.20 Smoking Behaviors and Cotinine Levels..............................178
xvi


CHAPTER 1
INTRODUCTION
Overview
Environmental tobacco smoke (ETS) exposure is an important source of
morbidity and mortality, being the third leading preventable cause of death in this
country (Hudzinski and Sirois, 1994). Unlike most other risk behaviors, tobacco
smoking is a health risk for those exposed to ETS who do not smoke directly
themselves (Sockrider, 1996). Children are more likely than adults to have adverse
health effects from ETS exposure (Ashley and Ferrence, 1998); they suffer from
lower respiratory illness, chronic middle ear effusion, pulmonary function changes,
asthma exacerbations, sudden infant death syndrome, and lung cancer (Cook and
Strachan, 1999; Etzel, 1997; Ey et al, 1995; Li, Peat, Xuan, and Berry, 1999; Samet,
1999). Between 8.7 and 12.4 million American children less than five years of age are
exposed to cigarette smoke in their homes (Etzel, 1997), resulting in substantial
public health and economic impacts (Ashley and Ferrence, 1998).
The household is a primary source of ETS exposure for young children. Little
is known about how families in households deal with ETS exposure in spite of the
fact that deleterious health effects are well established. Societal, economic, legal and
political factors contribute to lower levels of ETS control measures in homes
1


compared with workplaces and public places (Ashley and Ferrence, 1998) where
policies which restrict and/or prohibit smoking have effectively reduced ETS
exposure for non-smokers. There is a growing awareness of the dangers of ETS
exposure in households (Borland, Mullins, Trotter, and White, 1999; Farkas, Gilpin,
Distefan, and Pierce, 1999; Pizacani, Martin, Stark, Koepsell, Thompson, and Diehr,
2003), but little is known about the factors leading to ETS exposure and its reduction
in that environment.
The involuntary nature of ETS exposure in children places the burden of
protection on those who function in protective capacities, especially in homes. Young
children, vulnerable because of physiological and developmental differences, cannot
remove themselves from exposure and are dependent on other measures for
protection (Ashley and Ferrence, 1998). The examination of events and conditions
surrounding household policies related to tobacco use and smoke exposure will
increase understanding of how family members make decisions about smoking
behaviors and negotiate smoke free environments for their children. When adults act
to restrict smoking in their homes, they not only reduce morbidity in their young
children, but convey an important message about the dangers of tobacco smoke which
may have long term implications in smoking initiation in these children (Biener,
Cullen, Zhu, and Hammond, 1997). Reduction of exposure at the household level is
necessary for the health and protection of children living there.
2


Protecting children from the harm of smoke exposure is not only a
responsibility of the families with whom children live, but of the community in which
they reside, and the health professionals with whom they come in contact. This
research resulted from the interest of the researcher in addressing smoke exposure of
young children in families of diverse ethnicity, low income, and low educational
levels. A greater understanding of the dynamics of the household was needed before
specific measures for reducing smoke exposure could be implemented. The gap
between knowledge of harmful effects and actual controls in the home indicates a
need to examine situational factors that lead to smoke exposure (Goldstein, 1994).
In this study, the researcher was in the unique position to intervene with families in
maintaining smoke free environments through her role as the primary health care
provider of their children. This provided her the opportunity to investigate how
households address and deal with smoke exposure and what methods are utilized by
family members to reduce exposure. From these, recommendations, strategies, and
interventions for reducing exposure can be developed. Examining and describing
household health producing and health maintaining behaviors related to
environmental tobacco smoke exposure is a necessary prerequisite to analyzing their
health effects.
3


Hypothesis and Specific Aims
The purpose of this study was to investigate factors within households that
impact smoking practices and environmental tobacco smoke exposure of young
children living there. The household, a central unit for examining events and
conditions internally and externally, provides a basis for disease prevention and
health maintenance interventions (Berman, Kendall, and Bhattacharya, 1994).
Understanding how household decisions are made regarding potential health
risks/problems such as environmental tobacco smoke exposure is the first step in
protecting childrens health. How adults and family members in households make
decisions about smoking behaviors and negotiate smoke free environments for their
children provides the basis for development of effective interventions at the
individual and/or community level that lead to enhancement of tobacco-free family
environments.
This research pursued the following aims:
1. Explore household characteristics and relationships between factors
associated with reported smoking policies in households.
2. Identify means of implementation and enforcement of smoking policies
in households where young children reside.
3. Explore relationships between variables associated with household
smoking policies and actual smoke exposure of children as measured by cotinine
levels.
4


Using a social ecology perspective, it is hypothesized that household
characteristics will impact smoking policies as evidenced through smoking bans in
households. These household characteristics are impacted by factors both within the
household and the community in which it is located. It was predicted that cotinine
measurements of smoke exposure of young children residing in households with
no/partial home smoking bans would be greater than in those households with
complete home smoking bans. It was also predicted that children in households with
complete smoking bans would have fewer negative health effects than children in
households with no/partial bans. Five hypotheses were tested:
1. There are no differences in socio-demographic characteristics between
households with complete smoking bans and those without.
2. Children in households with complete smoking bans will have fewer
reported health effects related to smoke exposure than children in households with
no/partial smoking bans.
3. Households with complete home smoking bans will have greater
knowledge of environmental tobacco smoke exposure than households with no/partial
home smoking bans.
4. Households with complete smoking bans will report greater negative
attitudes toward environmental tobacco smoke exposure than households with
no/partial home smoking bans.
5. Households reporting complete home smoking bans will have less
5


smoke exposure as measured by cotinine levels than households reporting no/partial
home smoking bans.
Overview of Research Methods
In this study, factors impacting smoke exposure in households,
smoking policies (smoking bans) as reported by family members, and the
relationships between specified variables were investigated using a household
production of health framework. To achieve the aims of the study, research
employed both qualitative and quantitative methods; results were validated by
cotinine measurement of urine samples from children in the household.
Semi-structured interviews were conducted with parents in households
containing at least one child, newborn to age 5. The purpose of these interviews was
to provide information on the social and cultural context and personal meaning of
smoking (Erickson and Kaplan, 2000) necessary for the identification and
development of factors measured in the survey. The interviews explored health
behaviors and how they are implemented in households, smoking policies and
practices in the home and car, methods by which smoking policies are implemented
and enforced, and factors that families identified as being important in this process.
The survey questionnaire was developed to investigate the hypotheses
exploring household factors and health effects with smoking policies. The content of
questions included: household characteristics associated with smoking policies,
6


actual household smoking policies (rules), the implementation and enforcement of
restrictions in households, health indicators of children in the household, knowledge
of effects of tobacco smoke exposure, and attitudes/beliefs about ETS exposure of
children. The characteristics identified through the semi-structured interviews were
confirmed by literature findings, developed as variables, and measured in the survey.
The second hypothesis was tested by measuring actual smoke exposure of
urine obtained from a child, age newborn to school age, in a participating household.
Measurement of cotinine, a biomarker of nicotine, provided a quantitative measure
whereby actual smoke exposure of young children could be measured, validating
reported smoking policies and smoking behaviors in households. Identification of the
main sources of exposure and quantification of the dose of inhaled smoke are
fundamental to the study of passive smoking effects in children (Jarvis, 1999).
Surveys provide cumulative information about ETS exposure, but there is a danger of
misclassification through underreporting. The reliability of survey assessments is
enhanced when combined with biochemical verification such as cotinine analysis
(Wewers and Uno, 2002). Together they provide the best assessment of the extent of
ETS (Jarvis, 1999; Scherer, Meger-Kossien, Riedel, Renner, and Meger, 1999), more
accurate than the use of either of them alone.
Only households with children less than school-age were invited to participate
because the level of cotinine in body fluids of children in this age groups are a more
valid measurement of actual smoke exposure. Younger children are more confined to
7


their homes in which smoke exposure may occur and do not have the ability to
remove themselves from it (Ashley & Ferrence, 1998; Hopper & Kelly, 2000). They
also have less opportunity for smoke exposure outside the home environment by other
than their parents.
The following diagram illustrates the hypothesized relationships in the study:
Figure 1.1: Household Factors and Tobacco Smoke Exposure
Setting of the Study
The population selected for this study includes families with children, 5 years
or younger, who reside in Adams, Arapahoe, and Jefferson counties in metropolitan
Denver, Colorado. Specific areas within these counties were targeted because of their
accessibility to the researcher and their socioeconomic status as measured by income
8


levels of young families residing there. Because tobacco smoking is associated with
lower socioeconomic status (Arborelius et al, 2000) areas with known lower
household income were selected. These areas also included populations with ethnic
diversity.
Four faculty practice sites associated with the University of Colorado Health
Sciences Center, School of Nursing, located in Westminster (2), Sheridan (1) and
Arvada (1), provide primary health care services to families with children, newborn to
eighteen years of age in metropolitan Denver, Colorado. Two of these faculty practice
sites (Westminster and Sheridan) are located in areas with a federally qualified
underserved health status designation. Services are targeted for families who are
Medicaid eligible, CHP+ eligible or who have no health insurance. The school-based
health centers in Westminster and Sheridan are jointly administered with staffing
provided by Community Health Programs of The Childrens Hospital and University
of Colorado Health Sciences Center, School of Nursing in Denver, Colorado. Four
Head Start sites in Adams County and two preschools associated with District 50 and
District 14 providing preschool and early childhood services to three and four year-
old children in Adams County were utilized. One Head Start/preschool in Sheridan
School District (Arapahoe County) also participated.
9


Table 1.1: Study Sites
School-based Health Centers (SBHCs) County Pre-school/.Head Start Sties County
Sheridan Arapahoe Sheridan Preschool/Head Start Arapahoe
Westminster (2 sites) Adams Westminster District #50 Pre-School Adams
CarinClinic (Arvada) Jefferson Adams County Head Start (4 sites) Adams
Sanville Preschool Adams
The school districts in these locations include families of lower income with
40-60% of families eligible for free/reduced lunch. Many of these families who are
Medicaid eligible, CHP+ eligible, or uninsured utilize the school-based health clinics
for services. Families whose children attend Head Start met financial criteria for
enrollment. The majority of families utilizing services at preschools met financial
eligibility requirements necessary to access services.
The qualitative portion of the study was conducted using the school-based
health center sites. Families from all sites were recruited for participation in the
survey and measurement of cotinine levels.
10


Description of the Study Population
The population selected for the study consisted of low-income families with
young children, age five and under, who reside in School Districts 14, 50, and 27J of
Adams County, Sheridan School District in Arapahoe County, and in Arvada of
Jefferson County.
Families with children enrolled in four preschool/Head Start facilities and four
school-based health centers were eligible as study participants. These sites were
selected because of their geographic location, and the willingness of administrators to
facilitate participation of families. Two other preschools in Adams County were
invited to participate but declined. The school-based health centers were selected
because they have families with age appropriate children, are ethnically diverse, and
of low income. The researcher had access to these populations because she provided
primary health care services in these sites.
The Head Start sites and preschools were located in the same school districts
as the three school-based health centers. Head Start sites in two other school districts
with low income and ethnically diverse populations in Adams County also
participated. One school-based health center is located in Jefferson County,
providing services to uninsured families who attend three elementary schools and on
middle school of lower income in Arvadada. Approximately 1021 children were
enrolled in preschools/Head Start facilities and 3717 children were enrolled in school-
11


based health centers. Ethnicity/racial characteristics of the study population based on
enrollment in preschools or school-based clinics are as follows:
Table 1.2: Ethnicity of Study Population
Preschools SBHCs Total SBHCs And Preschools Total School Districts
N % N % N % N %
Not-Hispanic, White 1459 39 406 40 1865 39 11082 42
Hispanic 1843 50 515 50 2358 50 13114 50
Black 72 2 30 3 102 2 575 2
Asian 120 3 40 4 160 3 1281 5
Other 223 6 30 3 253 5 383 1
Total 3717 1021 4738 26435
Structure of Dissertation
Chapter 1 provides the overview of the need for research in ETS exposure in
homes, describes the purpose of the research, states the hypotheses and provides an
overview of the research methods, description of the study setting, and the study
population. Chapter 2 provides the background for the research including theoretical
framework, definition of terms and variables, and a review of the literature. Chapter
3 includes a description of the design with discussion of the qualitative and
quantitative methodologies: semi-structured interviews, survey questionnaire, and
12


testing of urine for cotinine to validate results from survey questionnaire. Chapter 4
discusses qualitative findings and use of these data in development of the survey.
Chapter 5 includes quantitative findings, with descriptive summaries, multivariate
analysis and conclusions. Chapter 6 further describes conclusions, study limitations
and provides recommendations for future research.
13


CHAPTER 2
BACKGROUND, THEORY AND LITERATURE REVIEW
Background
Much has been written about disease risks related to environmental tobacco
smoke exposure, and a growing awareness of this hazard has resulted in efforts to
reduce exposure in public places (Pizacani et al, 2003). The primary method of
protecting nonsmokers has been to restrict smoking through bans and non-smoking
policies; these have been shown to effectively limit exposure in public places and
thereby reduce health risks of exposure in these locations. Laws limiting ETS
exposure in public places do not directly impact exposure in private households.
Household smoking restrictions may have effects similar to those observed for
workplace and public place restrictions. Not only do these restrictions protect
children, but there is mounting evidence that restrictions in homes increase cessation
and reduce smoking (Farkas et al, 1999), prevent initiation (Biener et al, 1997), and
may be an indicator of the degree of anti-smoking social climate in a community
(Farkas et al, 1999).
Environmental tobacco smoke is a real, substantial threat to child health. For
the vast majority of children, exposure to tobacco smoke is involuntary, arising from
smoking, mainly by adults, in places where children live, work, and play. The World
14


Health Organization estimates that approximately one-half of the worlds children
breathe air polluted by tobacco smoke, particularly in their homes (WHO, 1999). In
the United States, the median prevalence for smoking in 1996 was 23.6% (MMWR,
1997); in 2000 the prevalence was 23.3% (Giovino, 2002); state specific prevalence
of in-home exposure of children in 1996 ranged froml 1.7% in Utah to 34.2% in
Kentucky (MMWR, 1997). The prevalence of current smoking in 2000 was highest
among persons involved in childbearing and childrearing age groups, 18-24 and 25-
44; the Center for Disease Control reported substantial decreases in current smoking
prevalence between 1993 and 2000 for all age groups except 18-24 years (MMWR,
2002).
In 2000, the prevalence of cigarette smoking among US adults was 28.6% for
those with less than 12 years of formal education, 29.5% among high school
graduates, 22.6% among persons with some college, and 11.2% among college
graduates. Prevalence of current smoking was high among blue-collar workers and
service workers, both indicators of lower socioeconomic status (Giovino, 2002). The
prevalence of current smoking continues to be higher for those below the poverty
level (1999) (31.7% [95%CI = 1.9]) versus those at or above the poverty level
(22.9% [95%CI = 0.7]) (MMWR, 2002) resulting in a greater burden of negative
health effects on children living in these environments.
Mannino et al. (1996) using 1991 National Health Interview Survey (NHIS)
data, reported that 31.2% of children in United States had daily ETS exposure and
15


37% had daily or less than daily exposure. This study also reported that the younger
children with ETS exposure experienced the strongest negative health effects.
Schuster et al. (2001) analyzed NHIS data (1994), and reported that 30% of children
live in homes with smoking by residents at least one day per week, 35% live in homes
with any smoking, and 34% of children, five years or younger, live in homes with
regular smoking. The Environmental Protection Agency estimates that every year in
the United States, between 150,000 and 300,000 children under ages one to one-and-
one-half years suffer respiratory ailments such as pneumonia and bronchitis from
breathing second hand smoke (EPA, 1994). According to the Colorado Department
of Health and Environment, 193,000 children in Colorado are exposed to ETS in their
homes each year (Tobacco Use Prevention and Reduction Plan for Colorado, 2000).
Smoking bans and restrictions are currently regarded as a primary means of
reducing non-smokers ETS exposure in the workplace and other public places
(Marcus, 1992; Schuster, 2002) while investigation of smoking restrictions in
households is just beginning (Gilpin et al, 1999; Pizacani et al, 2003). Home
smoking bans are a relatively new approach for dealing with ETS exposure, and
requiring parents to conform to smoke-free policies in their own homes may be
difficult (Wewers and Uno, 2002). The sanctity of the family unit restricts the
ability of policy actions to diminish tobacco smoke exposure and cultural values
sanction the individuals right to make rules in his/her home (Goldstein, 1994).
While a growing body of evidence on opinions and beliefs opposing ETS exposure in
16


home environments exists (Goldstein, 1994), the commonly held belief that
governments or other external agents should not interfere with behavior in private
settings has contributed to a lower level of support for ETS control measures in
homes compared to other settings (Ashley and Ferrence, 1998). This belief contrasts
with other laws and regulations protecting children from physical and sexual abuse in
the privacy of their homes, and those requiring school attendance, immunizations, etc.
(Ashley and Ferrence, 1998). Further investigation of household factors impacting
smoke exposure and evaluation of efforts to reduce ETS exposure are necessary if
efforts to provide smoke-free home environments for all children are to be successful
(Bek, Tomac, et al., 1999; Wewers and Uno, 2002).
The emergence of environmental tobacco smoke control in home
environments presents unique challenges and as a public health priority raises a host
of social, legal, and political issues. The potential for change rests not only on
supportive public, professional, and political attitudes with regard to protection of
children from harm, but also on the realities of housing, income, education, and child
care (Ashley et al, 1998). Understanding the factors impacting ETS in family
controlled spaces such as homes and automobiles is a beginning point from which to
address these issues. Identifying factors which influence smoke exposure can be the
first step in developing interventions appropriate for reducing ETS in the home.
Interventions are needed which will be effective in reducing the childs exposure
17


from all sources (Institute of Medicine, Clearing the Air: Asthma and Indoor Air
Exposure. 2000).
Theoretical Framework
Concepts of social ecology provide the umbrella in which environmental
tobacco smoke exposure in a household setting can be understood. Household
production of health framework organizes specific macro-level and micro-level
factors from social ecology so that they can be measured, and inputs and influences
can be accounted for as they impact factors. Finally, household production of health
provides a method by which relationships between factors and outcomes can be
addressed.
Social Ecology
Social ecology considers the nature of peoples transactions with their
physical and sociocultural surroundings (Sallis and Owens, 1997). This approach
considers micro-level to macro-level factors operating in a synergistic fashion with
individuals, groups, organizations, communities and populations. As such, it honors
factors that influence individuals through their social and physical environments, and
allows for consideration of health problems such as tobacco smoke exposure and
solutions at a variety of levels (Sallis and Owen, 1997; Corbett, 2001). People are but
one component of the larger behavior-setting system, which restricts the range of
18


their behavior by promoting and sometimes demanding certain actions and by
discouraging or prohibiting others (Sallis and Owen, 1997).
The social ecology perspective maintains that multiple levels of influence are
important for understanding the problem of environmental tobacco smoke exposure.
Smoking behaviors of individuals and households are considered within the context
of cultural norms, environmental cues, and infrastructure constraints including costs
and restrictive polices (Corbett, 2001). As a result, comprehensive strategies targeting
prevention or reduction of environmental tobacco smoke exposure at the level of
individuals, groups or social networks, organizations, communities, and populations
may be implemented.
Household Production of Health
While a social ecological perspective provides overall guidance for
understanding problems and solutions associated with environmental tobacco smoke
exposure, household production of health provides a basis upon which factors directly
related to households can be studied as they relate to ETS exposure. Berman,
Kendall, and Bhattacharyya (1994) describe use of the household and its application
in research and social action. Household production of health is a dynamic
behavioral process through which households combine their knowledge, resources,
behavioral norms and patterns with available technologies, services, information, and
skills to maintain and promote the health of their members (Berman, Kendall, and
19


Bhattacharya, 1994). This approach focuses on the presence and maintenance of
health; because the concept of health has multiple determinants, there may be various
pathways through which a household can maintain a level of health. Behaviorally
chosen determinants have a biological impact on health (DaVanzo and Gertler, 1999),
and integration of disciplines allows for a broad investigation of these determinants as
well as health changes in the household unit (Berman, Kendall, and Bhattacharya,
1994).
Household processes are becoming more critical as determinants of impact as
health interventions increasingly rely on behavior changes to produce benefits
(Berman et al, 1994). Households are apart of a larger social and economic
environment and are best analyzed in that context. Utilizing the household as a central
unit of focus allows us to examine both internal and external events and conditions
that impact the health of householders (Berman et al, 1994). In the household,
individual factors such as knowledge, attitudes, beliefs, and economic considerations
(Mattila-Wiro, 1999) interface with social networks, social support and other factors
from the environment to further explain health behaviors.
Social ecology provides the context in which all events, both internal and
external can be considered in assessing the household production of health. Intra-
household health behaviors occur or arise from within the household and are well
defined in the household production of health model. These include feeding practices,
child-care, health seeking behaviors at home, home hygiene and sanitation behavior,
20


use of preventive and curative services for health. They may be congruent with or
differ from external social norms.
External factors including household income, education, social status arise
from social and economic environments in which households find themselves. While
these are defined by environments outside of the household, they impact how
households view themselves and determine health behaviors. Both intra-household
and community factors are influenced by patterns of knowledge, cultural norms, and
expectations of efficacy, described by Berman et al. (1994) as inputs to health. All
three are important in influencing health behaviors of the household. The complex,
dynamic process resulting from the interaction of these ubiquitous factors is not easily
studied. Household production of health provides the framework for studying
specific factors influential in environmental tobacco smoke exposure and social
ecology provides the overall umbrella through the problem, solution, and
interventions can be addressed at many levels.
Households and Smoking
Households are frequently used as the unit of analysis in studies associated
with health behaviors as they contain several defining characteristics, relevant to
determination of health indicators and health responses. Household is defined as a
task-oriented residence associated with physical location (Netting, Wilk, and
Amould, 1984) composed of members living together by mutual consent (Dwyer and
21


Bruce, 1988), a social environment for child rearing, and a setting for child health
interventions (Berman, Kendall, and Bhattacharya, 1994). It is an economic and
social unit in which individual and environmental factors shape health outcomes,
health promoting behaviors and behavior change (Netting, Wilk, and Amould, 1984).
The household is a primary locale within which daily life takes place as well as the
institution for primary social and biological reproduction. It is usually the physical
locale and social environment for child rearing and child health interventions. Ideas
of social order emerge from households and manifest themselves within the context
of intra-household relations. This social order may reflect the larger society or it may
offer exceptions from those norms (Berman et al, 1994). Consequently,
understanding how households function within the social and cultural environments
in which they exist is essential in understanding behaviors related to health problems
such as environmental tobacco smoke exposure.
The concept of household policies is considered in this project to highlight
and target family and household practices and rules based on beliefs, knowledge, and
communication about exposure to tobacco smoke. It is at the level of the household
that behaviors emerge, directly impacting the health of residents living within. These
behaviors do not occur in a vacuum, but are influenced by cultural norms and
environmental cues arising from the surrounding environment. Individual factors
influence household policies and these are considered as influences on household
policies. From the complexity of the household and its surrounding ecological
22


influences, policies are determined and implemented directly impacting the health of
those within. The concept of household was selected because it provides a broader
perspective and a more accurate and comprehensive view of smoke exposure as a
health risk for children. Individual and family factors are considered under the
umbrella of the household as they affect health practices and behaviors. The
household production of health framework is the organizing framework through
which the relationships between these variables, measured as household and
individual factors, can be assessed.
The Household as a Unit of Analysis
Household has been defined as a place, a mode of social organization, or a
cluster of functions (Berman et al, 1994). Inherent in the household are people who
live together by consent and who perform certain social and economic functions
(Berman et al, 1994). This provides a broader, more comprehensive perspective in
addressing factors in a childs life impacting his/her health than does considering just
a family unit. However, the defining characteristics of a household are complex and
do not lend themselves easily to measurement. Consideration must be given to
physical locale, functions within the household, and household patterns in
understanding health indicators and health effects of the household.
This research focused primarily on the home as a household environment.
The home (for which the household is a setting) is the greatest single source of
23


environmental tobacco smoke exposure of young children and little is known about
how smoking policies are determined or implemented in that setting (Ashley and
Ferrence, 1998). Home and household are used interchangeably. Families are
two or more individuals who live collectively by consent in a household, including
adults and children. Households include the concept of families in a physical locale,
with tools for social control (Netting, Wilk, and Amould, 1984) and social
reproduction. While households have been studied extensively from an economic
perspective, little is known about composition, functions, or behavioral processes as
they relate to environmental tobacco smoke exposure.
Little is known about how inputs and influences impact environmental
tobacco smoke exposure of young children within the household context. These
cannot be studied in isolation; rather, they are considered in context of outside
influences impacting the household. The knowledge and range of importance of
health effects, beliefs about efficacy of environmental tobacco smoke reduction, and
the social and economic factors impacting ETS exposure are explored within the
framework. Factors that influence behavior change in social and policy environments
and are applicable in the household context need further consideration as well. This
research will provide an opportunity to begin investigating possibilities for policy
options for families as a basis for behavior change related to environmental tobacco
smoke exposure.
24


Environmental Tobacco Smoke Exposure
Environmental tobacco smoke is composed of two types of smoke: side
stream smoke and exhaled mainstream smoke. Side stream smoke is emitted from the
burning end of a cigarette in between puffs. It contains higher concentrations of
chemicals such as ammonia, nicotine, carbon monoxide and carcinogens. Exhaled
mainstream smoke consists of smoke which escapes from the burning end during
puff-drawing and gases which diffuse during smoking through the cigarette paper
(Brownson, Figgs, and Caisley, 2002). Environmental tobacco smoke contains forty-
three chemicals that are known human or animal carcinogens (Environmental
Protection Agency, 1994): eye and respiratory irritants, systemic toxicants, solid
particles, and semi-volatile and volatile organic compounds. Significant amounts of
nearly thirty volatile organic compounds have been measured, remaining in the air for
prolonged periods of time following the smoking of a cigarette (Institute of Medicine.
Clearing the Air: Asthma and Indoor Air Exposures. 2000). Exposure to side stream
and exhaled mainstream smoke are a threat and result in similar adverse health-
related affects to active smokers and those non-smokers exposed to ETS (Wewers and
Uno, 2002).
Cotinine and Smoke Exposure
Estimating the public health impact of ETS exposure in children and the
extent of that exposure are essential in providing the information necessary for action
25


to reduce health consequences of ETS (WHO: International Consultation of
Environmental Tobacco Smoke (ETS) and Child Health, 1999). There are currently
no means by which harmfiil components of ETS can be directly measured in the
organs of interest. Indirect measures of ETS exposure have been developed including
self-reports, biological markers and environmental air monitors (Matt et al, 1999).
These measures differ considerable in terms of reliability, validity, potential biases,
cost, and ease of administration.
Cotinine, the major proximate metabolite of nicotine, is the most widely used
and considered to be the best biological marker of ETS exposure (Rickert, 1999). It
can be detected in blood, saliva, urine, semen, and hair. Its presence indicates that a
person has been exposed to nicotine, but does not measure direct exposure to disease-
causing constituents. The concentration of cotinine is affected by individual
differences in uptake, distribution, metabolism, and excretion of nicotine. In
breastfeeding infants, the concentration is also influenced by mothers frequency of
breastfeeding, smoking behavior, ETS exposure and use of nicotine replacement
therapies (Matt et al, 1999; Haufroid and Lison, 1998).
Cotinine is shown to be valid over time, with a half-life of32-82 hours
(Rickett, 1999; Jarvis, 1999) with variations from 16-82 hours (Haufroid and Lison,
1998; Peterson et al, 1997) compared to nicotine that has a half-life of 30 minutes to 2
hours. It provides evidence of smoke exposure from several days to a week, but
cannot measure cumulative exposure over previous months and years. The half-life of
26


cotinine is typically longer in infants and young children averaging from 40 hours (18
month) to 65 hours (neonates) (US EPA, 1992). In spite of its limitations, cotinine is
recognized as the most sensitive and specific biomarker readily available (Manuel,
1999). Repeated measurements of cotinine have been shown to provide a more
accurate descriptions of an infants ETS exposure but a single measurement may be
used as binary marker of passive smoke exposure as well (Woodward and
AlDelaimy, 1999; Peterson et al. 1997).
Urinary cotinine excretion is variable across and within individuals,
depending on renal function, urinary flow rate, and urinary PH. Urinary results may
be expressed as nanograms of cotinine per milligram of creatinine in order to correct
for differences in dilution effects. Low levels of creatinine in infants compared to
adults may result in cotinine to creatinine ratios that are higher than for adults
(Tobacco Monographs, 27; Watts, 1990; Haufroid and Lison, 1998). Racial
differences among children may also affect cotinine levels with black children having
higher urinary cotinine levels than white children (Knight, Eliopoulos, Kelin, and
Greenwald, 1996). The presence of nicotine in some foods such as eggplant,
potatoes, tomatoes have raised concern regarding impact on urinary cotinine values
but studies thus far indicate negligible interferences from these substances (Haufroid
and Lison, 1998; Jarvis, 1994).
Using a biomarker such as cotinine as a quantitative measure of ETS exposure
validates self-reports of ETS exposure in households because it is specific for tobacco
27


smoke exposure. Because of the longer urinary half-life of cotinine as compared with
nicotine and of the absence of sample contamination during acquisition, it is currently
considered the marker of choice. It is a reliable indicator of health outcomes related
to smoking practices in households because of the consistent correlations between
urinary cotinine and daily tobacco consumption (0.39 to 0.99 and usually greater than
0.75) as reported by Haufroid and Lison, (1998); 0.62 reported by Seifert et al,
(2002); 0.50-0.63 for smoking mothers reported by Matt et al, (1999).
Definition of Terms
Health Behaviors. Indicators, and Outcomes
Health behaviors include the actions of individual, household, and groups and
the determinants, correlates, and consequences of those activities (Glanz et al, 1997)
to improve and enhance the quality of life. Included are observable, overt actions and
mental events and feeling states surrounding these actions which can be reported and
measured as (Glanz et al, 1997). In this research, the individual and household
activities and the perceptions of these actions will provide measures of health.
Indicators reflecting these measures include the parental rating of children in the
household as healthy or not healthy, reported numbers of minor illnesses occurring in
children in the past year, and the incidence of chronic illnesses and asthma among
28


children in the household. Inputs to health and influences behind those inputs are also
considered as they produce or influence health activities.
Health outcomes in households may be quantified in terms of morbidity and
mortality, resulting from health behaviors measured over time. For the purposes of
this study, health outcomes are not measured directly but quantified through health
indicators. Health indicators are those measurable factors demonstrating behaviors
associated with prevention and/or reduction of smoke exposure. They include
reported smoking bans in households and reported health of children in households.
Measurement of cotinine is a health risk indicator as it is used to measure smoke
exposure, validating parent reports of smoking bans or no smoking bans. Newborn
through pre-school age children live in close proximity to their parents, and do not
have the ability to voluntarily remove themselves from the household. Consequently,
testing of their body fluids provides the most reliable and valid indicators of smoke
exposure in the household.
Smoking Policies in Households
Smoking policies refer to the rules and practices within the household
impacting the environmental tobacco smoke exposure of young children who reside
there. These involve rules about whether or not any smoking by anyone is permitted
in homes or vehicles. They have been identified as informal controls by Goldstein
(1994) as opposed to formal controls (laws, fiscal measures, and bureaucratic
29


regulations) implemented in workplaces or public places to decrease ETS exposure.
Household smoking rules take many forms and include smoking behaviors allowed in
households. Examples of smoking policies include: the last time that someone
smoked in the home, number of cigarettes smoked in the home, number of smokers
living in the home, locations in home where people are allowed to smoke, or
situations in which people are allowed to smoke.
The rules that are articulated or practiced related to smoking within a
household are considered as they modify health practices and produce health effects
within that household. Smoking policies are categorized in three levels in
households: no smoking restrictions (no home smoking ban), some smoking but with
restrictions (partial home smoking ban), no smoking allowed (complete home
smoking ban) (Gilpin et al, 1999; Pizacani et al, 2002; Wewers and Uno, 2002).
Smoking policies comprise the health behaviors in the household production of health
framework as specified for this research.
Smoking Bans
For the purposes of this research, smoking bans are selected as the measurable
indicator for smoking policies in households. Other smoking behaviors are measured
and compared to smoking bans. A complete home smoking ban is defined as no
reported smoking in the household at any time; a partial home smoking ban is defined
as smoking with restrictions to either time or place in the home; a no home smoking
30


ban is defined as smoking with no restrictions in the home. For purposes of analysis,
the partial and no home smoking bans are combined into one category identified as no
home smoking ban; complete smoking bans are reported as smoking bans. Socio-
demographic characteristics (independent variables) are investigated as they relate to
smoking bans (dependent variable). In later analysis, smoking bans are the
independent variable and cotinine measurements of smoke exposure are the
dependent variable.
Smoking bans in vehicles include the same categories as home smoking bans
and are grouped using the same criteria. Complete vehicle smoking bans allow no
smoking in vehicles at any time, partial vehicle smoking bans allow smoking with
some restrictions including when children are not in the car or only with the window
down, and no vehicle smoking bans allow unrestricted smoking in vehicles regardless
of who is in the car or if the windows are down or not.
The following schema will be utilized as a framework through which
household actors and their relationships to household production of health and health
outcomes related to smoke exposure are examined.
31


Figure 2.1: Social Ecology, Household Production of Health and Environmental
Tobacco Smoke Exposure in Households
Social Ecology and Household Production of Health
HOUSEHOLD
Intra-household factors are identified through exploration of household
smoking policies, attitudes, actual smoking behaviors in homes, knowledge of smoke
exposure with families containing young children. Macro-level or community factors
are identified through demographic data. Smoking behaviors are categorized in
households and measured as health behaviors. Health indicators related to smoking
behaviors are measured through health of children in the home and cotinine analysis
of urine of an age appropriate child in the household.
The emergence of environmental tobacco smoke control in home
environments presents unique challenges and as a public health priority raises a host
32


of social, legal, and political issues. Reducing or eliminating tobacco smoke
exposure in homes rests not only on supportive public, professional, and political
attitudes with regard to protection of children from harm, but also on the realities of
housing, income, education, and child care (Ashley et al, 1998). Identifying factors
contributing to ETS in family controlled spaces such as homes and automobiles is a
beginning point from which to support and encourage household smoking policies
with known effects. Identifying factors which influence smoke exposure can be the
first step in developing interventions appropriate for reducing ETS in the home.
Interventions are needed which will be effective in reducing the childs exposure
from all sources (Institute of Medicine, Clearing the Air: Asthma and Indoor Air
Exposure, 2000).
In summary, within the context of social ecology, the household production of
health as a dynamic behavioral process combines both intra-household factors with
external health behaviors (resources, skills) and leads to health outcomes (Berman,
Kendall, and Bhattacharya, 1994). In this research, intra-household and community
factors, and inputs to these factors are examined using a qualitative approach. These,
along with actual health behaviors related to smoke exposure, are measured through a
survey. Health behaviors are then compared to health indicators, through cotinine
analysis. Understanding the impact of events and conditions internally and externally
can reveal the domestic strategies that people employ to deal with factors in their
environment to constrain or promote health (Clark, 1998).
33


Review of Literature
Developing a model that includes household factors impacting smoke
exposure of young children is a daunting task. The epidemiological literature
establishing associations between ETS exposure and health effects focused on
predictors of ETS exposure, development of new and improved methods of
measurement of exposure, relationships of ETS exposure and measurement with
incidence, prevalence and morbidity of illnesses. Many studies were population-
based utilizing data from national surveys (NHIS, NHANES) with subjects of all
ages, while others focused on pre-school or school-aged children. Studies were
reported from the US, Great Britain, Scandanavian and Nordic countries, Europe,
Africa, New Zealand, and Australia. While becoming more methodologically
rigorous over the past decade, study designs include retrospective, cross-sectional,
case cohort, longitudinal, and a few randomized controlled clinical trials. Analysis
included univariate, bivariate, and multivariate, as well as linear regression, logistic
regression, correlation and analysis of variance. Relationships between smoking by
intensity and measured exposure through cotinine analysis were found in most studies
in which measurement occurred. Behavioral factors were identified less consistently.
However, knowledge, attitudes, and beliefs, and health protection of children were
routinely identified. The following review further expands the epidemiological
components of smoke exposure, identifies smoking policies and practices in
34


households, and explores the impact of knowledge, attitudes, and beliefs on behaviors
related to smoke exposure.
Epidemiology of Smoke Exposure
Morbidity and mortality associated with tobacco use was first recognized in
the early twentieth century. The effects of tobacco smoke exposure on the nonsmoker
was recognized as early as 1928 (Doll, 1998). Medical evidence of the harm done by
smoking had been accumulating for two hundred years but was largely ignored until
five case control studies were published relating smoking to development of lung
cancer in 1950. Studies in the next two decades showed that active smoking was
associated with other diseases as well (Doll, 1998). In the early 1970s, mounting
evidence linked parental smoking to increased risk for more severe lower respiratory
illnesses during the first years of life. The first studied effects of ETS exposure in
children were increased risks of ear, nose and throat diseases which evidenced
themselves as a result of having spent part of a Sunday in a smoke-filled car (Sasco
and Vainio, 1999). Colley (1971) found an increased risk of bronchitis and
pneumonia in children during their first year of life if their parents smoked.
In the late 1970s, reduced lung function in children associated with smoking
in the home was reported, and confirmed over the next decade. The Surgeon
Generals report on Health Effects of Involuntary Smoking (1986) declared a causal
relationship between involuntary smoking and lung cancer. US Environmental
35


Protection Agency (1992) classified tobacco smoke as a known carcinogen. Rubin
and Damus (1988) reviewed studies investigating possible associations between
passive smoking and health effects in children to explain the wide range of effects
reported in the literature. They concluded that studies confirmed the effects of
passive smoke on child health but more methodological rigor was needed to delineate
the dose-effect relationship of the toxin.
Additional adverse health effects have now been linked to involuntary
exposure of children to tobacco smoke: increased prevalence and exacerbations of
asthma (Cook and Strachan 1999; DiFranzia and Lew, 1996; Joad, 2000; Wahlgren et
al., 2000), upper respiratory irritation (Cook and Strachan, 1999; Joad, 2000),
decrease in lung function (Charlton, 1994;Cook and Strachan, 1999; Sasco and
Vainio, 1999), middle ear disease (Ey et al., 1995; DiFranzia and Lew, 1996; Gaffney
and Lynch, 2000), Sudden Infant Death Syndrome (Charlton, 1994; Joad, 2000;
Samet, 1999), childhood cancer (Sasco and Vainio, 1999) and lung cancer (Charlton,
1994; Cook and Strachan, 1999; DiFranzia and Lew, 1996; Doll, 1998; Ey et al.,
1995; Li et al., 1999; Norman et al., 2000; Strachan and Cook, 1998). These reviews
include both systematic, quantitative meta-analyses (Strachan & Cook, 1998; Ey et
al., 1995) and narrative reviews as developed by EPA (1992) and WHO (1999).
Overall, Cook and Strachan (1999) identify a consistent picture with odds ratios for
respiratory illness and symptoms and middle ear disease between 1.2 and 1.6 for
either parent smoking. Odds ratios in children age 0-2 years (1.55 [Cl 1.16-2.08])
36


were higher than in pre-school children who in turn were higher than school age
children. Cook and Strachan (1998) further suggest a dose-response relationship as
evidenced by studies where odds ratios for asthma prevalence are higher when both
parents smoke (OR 1.5[CI 1.29-1.73]) than when just mother smokes (OR 1.36 [Cl
1.20-1.55]) or just when father smokes (OR 1.07 [Cl 0.92-1.24]). There is no
evidence identifying at what, if any, level of ETS exposure for a child could be risk
free (Institute of Medicine, 2000).
The health effects of smoke exposure on young children have been well
established; other factors that strongly influence actual smoke exposure of young
children have been explored. Community factors include demographic phenomena,
economic costs, social support and cultural norms resulting in expected normative
behaviors. Intra-household factors reviewed are smoking policies within households
and in vehicles, actual smoking practices in households, and the knowledge of
environmental tobacco smoke exposure. Other influences (inputs) are beliefs,
attitudes toward ETS exposure impacting these factors.
Passive Smoke Exposure in Homes: Intra-Household Factors
Childrens vulnerability to ETS and the difficulty that they have in protecting
themselves from the threat imposed by adults places the burden of reducing or
eliminating ETS exposure on adults in protective capacities (Ashley and Ferrence,
1998). Most ETS exposure of young children occurs at home with parents who
37


smoke or allow smoking in their home. Maternal and paternal smoking habits affect
ETS exposure of children. (Cook et al,1994; Willers, Axmon, Feyerabend, Nielsen,
Skaiping and Skerfving, 2000). Dose-response relationships that were associated
with the amount smoked by both parents as measured by cotinine levels in children
were reported by Cook et al, (1994). Mothers were less likely to smoke than fathers,
but when they did, the effect on their childrens cotinine concentrations was greater.
Mothers smoking status was the most important predictor of urinary cotinine
concentration (Cook et al, 1994; Jordaan et al, 1999: Willers et al, 2000).
A significant relationship between cotinine levels of children and the total
reported amount of tobacco smoked indoors by parents and others was also found.
Jordaan et al, (1999) found maternal smoking to account for 21.8% of variation in
urinary cotinine studies of school age children while male parents or other household
smokers accounted for 12.7% of variation. Bahceciler, Barlan, Nuhoglu, and Basaran
(1999) evaluated the effects of parental smoking modification on cotinine levels of
children. Although they had a very small group of subjects (n=77), they found that
children whose parents reported smoking indoors had significantly higher cotinine
levels than those in whose homes there was no smoking (p<.001). Children whose
parents reported smoking on the balcony also had significantly higher cotinine levels
(p<.002) than those in homes with no smoking. Parental reports of exposure and no
exposure were consistent with urinary cotinine levels in 84% and 82% of children,
respectively. They concluded that there was a strong relationship between parental
38


reports of smoking indoors and urinary cotinine levels but between parent reports of
smoking outdoors and cotinine levels.
Emmons et al. (1994) compared levels of exposure of adult non-smokers who
lived with a smoker and non-smokers who lived with non-smokers when smokers quit
smoking. Passive nicotine monitor measurements showed significant reduction in
ETS exposure among non-smokers who lived with smokers who quit. Borland
(1999) reported cotinine concentrations in children to be affected by frequency with
which parents smoke in the same room as the child. He also found that presence of
open ventilation and smoking only in restricted home areas reduced cotinine
concentrations in the child. Smoking by household members other than parents and
smoking by visitors resulted in increases in childrens cotinine levels but the
magnitude of effect was small compared to parents.
Community factors including exposure from outside of the home have also
received some attention in the literature. Jordaan et al, (1999) quantified community
exposure through measuring exposure of children at school and found that it
accounted for 3.3% of variance in cotinine levels. Ownby, Johnson, and Peterson
(2000) investigated passive smoke exposure of infants (birth to 2 years of age) from
parents and other sources of exposure i.e. daycare workers, visitors in the home, and
residents of the home other than childs parents in a longitudinal study conducted
over a two year period. Data were analyzed to determine relative contributions from
different groups of smokers to urinary cotinine concentrations in the infants.
39


Frequency of smoking by those other than parents in the household was a significant
contributor to urinary cotinine concentrations of infants. A highly significant
correlation existed between maternal smoking and quantity of cotinine in an infants
urine; furthermore, smoking by other adults was also significantly correlated with
urinary cotinine measurements.
Smoking Policies in Households
Home smoking restrictions and household bans are a new and important area
of research as these are increasingly recognized as the most effective steps that
parents can take to reduce ETS exposure of children (Gilpin et al, 1999). Gilpin et al
(1999) analyzed Tobacco Surveys (1996) in California looking at smoking status and
behavior, household smoking restrictions, and other social variables. Sixty-four
percent of households surveyed had a total or partial smoking ban. The presence of
non-smokers in the household was a major determinant of whether or not the home
was smoke free. Smokers with children in the household were more likely to have
smoke-free households than smokers with no children, and the younger the age of the
youngest child, the more likely the home was to be smoke-free. Households with
both a child and an adult were 5.7 times more likely to be smoke-free than households
with neither. A belief in the harmfulness of second hand smoke was also related to
smoke-free homes. The results of this study suggest that tobacco control policies
promote smoke-free homes. A similar study in Oregon (1997) investigated
40


households according to type and degree of smoking restrictions and explored
whether smoking restrictions are associated with decreased environmental tobacco
smoke exposure using a cross-sectional telephone survey (Pizacani et al, 2002). This
study also found that the presence of children in the home (OR=4.6) and awareness
of the harm of ETS (OR =12.8) were closely associated with full bans on smoking in
the home. In contrast, 50% of households with a smoker and children did not have a
full ban on indoor smoking.
Okah, Choi, Okuyemi, and Ahluwalia, (2002) found that home smoking
restriction was associated with presence of children (P<.0001), and a non-smoking
partner in the home (P=.002). Restrictions were not associated with age, gender, race,
education number of best Mends who smoke or perceived harm from smoking which
have been previously identified as impacting smoking in homes. This suggests that
inner city smokers are concerned about the effects of ETS and take steps to limit
exposure in their children. Furthermore, significantly more steps are taken to limit
ETS exposure when there is a nonsmoking partner in the home.
Demographic Characteristics
Smoke exposure in households is related to socioeconomic status of families,
level of education, income and family structure, age of children living in home. Low
income families may encounter more difficulty providing smoke free environments
for their children as they are more likely to smoke, associate with people who are also
41


smokers, live in small housing units which may have limited access to the outdoors,
shared rooms, and ventilation systems (Ashley, 1999). Eriksen and Bruusgaard
(1995) found, in a cross-sectional study in Norway, that parents were less likely to
smoke if they were more than 35 years of age, had a child less than one year of age,
had a spouse, and had a long education. Smoking parents also smoked less if they
had a spouse/co-habitee, had a child under one year of age or had few children.
Arborelius, Hallberg, and Hakansson (2000) reported that the Swedish
Medical Birth Registry showed a significantly higher proportion of women with less
education and from lower socioeconomic groups that were smokers. Lund, Skrondal,
Vertio, and Helgason (1998) likewise found in a population based study in Nordic
countries that parents with lower socioeconomic status and single parents expose their
children to more ETS, in spite of the fact that they are just as likely to report having
tried to change their smoking behavior because of their children. Whitlock et al.
(1998) found an inverse relationship between socioeconomic status and smoke
exposure. He assessed ETS exposure through a self-report survey among adults and
measured socioeconomic status through educational level, occupational status, and
median household income. Two measures were reportedthe number of hours per
week spent near someone who is smoking and prevalence of regular exposure to ETS.
Both were inversely associated with all three indicators of socioeconomic status
(pO.OOOl). Occupational level and educational level showed steeper associations
with ETS exposure than median income. Jordaan, Ehrlich, and Potter (1999) found
42


that socioeconomic indicators explained 4.8% of variation in cotinine levels of a
population study including children age 6-11 years in South Africa. Interaction
affects were found between smoking variables (mothers smoking) and
socioeconomic variables (number of people living in the home) (F=4.24, P=015).
They also found that the number of people who lived in the house but not density
(number of individuals/room) was a significant predictor of cotinine levels.
Mannino (2001) analyzed NHANES III data (1988-1994) in an attempt to
identify predictors of cotinine levels in young children age 4-16 years. He found that
demographic factors such as age, race/ethnicity, poverty status, and region of the US
predict cotinine levels in children with the strongest predictor being the reported
number of cigarettes smoked in the home daily. Teen-agers without smoke exposure
at home had higher cotinine levels than did younger children who had no smoke
exposure at home, suggesting that they were exposed to sources outside the home.
Health Care Utilization and Economic Costs
Closely associated with the relationship of passive smoke exposure and health
affects in children, is increased medical care utilization by children and greater
related costs such as prescription drugs for respiratory complaints, more school
absences, and increased health care costs (Archives of Family Medicine, 1994;
Stoddard and Gray, 1997; Wewers and Uno, 2002). Passive smoke exposure is
associated with 19% of all expenditures for childhood respiratory conditions
43


(Stoddard and Gray, 1997). DiFranizia and Lew (1996) estimated that 15%-25% of
all hospitalizations of infants and children with lower respiratory illnesses were
associated with ETS exposure. Active smoking is known to vary by socioeconomic
and educational status with greater representation among low income and less
educated (Ashley, 1999) suggesting that children living in these households have
greater risk for adverse effects related to smoke exposure.
Lam, Leung, and Lai-Ming (2001) examined the association between ETS
patterns and doctor consultations and hospitalizations of infant participants in a
longitudinal study in Hong Kong. Data were collected from families at visits when
the infant was 3 months, 9 months, and 18 months. Using multivariate logistic
regression analyses in a population based study, they found a higher level of doctor
consultation visits for respiratory and febrile illnesses (P<.001) in infants who had
been exposed to ETS in utero and high hospital admission rates among infants
exposed to ETS either before or after birth. They also demonstrated a clear dose-
response gradient between the total number of smokers at home and increased
hospitalizations for respiratory and febrile illnesses (P=.0003) and any illness (P
<.001). Economic ramifications of increased utilization were not addressed in this
study; however, increased utilization has economic consequences for the health care
system as a whole (Lam et al, 2001).
44


Knowledge of Health Effects of Smoking
Knowledge of health risks of smoking and passive smoke exposure have
increased substantially over the past several decades. Population data collected from
high-risk urban population in St. Louis and Kansas City, Missouri (Brownson et al.,
1992) show that a majority of population were aware of health hazards of smoking
(78.2%) and 87.2% knew that passive smoke exposure was hazardous to young
children. The belief in harmful effects was inversely related to age and positively
correlated with educational level. Current smokers were significantly less likely than
never smokers to acknowledge the effects of passive smoking on non-smokers health
(OR = 0.5) and were less annoyed by passive smoke. Knowledge of health effects of
smoking and passive smoking were lower among older age groups, women,
respondents with less education and current smokers. In a randomized clinical trial,
Sorum and Bruusgaard (1996) tested the effects of an anti-smoking information
program with smoking parents of young children during well-child visits. Information
(minimal intervention program) was given to one group of parents by a health visitor
and the control group received no information unless they asked. No differences
were found in smoking behaviors of the two groups at a one-month follow-up after
the interventions. There was possibility of contamination in this study and a large
attrition rate that may have impacted results.
Several studies have investigated the relationship between knowledge and
behaviors that protect children from ETS. Goldstein (1994) found that although
45


households had knowledge of harmful effects (90%) of ETS, only (24%) had actual
controls (24%) in place, restricting smoke exposure.
Household Behaviors. Attitudes and Smoke Exposure
Results of studies in which attitudes, beliefs, and practices in homes with
small children have been investigated show that awareness of the dangers of passive
smoke of children is increasing (Al-Delaimy et al, 1999; Ashley et al, 1998; Eriksen
and Bruusgaard, 1995; Goldstein, 1994; Norman et al, 1999). These same studies
show that from 24% (Goldstein, 1994) to 76% (Norman et al, 1999) of households
surveyed have some smoking restrictions in the home. In spite of the increasing
awareness of dangers of passive smoke exposure, Ashley et al. (1998), in analyzing
population based surveys in Canada, reported that 34% of homes with non-smokers
and 20% of homes with daily smokers were reported to be smoke free. Of these, non-
smokers who thought that parents should not smoke when spending time with
children increased from 62.6% in 1992 to 78.0% in 1996. Smokers answering the
same questions increased from 16.7% (1992) to 42.6% (1996). Results of a telephone
survey in New Zealand examining attitudes about protection of children from passive
smoking (Al-Delaimy et al. (1998) found similar discrepancies. Both smokers and
non-smokers indicated that smoking at home and in private cars was almost as
unacceptable as smoking in public places with children around. Fewer than two-
thirds of the smokers reported that they refrained from or reduced their smoking in
46


presence of children. The high prevalence of smoking in the presence of children in
spite of knowledge and changes in attitudes indicates the continued need for
investigation of factors influencing smoking behaviors resulting in ETS exposure.
Feamow et al. (1998) examined parent activism (how much parent
discourages, talks about, and monitors child smoking) and parental permissiveness
about children smoking at home in homes containing teen-age children. Data were
obtained from a cohort-sequential study of cigarette smoking from adolescence to
adulthood. Study variables included smoking status, educational attainment of
parents, parental values concerning child nonsmoking, health beliefs about smoking,
parents belief about addictiveness of cigarette smoking and stress. Four outcome
variables in which parents were asked how they would deal with a childs smoking
were measured on 5 point scale; three of these were combined to measure parental
activism. Another measure, parent permissiveness, was measured as a single item
outcome. Correlations between study variables and parental beliefs about health
consequences of smoking were associated with parental activism but not
permissiveness. Variables were examined through three regression models including
health beliefs, perceptions of addiction, and environmental stress. Health dangers of
smoking were more strongly correlated with smoking activism at high levels of
parental education. Health beliefs about smoking and parents values on childs non-
smoking were only marginally significant. Relationships between parent values and
actions were stronger for parents who had negative health beliefs about smoking than
47


for parents who had more positive health beliefs about smoking. Marginally
significant interactions between parents smoking status and their perceived personal
health in predicting parent permissiveness. Smokers who perceived themselves as
healthy were more permissive of child smoking; ex-smokers who perceived
themselves as healthier were less permissive of their childs smoking. Personal
health status was unrelated to parent permissiveness for non-smokers.
Arborelius, Hallberg, and Hakansson (2000) investigated methods of
preventing smoke exposure in small children focusing on effectiveness of
interventions to determine what measures were most effective. They reviewed studies
in which varying interventions were used with parents who smoked and evaluated
outcomes of these interventions. Demonstrable effects were found for interventions
geared to behavioral strategies with parents, patients beliefs about smoking and
effects on children, counseling efforts for stress reduction, strengthening parental self-
efficacy. No effects were found for interventions focusing on providing factual
information about dangers of ETS exposure or interventions in which pediatricians
provided feedback regarding cotinine levels of children exposed to ETS. They
concluded that interventions effectively focus on provision of a smoke-free
environment for children and not on helping parents stop smoking.
48



Social Support and Smoking
The discrepancy between the awareness of the dangers of tobacco smoke
exposure and the smoking behaviors of adults in the presence of children suggests
that other factors must be addressed if ETS exposure of children is to be reduced.
The prescriptions and proscriptions of parents, friends, and peers have shown to be
related to the use of tobacco. Goldstein (1994) investigated smoking rules within
homes through informal smoking controls which he defines as the rules about
smoking that individuals impose on one another in their everyday lives. He found
that informal smoking norms were directly related to negative attitudes toward
smoking and smokers, and to believing that ETS is harmful. He also found that
informal smoking controls were less likely to be found as the number of friends who
smoke increased and when the partner of the respondent smoked. Formal controls
(smoking restrictions and/or smoking bans in public places including work sites) have
been emphasized in ETS exposure much more than informal controls. Less attention
has been paid to the factors that impact how informal controls are developed and
articulated. The sanctity of the family unit restricts the ability of policy actions to
diminish tobacco exposure in the home; the political will to monitor, enforce, or
change behavior related to tobacco smoke exposure is also lacking. Other efforts are
necessary if providing smoke free home environments for all children are to continue
and increase (Bek, Tomac, et al, 1999).
49


Erikson and Bruusgaard (1995) investigated social support as a variable in
smoking prevalence among parents with young children. A longitudinal study in
which smoking behaviors of parents of small children were assessed at age of 6
weeks, 2 years, and 4 years by self-report of parents. Parents were categorized into
levels of social support (low, medium and high) and associations between levels of
support and smoking behaviors were investigated. Daily smoking was not related to
level of social support. Associations were found between high social support and
smoking less than ten cigarettes per day in parents with several children. Smoking
ten or more cigarettes per day was associated with medium and low social support. A
significant interaction was found between the level of social support and the number
of children in the family. Smoking parents were also less inclined to smoke indoors if
they had high social support suggesting the importance of strong and available social
support in reducing passive smoke exposure.
Self-efficacy
In addition to knowledge and attitudes, self-efficacy has been identified more
recently as an important factor related to the prevention of passive smoking (Crone et
al, 2001). Mothers lack confidence in their ability to ask others to refrain from
smoking around children. Arborelius et al, (2000) report that in Sweden interventions
geared to behavioral strategies resulted in demonstrable effects in reducing passive
smoke exposure of children. Specifically, a methodology aimed at strengthening self-
50


efficacy of parents demonstrated that 66% of parents involved in the study by self-
report were successful in reducing smoke exposure (Aborelius et al, 2000).
Oakley (1993), in a qualitative study investigating social support in pregnant
women, found that in addition to low income, unemployment, and single marital
status, smoking was associated with stress and crises involving family illness,
relationship problems, financial difficulties, and violence. All women expressed
awareness of the health effects of smoking regardless of their own behavior and
Oakley concluded that health promotion strategies need to do more than reinforce the
moral message to stop smoking for the sake of someone else (the baby). Women in
these situations were involved in what Oakley described as the paradox of health-
promoting work that may be health-damaging to them; providing an environment
conducive to the health of their children came at the expense of their own health.
Interventions for Smoke Exposure
Studies of attempts to reduce passive smoke exposure of children report
varying success rates. Greenburg et al. (1994) provided an education intervention
based on social learning theory delivered over a 6-month period to mothers in their
homes. Intervention effects were studied for smoking mothers and non-smoking
mothers. Differences in self-reports of smoke exposure between smoking and non-
smoking mothers were reported by not supported by similar differences in cotinine-
to-creatinine ratios of children. Hovell et al (2000) provided counseling for smoking
51


cessation to smoking mothers in a controlled clinical trial in which one group of
mothers received seven individual counseling sessions during a three month period
and a control group received brief advice to quit smoking and not to expose their
children to ETS. Mothers self-reports of tobacco use and childs exposure to ETS,
cotinine levels of urine samples from children, saliva cotinine concentrations from
mothers, nicotine monitoring to validate mothers self-reports were done at baseline,
three months, and twelve months. Exposure declined more significantly in counseled
group than in control group from baseline to three months (P=0.011) and at twelve
months (P=0.0002). Significant differences remained at twelve months but neither
group showed significant change over time, suggesting that counseling effect was
maintained but no later improvement occurred. Urine cotinine concentrations were
significantly different for the two groups (P=0.0008) with a slight decrease in urine
from children in families counseled over twelve months and a substantial increase in
cotinine in urine of children in control families. Both of these studies confirm the
efficacy of parental smoking related counseling to reduce childrens exposure to
environmental tobacco smoke.
Summary
For the purpose of this research, factors in households were explored and their
relationship to presence of household smoking bans and measurement of actual ETS
exposure were investigated. The household factors identified from the research
52


which warrant further investigation include smoking behaviors, demographic factors,
knowledge of health effects of ETS exposure, health promoting behaviors related to
smoke exposure in households, social support, and attitudes and beliefs regarding
health protective behaviors and ETS exposure. In the household production of health
framework, these include both intra-household factors and community factors and are
investigated equally in the study.
Very little research has occurred investigating environmental tobacco smoke
exposure as an issue of health within the household context. The household
production of health (HHPH) as a conceptual framework has been little utilized in
designing programs which focus on presence and maintenance of health related to
environmental tobacco smoke exposure. Rather, research in public health has
focused on social and behavioral sciences addressing only issues set by the
development of technology (Berman, Kendall, and Bhattacharya, 1994).
53


CHAPTER 3
METHODOLOGY
Overview
This research used qualitative methods in the form of semi-structured
interviews and observations, and quantitative methods in the form of a survey
questionnaire and measurement of cotinine as an indicator of smoke exposure. The
qualitative investigation provided description and understanding, thus informing
instrumentation used in the survey questionnaire. Sequential mixed methods,
including triangulation, were used to study behavioral issues related to tobacco smoke
exposure. Qualitative and quantitative procedures operate from different
assumptions and seek to answer questions in different ways (Miles and Huberman,
1994). They elicit different but complimentary kinds of information, which are
important to the understanding of the issues associated with smoke exposure in
households (Erikson and Kaplan, 2000). Qualitative description is a prerequisite of
good quantitative research, particularly in areas that have had little previous research
(Pope and Mays, 1995).
Parents have been encouraged to ban smoking in homes to reduce
environmental tobacco smoke exposure (Wakefield et al, 2000). However, it is
54


unclear whether restrictions and bans by themselves offer protection from ETS
exposure. Little information is available on how families negotiate smoke free
environments for their children. Qualitative methods investigated how families
address smoking issues in their households with children present, including
description and understanding of behaviors and attitudes regarding smoke exposure in
households. From these data, information emerged which identified and confirmed
factors to be investigated further through the survey questionnaire.
Quantitative methods in a descriptive, cross-sectional design examined factors
in households associated with smoking rules that affect environmental tobacco smoke
exposure of children, age newborn to five years. The design provided for quantitative
follow-up of discoveries (Crabtree and Miller, 1992) of the semi-structured
interviews through a questionnaire investigating factors and smoking
behaviors/smoke exposure in households. Reports of smoke exposure in the
household were validated through measurement of cotinine, a biomarker for nicotine.
The design allowed further investigation of a model examining health factors
associated with the presence of home smoking restrictions, using the household
primarily as a unit of analysis. In such a context, smoke exposure in a childs
environment could be more accurately measured. As factors were identified and
quantified, this design allowed for examination of possible relationships between
factors and smoking policies in households.
55


Qualitative Methods
Qualitative methods in this study provided description and increased
understanding of smoking behaviors in households to explicate ways that people in
particular settings come to understand, account for, take action, and manage their
day-to-day situations (Miles and Huberman, 1994). The identification of ideas,
values, and beliefs and how people draw on them to make sense of situations, actions,
and processes in their lives were explored through semi-structured interviews
(Williams, 1998).
According to Pope and Mays (1995), qualitative techniques provide
description and understanding of situations or behavior preceding more quantitative
investigation, are essential in triangulation for validation purposes, and provide a
means for exploring complex phenomena not amenable to quantitative research. The
inclusion of more qualitative strategies in smoking research provides information on
the social and cultural context and personal meaning of smoking necessary for the
development of prevention and cessation strategies (Erickson and Kaplan, 2000).
Semi-structured interviews were conducted to explore behaviors, attitudes,
and experiences regarding smoking in households. The interviews attempted to elicit
contextual meaning of smoking from respondents as a prerequisite for the
identification of variables and testing relationships between them (Erickson and
Kaplan, 2000). The use of semi-structured interviews combined the advantages of
obtaining the responses needed through a structured interview approach and the
56


breadth and richness afforded through a more human-to-human relationship of
unstructured interviewing (Denzin and Lincoln, 1998). The interviews attempted to
increase understanding about how families really behave related to smoke exposure
in their homes and to describe their experiences and associated meanings regarding
smoke exposure in the context of their household. Household smoking rules or
*
policies, implementation and enforcement of these policies, and factors which
families identified as being important in this process were explored. Dialogue with
families provided opportunity for rich description and detail, strategic comparison
across cases, and to initiate new lines of thinking (Miles and Huberman, 1994).
In addition to the semi-structured interviews, observation was used as an
adjunct method of data collection. This involved systematic, detailed observations of
behavior and talk (Mays and Pope, 1995), and watching and recording the subjects
behaviors and interactions. The concerns with validity and reliability with the
observational techniques able to be employed in this study limited its usefulness as a
data collection method. There was no member validation of findings; consistent
trends were sought in each situation to ensure reliability. The original intent was to
conduct the activities in the natural home setting to facilitate observation, but only
three households agreed to be interviewed at home. Consequently, observation
provided limited information about the working and functioning of individuals within
households related to smoke exposure.
57


Issues of Validity
Qualitative approaches require specific aims and a clear purpose to guide data
collection and systematic analysis of the data. Included are explicit sampling
strategies, systematic analysis of data, and a commitment to examining counter
explanations (Green and Britten, 1998, Pope and Mays, 2000). Procedures to ensure
validity and reliability of qualitative findings have been the subject of much
controversy. Appropriate criteria and how they should be assessed have been topics
of much debate.
Lincoln and Guba (1985) proposed criteria for establishing measures of
validity and reliability in qualitative research. These include: credibility,
confirmability, dependability, and transferability. Credibility is associated with
internal validity, confirmability with objectivity, dependability with appropriateness
of science behind the method, and transferability with generalizability (Lincoln and
Guba, 1985). In a similar fashion, Pope and Mays (2000) recommend that relevance
and validity be considered using the following criteria:
1. Triangulation involving the use of two or more methods of data collection.
2. Respondent validation in which researchers findings are reviewed and
compared with research subjects and experts in the field.
3. Clear exposition of methods of data collection and analysis.
4. Reflexivitysensitivity and assumptions of researcher in data collection.
5. Attention to negative casesdeviant case analysis.
58


6. Transferability of findings to other settings.
Consistent with these recommendations, this analysis meets the criteria for
triangulation, expert validation, and clarification of researcher bias. Observations, in
addition to interviews, were used to overcome the discrepancy between what people
say and what they actually do; they circumvent the biases inherent in the accounts
people give of their actions caused by factors such as the wish to present themselves
in a good light (Mays and Pope, 1995). Member checks were conducted with two
nurse colleagues and one PhD expert researcher with whom the researcher worked
during data collection. Respondent validation with research subjects was not possible
because of the confidential nature of the interviews.
Data analysis involved referring back to original transcripts and reviewing
coding and themes with conversations of the participants, using Atlas/ti software.
The researcher also carefully reviewed differences in responses of households that
allowed smoking and those who did not.
Interview Instrument
Questions composing the semi-structured interview (See Appendix A) were
based on the household production of health framework, focusing on intra-household
and community factors. Specific content was organized using the HHPH framework
including intra-household factors, inputs to health, and community factors. Once
developed, the questions were reviewed by an expert researcher for validity and
59


accuracy. After revisions, the finalized version of the interview was conducted with
one household before actual data collection began to test the usability of the guide for
questions.
Questions in the interview were sequenced to facilitate and maintain rapport
and good feelings between the interviewer and the respondent (Jones, 1996) to elicit
as much description about smoking in households as possible. A modified funnel
sequence (Jones, 1996) was used, progressing from general to more specific
questions, which were related to questions before them. The interviews began with
general questions focusing on health activities in household. More specific inquiries
followed about promoting health of children and protection from harm. Next, there
were questions on health protective behaviors in the household, and then on sensitive
topics based on the assumption that sufficient time for gaining trust and establishment
of rapport had occurred (Fontana and Frey, 1998). Smoking behaviors and practices
in the home, attitudes and/or beliefs related to smoke exposure, health protection and
smoke exposure, rules for smoking in the home, and methods of implementation
comprised the more sensitive topics to be included (See Appendix A).
In addition to open-ended questions of the semi-structured interview,
household observations were made using a checklist (See Appendix A). The size of
dwelling as measured by the number of rooms, number of people living in the home,
number of smokers living in the home, and evidence of smoking behaviors in home
(ashtrays, cigarette paraphernalia, smell of cigarette smoke, etc) were recorded.
60


These data were obtained at the end of the interview. In instances where observation
in the home was not possible, the researcher asked questions to obtain information. A
contact summary form was used to collect demographic data, write short field notes,
and identify preliminary themes of the interview. Also included were field notes of
visual observations of responses and activities occurring during the interview for all
interviews. Respondent behaviors and visual expressions were recorded as field notes
immediately following the interviews.
Samnle Selection for Interviews
Through systematic sampling, twenty families in households were selected
using the following inclusion criteria:
1. At least one child under school-age (age 5) lived in the home.
2. Children in the household were registered to receive services at
school-based health centers associated with the School of Nursing Faculty Practice at
University of Colorado Health Sciences Center.
3. Participant of household being interviewed could understand,
converse, and write in English.
Families were asked to participate in the study either by telephone or when
they utilized the school-based health center for services. The researcher explained the
purpose of the study and invited participation, carefully explaining that it made no
difference for research purposes if anyone in the home smoked. As the interviews
61


progressed, the researcher selected families to insure that households with smoking
members and households with no smoking members were included in the sample.
Structure of the Interview
An adult member, usually a parent, was the spokesperson for each of the
households that participated. After completion of the written consent, the researcher
verbally confirmed with the person being interviewed that he/she agreed to being
taped and that the researcher might make other notes as they talked together. The
respondent was also informed that he or she could stop the interview at any time. The
interview began with basic general questions, proceeding to more specific questions,
and subjects were encouraged to share experiences related to their responses. Probes
such as direct questions, repetition of questions, and silence were used as needed to
encourage respondents to continue, amplify, or clarify answers (Jones, 1996).
The nature of these interviews was such that the person being interviewed
became a spokesperson for the household, in keeping with the study design in which
the household was established as the unit of analysis. The researcher directed
questions toward activities and behaviors of the household; however, it was not
always possible to ensure that the responses reflected the characteristics of the
household as a unit. In households where there were two parents, both parents were
invited to participate in the interviews.
62


The interviews included but were not limited to the topics of discussion
outlined in the semi-structured interview instrument. Questions were asked in similar
order with potential variations if the family member being interviewed addressed
questions before the interviewer asked them. All questions included in the instrument
were asked; respondents could choose not to answer if they so desired.
Because families with young children were selected for these interviews,
children were usually present during the interview, sometimes interrupting the
dialogue between the researcher and spokesperson. Although these interruptions
interfered with the interview process, they provided opportunity for making notes and
observations about the interview. Interviews lasted from twenty minutes to one hour.
Role of the Researcher
The researcher explained in detail differences in her role as a researcher and
as a care provider for the participating families so that subjects could share truths
regarding their households without being evaluated or judged. Because many of the
families were familiar with the researcher as a care provider, establishment of rapport
was not problematic. However, the familiarity and comfort level of the family with
the researcher as a care provider led to potential bias in the interviews. To address
this, the interviewer explained her role as a researcher as being one of hearing and
understanding how families deal with smoking in their homes. The researcher
prefaced each interview with a statement indicating that how each family deals with
63


smoking is a household decision. She reiterated several times that, as a researcher, it
made no difference to her whether or not there was smoking in the home, and that she
wanted to talk both with families who allow smoking and those who do not. It was
necessary to repeat and/or interpret that objective role in interviews. The researcher
affirmed how families managed their health, encouraged expression of both positive
and negative perceptions, attitudes, and feelings, and asked for further explanations
on various answers.
Subjects frequently asked the researcher for her opinion regarding the
questions. The researcher responded by stating that the purpose of the interview was
not to provide her opinion about smoking issues but to discover how families deal
with smoking in real life. For some interviews it was necessary to reiterate the
importance of the family behaviors repeatedly until they became comfortable sharing
their information.
If the person being interviewed asked for the opinion of the researcher, she
responded by stating that she was more interested in hearing the opinions of the
person being interviewed. Each interview ended with the question, Is there anything
else that would be helpful for me to know about you and your family? On several
occasions, the family asked for information about smoking cessation or the dangers of
environmental tobacco smoke exposure after the interview was complete and the tape
recorder turned off. Several families asked, So what does smoke exposure really do
64


to children? The researcher provided information as requested and in several
instances provided resources on smoking cessation after the interview was completed.
Data Collection
After obtaining informed consent from the subjects including permission to
audiotape, the researcher audio taped the interviews; they were later transcribed and
analyzed. The researcher completed all transcription to insure that the data were
accurately transcribed; many of the tapes contained background noise making
understanding the dialogue difficult at times. Numbers were assigned to study
participants for transcription purposes. Names mentioned in the taped interviews
were omitted in the transcription, identified instead by individuals role in the
household (i.e. husband, child etc.). Once transcribed, data were stored in a
computer, protected by password. A form used for tracking purposes including
participant name, contact information, and number of the interview was attached to
the informed consent and kept with research files. This was completed prior to the
interview and utilized when necessary to contact households if changes in time and
date of interview were necessary. Subjects were given one copy of the informed
consent and the other copy was retained in a locked filing cabinet.
65


Quantitative Methods: Survey Research
A survey questionnaire was developed to examine the household factors
identified in literature and through the semi-structured interviews with a larger
sample. Self-reports obtained through the survey were validated by cotinine
measurement of urine samples from a child in the household. No names were placed
on surveys to encourage truthfiil responses to questions which might be considered
sensitive and which household members might answer as they perceived the correct
answer to be. Each survey was numbered so that cotinine measurement of urine
could be compared with survey results. The survey number was recorded on the
specimen cup for urine and on the consent form completed prior to collection of urine
so that accuracy of records could be maintained and results could be shared with
parents if they requested. Consent forms and surveys were stored separately so that
surveys and cotinine results remained confidential.
The study instrument investigated individual and household demographics,
smoking behaviors in households, home smoking restrictions, knowledge of harms
and health effects of smoke exposure, and attitudes/beliefs about smoke exposure.
Variables examined in the survey included behavioral norms and patterns in
households as measured by individual and household smoking behaviors, health
status of children in household, household smoking bans, attitudes/beliefs related to
smoke exposure, knowledge of smoke exposure, and demographics.
66


Outcome variables were selected based on research questions; smoking bans,
used to measure smoking policies of households, were categorized by self-report, and
actual smoke exposure of household was measured by cotinine level in urine samples.
Household smoking bans as a predictor variable were compared with cotinine level of
urine. Predictor variables included both individual and household measures.
Parental reports of exposure of older children (over 6 years) dependent on
memory and physical parameters have demonstrated acceptable reliability and
validity in surveys (Fried et al, 1995). The same has not been documented for young
children and infants because of the perceived social desirability not to expose infants
to ETS and limitations of the reporter in documenting accurately the duration,
proximity, and frequency of exposure (Matt et al, 1999). In this study, memory-based
reports by parents were validated by measurement of cotinine, a biomarker of
nicotine, providing a quantitative measure of smoke exposure in their children.
Instrument Development
Development of the survey instrument followed procedures recommended by
Aday (1996) and Czaja and Blair (1996) and included survey design, preliminary
planning, and pre-testing. A continual process of development, testing, and revision
resulted in the final survey instrument (See Appendix B).
67


Items in the survey were selected based on household production of health
framework within a human ecology context. They were divided into three main
groups based on definitions of household production of health framework (HHPH):
1. Intra-household factors included actual smoking behaviors, both
individual and household. Individual measures include smoking status of
respondent. Household smoking practices include number of smokers living in
household, number of cigarettes smoked in home in past week, last time smoke
exposure occurred in household, situations and/or locations in which smoking is
permitted in the household, exceptions to smoking policies in household.
2. Inputs to household policies include knowledge of respondent,
attitudes/beliefs of respondent, and factors identified in HHPH that influence smoking
behaviors.
3. Macro-level factors included race/ethnicity variables, household size as
measured by number of residents in home and number of rooms in household,
household income, gender of respondent, educational level of parents, health status of
children in household.
In addition to the framework of the household production of health,
individual survey questions were informed by results of the semi-structured
interviews. Sixty-seven questions included multiple choice, fill-in-blank, check all
that apply, yes and no options, true/false, and identification of attitudes using a 5-
point Likert scale. Included in the 5-point Likert scale was a neutral category to
68


allow respondents to indicate a more neutral attitude toward the four components of
smoke exposure. Aday (1996) suggests that scales with five to seven points are more
valid and reliable than those containing fewer categories. The survey incorporated
items from previous interviews with known validity and reliability; new questions
were included based on literature review and results of qualitative analysis. Smoking
behaviors in households and home smoking restrictions questions were published
previously by the National Health and Nutritional Examination Survey and by the
Centers for Disease Control (MMWR, 1999).
A self-administered questionnaire was selected for this study to facilitate
disclosure of information which respondents might otherwise be hesitant to provide
due to the somewhat sensitive nature of the topic and their desire to present
themselves in the best way possible. Because information was collected from
subjects, some of whom read English and some of whom only read Spanish, the
survey was developed at a fifth grade reading level using the SMOG readability tool
(US Department of Health and Human Services, 1989) to facilitate completion by
subjects with little or no assistance from the researcher. The survey was finalized in
English and transformed into teleform format so that it could be scanned for data
entry. The Spanish survey could not be transformed because of the lack of accents
and tilde in the computer program. Care was taken to ensure that each survey item
had comparable responses in Spanish and English to minimize bias and facilitate
statistical description and analysis. A cover letter informed the subjects that by
69


completing the survey, they were providing permission to participate in the study.
This further facilitated self-administration of the survey.
Measures based on similar items used in previous research (Norman et al,
1999) relating to household factors and smoking practices were incorporated into the
self-administered questionnaire. Some measures were assessed as household
measures; others could only be measured at the individual level. The framework for
the research established household as the unit of analysis, and this was used whenever
possible in reporting results and findings. When characteristics could only be
measured as individual factors, they were reported as such in the summarization and
analysis.
Items were grouped into variables, either predictor or outcome. Outcome
variables include self-reported smoking policies in households: no smoking allowed
(complete ban), smoking allowed in some places or at some times (partial ban), no
smoking ever allowed (complete ban). Cotinine measure of smoke exposure in
household is used to validate identified smoking policies in households.


Table 3.1 Predictor and Outcome Variables in Survey
Domain Variable Level of Measures Proposed Measures Measurement
Outcome Household smoking policy Household Measure Complete ban, partial ban, no ban. Reported Parent measurement
Household smoke exposure Child Measure Cotinine measurement of urine Quantitative Child measurement
Predictors Demographic Individual Measure Age of respondent Gender of respondent Parent self report
Household measure -Educational level of parents -Race/ethnicity -Household income -Number of residents -Type dwelling -Number of rooms in dwelling -Perceived health of children living in household
Predictors Smoking Behaviors Individual Measure -Smoking status of respondent Parent self report
Household measure -Number of smokers living in household -No. of cigarettes smoked in house in past week -Last time in which smoking occurred in household -No. of rooms in which smoking allowed -Situations in which smoking allowed in household Parent self report
71


Table 3.1 Predictor and Outcome Variables in Survey (Cont.l
Domain Variable
Level of Proposed Measures
Measures Measurement
Predictors Smoking Individual
Behaviors Measure
Household
measure
Attitudes/Beliefs Individual
Measure
Knowledge Individual
Measure
-Smoking status of respondent Parent self report
-Number of smokers living in household -No. of cigarettes smoked in house in past week -Last time in which smoking occurred in household -No. of rooms in which smoking allowed -Situations in which smoking allowed in household Parent self report
-Attitudes toward smoke Parent self report
exposure -Attitude toward health protection of children -Attitudes toward legal control on smoke exposure -Attitudes toward smoke exposure of children Likert scale
Health effects of smoke Parent self report
exposure Harms/dangers of smoke True/false
exposure
72


Measures
Several measures of smoking behaviors in households were included in the
survey, all of which have been documented in previous research. These included the
number of smokers living in the household (Borland, 1999), frequency of passive
smoke exposure in household as measured by number of cigarettes smoked in
household in past week, (Al-Delaimy et al., 1999), indoor locations of smoke
exposure by rooms (Goldstein, 1994), smoking status of respondent as a current
smoker (smoked within the past thirty days), ex-smoker (smoked, but not in past
thirty days) and never smoker. Smoke exposure in the household is measured by the
last time that someone smoked in the home from within the past 24 hours to more
than a month ago (Crone et al., 2001). Smoking status of the parent and the number
of children over age 10 who smoked in the home were also measured.
To insure that cotinine measurements of urine were accurate and reflective of
smoke exposure based on half-life of cotinine in young children, one month was
selected as the cut-off point for current exposure. Past smoke exposure measured
smoke exposure between one month and one year ago. No smoke exposure
represented smoke exposure of more than one year ago. Past smoke exposure and
no smoke exposure were reduced to one category of no smoke exposure for
multivariate analysis.
73


The number of cigarettes smoked in the house in the past week assessed a
second measure of smoke exposure. The categories were reduced to none, two packs
or less per day, and more than two packs per day for summarization and analysis.
Smoking rules in household, as identified in the National Health Information
Survey (1994), comprised the outcome indicator of smoking behaviors in the
household and included three categories: no smoking allowed anywhere or anytime,
smoking sometimes or in some places in the household, and smoking allowed without
restriction anywhere and anytime in the household. These data were reduced into two
categories for data analysis: complete smoking ban defined as no smoking occurs in
the household, and no smoking ban defined as smoking is allowed in the home with
or without restrictions. This categorization of smoking in the household for
purposes of analysis is consistent with other studies (Ashley et al., 1998; Crone et al.,
2001; Norman et al., 1999; Okah et al., 2002). The strength of the relationship
between the smoking ban variable and cotinine level of urine sample was also
investigated as a validation measure.
Five questions comprised a measure of situations in households when
smoking was allowed, including presence of children and permission for relatives and
friends to smoke in home.
74


Table 3.2: Survey Questions For Situations in Households
When Smoking Is Allowed
Smoking is allowed indoors when children are at home. Yes or No
Smoking is allowed indoors when children are in the room. Yes or No
Smoking is allowed indoors when children are asleep. Yes or No
Some relatives are allowed to smoke in my home. Yes or No
Some friends are allowed to smoke in my home. Yes or No
Higher scores indicated more situations in the home in which smoking was
allowed, reflecting a higher level of smoke exposure. The five questions were
evaluated for reliability, resulting in a Cronbachs alpha of .8948. Corrected item-to-
total correlation of individual items was .6988 or greater. Both of these measures
suggest that the five questions provide an acceptable measure of household situations
resulting in smoke exposure of children.
Smoking practices of friends were investigated through two items: if friends
were allowed to smoke in home and if friends allowed smoking in their own homes.
Several previous studies (Goldstein, 1994; Erikson & Bruusgard, 1995) emphasized
the importance of effects of friends on smoking practices in households. Although
this did not arise as a major theme in the qualitative portion of the study, it was
investigated as a part of situations in households in which smoking is allowed.
75


Knowledge of Smoke Exposure. Inputs to smoking behaviors in HHPH were
measured by knowledge and attitudes/beliefs reported by households. Knowledge of
ETS exposure was measured using a twelve-item true/false tool. Two concepts of
knowledge were measuredknowledge of health effects of ETS exposure (Brownson
et al., 1992) and knowledge of harms/dangers of smoke exposure (Feamow, Chassin,
and Presson, 1998; Kegler and Malcoe, 2002). The items identifying health effects
were derived from analysis of the semi-structured interviews in which illnesses,
commonly associated with smoke exposure, were identified by parents.
Table 3.3: Knowledge Assessment
CODE ITEM CORRECT ANSWER CATEGORY
HARMTF If a child is healthy, tobacco smoke will not harm them. False Harms
SCIENTF The scientific evidence doesnt really prove tobacco smoke is harmful. False Harms
ASTHTF Indoor tobacco smoke makes childrens asthma worse. True Health Effects
COLDSTF Children who are often around indoor tobacco smoke have more colds and coughs. True Health Effects
CANCERTF People who are exposed to tobacco smoke when they are children are more likely to get cancer as adults. True Health Effects
EARTF Children who are often around indoor tobacco smoke have more earaches. True Health Effects
SICKTF Smoke exposure hurts children only if they are already sick. False Health Effects
76


Table 3.3: Knowledge Assessment fCont.l
CODE ITEM CORRECT ANSWER CATEGORY
CHRONTF Smoke exposure hurts children only if they have a chronic illness like asthma. False Health Effects
BABYTF A little smoke exposure will not harm a baby. False Health Effects
CHILDTF A little smoke exposure will not harm a child. False Health Effects
ETSTF Smoke exposure harms children whether they are sick or healthy. True Harms
DANGER The dangers of tobacco smoke have been exaggerated False Harms
Attitudes/ Beliefs. A fourteen-item scale assessed attitude/beliefs of
households using a Likert scale composed of five points. Reponses ranged from
strongly agree to strongly disagree and were categorized as to whether they
represented negative attitudes, neutral or positive attitudes toward smoke exposure.
Higher scores indicated more negative attitudes toward smoke exposure. Four
measures of attitudes/beliefs were assessed:
1. Health Protection of Children against ETS exposure. Participants
completed five items assessing attitudes and responsibilities in protecting children
from ETS (Feamow, Chassin and Presson, 1998).
2. Parent beliefs about smoke exposure related to selves or children. Three
77


items assessed parent response to general statements indicating attitudes toward
smoke exposure.
3. Beliefs about effects of smoke exposure on children in household. (Ashley
et al., 1998; Al-Delaimy et al., 1999). Three items assessed parent responses to
smoke exposure of children in households.
4. Beliefs about legality protection of children from smoke exposure in
households. Two items assessed parent beliefs about legal restrictions on household
smoke exposure.
Table 3.4: Attitude/Beliefs Assessment
Code Item Category
HARMFUL I believe that smoke exposure is harmful to children. Parent beliefs about smoke exposure of children.
HATE I hate it when I see adults smoking around children Parent beliefs about smoke exposure of children.
MAD It makes me mad when people smoke indoors around children. Parent beliefs about smoke exposure of children.
DEAL Its not a big deal if adults smoke around children. Parent beliefs about smoke exposure
OKAY Its okay for people to smoke around children as long as they dont smoke around my kids. Parent beliefs about smoke exposure
NOPROB I dont mind when people smoke around me. Parent beliefs about smoke exposure
NEVER Children should never be exposed to environmental tobacco smoke. Health protection of children against smoke exposure
78


Table 3.4: Attitudes/Beliefs Assessment (ConO
Code Item Category
UPSET I get mad or upset when I see someone smoking close to a baby. Parent beliefs about smoke exposure of children.
PROTECT A parent should protect their child from smoke exposure. Health protection of children against smoke exposure
RIGHTS Parents have the right to decide whether or not they will smoke around their children. Beliefs about laws protecting children from smoke exposure.
LAWS We should have laws against smoke exposure in home just like in public places and work places. Beliefs about laws protecting children from smoke exposure.
ILLEGAL Smoking indoors should be illegal where children live. Beliefs about laws protecting children from smoke exposure.
TEACH It is my job to teach my children about dangers and health effects of tobacco smoke. Health protection of children against smoke exposure
Community or macro-level factors were assessed by demographic
characteristics of households including age/gender/marital status of respondent,
number of adults/children living in household, educational level of mother and father,
household income, type of housing, number of rooms in home, and ethnicity/race of
household.
Health of children living in the household was assessed by three measures:
1. Number of minor acute illnesses in past year of each child living in
79


household.
2. Presence of chronic illnesses in children in household.
3. Presence of children with asthma in the household.
4. Perceived health of children in household (healthy or unhealthy).
The number of minor acute illnesses in the household in the past year was
measured as a continuous variable and the other three were measured as categorical
variables. Households were categorized as to whether they contained children with
chronic illnesses or children with asthma. Households who identified all children as
being healthy were categorized as healthy and those who identified at least one child
as not healthy were categorized as not healthy. Extensive findings in the literature
demonstrate relationships between childrens health and smoke exposure (Cook &
Strachan, 1999; DiFranzia & Lew, 1996; Etzel, 1997; Gergen et al, 1998; Mannino et
al., 2001; Pershagen, 1999).
Exceptions to smoking rules were investigated by asking parents whether
they made exceptions for smoking in the home. Weather-related exceptions, relatives
visiting, discomfort in telling someone not to smoke in house were identified in semi-
structured interviews and included as items in the survey.
Reliability and Validity of the Instrument
Content Validity. Questions adopted from the NHIS survey included smoking
categories, number of cigarettes smoked in the home and demographics. The
80


remaining survey questions (in English) were reviewed for content and clarity by two
PhD level researchers, two masters level nurse practitioners, and two patient service
coordinators at clinic sites, and revised based on recommendations made by
reviewers. The Spanish version of the survey was reviewed by one bilingual PhD
researcher and two bilingual patient service coordinators who had Spanish as a first
language and were employed in school-based clinics. Changes in grammar and
sentence structure to make the survey more understandable were done based on their
recommendations.
The surveys were pilot-tested by twenty families in English and ten families in
Spanish. Interviews were conducted with families who completed the surveys,
problem questions were identified and changes were made to clarify items. The
English and Spanish versions of the survey were approved by human subjects
committee (COMIRB) before they were administered.
Reliability Analysis. Reliability analysis was used to determine the ability of
the true/false items to measure the constructs they were intended to measure. A Kuder
Richardson 20 (KR 20) was used to determine the alpha coefficient because of the
dichotomous nature of the true/false items. Reliability analysis revealed a Kuder-
Richardson 20 (KR 20) measure of .6003. Of the two measures tested, the corrected
item-to-total correlation suggested that harms/danger questions were better measures
of knowledge than the health effects questions.
81


Table 3.5: Reliability Analysis of Knowledge Scale
Mean of Correct Answers Corrected Item to Total Correlation KR20
Health Effects (6 items) 5.39 (90%) 6 of 6 items Harms/Dangers (6 items) 5.34 (89%) 3 of 6 items <.4 .5746
Total Knowledge (12 items) 10.73 (89%) 8 of 12 items <.4 .6003
Harms/Dangers (5 items) 4.42 (88%) 2 of 5 items <.4 .6315
Health Effects (2 items) 1.82 (91%) 0 of 2 items <.4 .6618
Revised Scale (7 items) 6.24 (89%) 2 of 7 items <.4 .7174
Removing questions with low corrected item-to-total correlation (less than
.4000) in the original reliability analysis improved the KR 20 (See Table 3.6).
Removing four items measuring health effects and one item measuring harm/dangers
resulted in improvement of KR 20 (alpha coefficient) to .7174. All questions were
kept as a measure of overall knowledge and utilized in later analysis. Reducing the
number of questions increased reliability of the scale but reduced correlations with
dependent variables. Consequently, all items were kept for analysis in which level of
knowledge was a predictor of health indicators related to smoke exposure.
82


Factor Analysis of Attitude Measure. Factor analysis was performed on the
14 items measuring attitude/beliefs. The total number of subjects provided a ratio of
10-15 subjects for each attitude/belief variable measured. A Kaiser-Meyer-Olkin
measure of sampling adequacy was .815, making factor analysis an appropriate
measure of construct validity for this scale. The 14 items were contained in 4
theoretical constructs (See Table 3.7). The reliability analysis of the 14-item scale
was .7906 (Cronbachs alpha). The question pertaining to the rights of parents
regarding smoking (RIGHTS) had a low item-to-total correlation, and when the item
was removed, the 13-item scale had Cronbachs alpha of .8012. Although the
difference was small, the item also had a low reliability coefficient and was poorly
written, so it was removed and the thirteen-item scale was used for further analysis.
Table 3.6: Factor Analysis: Rotated Component Matrix
ITEM FACTOR 1 FACTOR 2 FACTOR 3 FACTOR4 COMMUNALITIES
JOB .755 .655
TEACH .704 .518
PROTECT .698 .541
HARMFUL .676 .480
HATE .619 .677
ILLEGAL .936 .894
LAWS .914 .881
MAD .552 .622
NEVER .685 .637
UPSET .902 .820
NOPROB .615 .631
OKAY .882 .676
DEAL .761 .653
83


Four components were extracted from the 13-item scale with Eigen values of greater
than 1, accounting for 67% of cumulative variance in 5 iterations. These
components were consistent with the four theoretical constructs upon which the
original scale was developed although several items factored differently into groups
as a result of the analysis (See Table 3.6). Four groupings of variables remained and
these groupings were maintained for further analysis.
Table 3.7: Comparison of Item Grouping in Factor Analysis
Theoretical Constructs Item In Original Item in Group After
Group Factor Analysis
(14 items) (13 items)
Beliefs of parents regarding health protection of children. NEVER* PROTECT JOB TEACH PROTECT* JOB TEACH HARMFUL HATE
Attitudes of parents regarding smoke exposure. DEAL OKAY NOPROBLEMS DEAL OKAY NOPROBLEMS
Attitudes of parents regarding smoke exposure of their children. HARMFUL HATE MAD UPSET NEVER UPSET MAD
Attitudes of parents regarding use of legal means to protect children from smoke exposure. LAWS RIGHTS ILLEGAL LAWS ILLEGAL
*See Table 3.4 (page 78,79) for explanation of codes.
The final step involved computing reliability analysis of the 13-item scale
using the groups of variables, which emerged from the factor analysis. The four
84


Full Text

PAGE 1

... .;_*': HOUSEHOLD FACTORS ASSOCIATED WITH ENVIRONMENTAL TOBACCO SMOKE EXPOSURE OF YOUNG CIDLDREN by Yvonne K. Yousey B.S.N., Eastern Mennonite University, 1970 M.S., University of Colorado Health Sciences Center, 1977 A thesis submitted to the University of Colorado at Denver in partial fulfillment of the requirements for the degree of Doctor of Philosophy Health and Behavioral Sciences 2003 I (

PAGE 2

This thesis for the Doctor of Philosophy degree by Yvonne K. Y ousey has been approved by Lauren Clark I Date I

PAGE 3

Yousey, Yvonne K., (Ph.D., Health and Behavioral Sciences) Household Factors Associated With Environmental Tobacco Smoke Exposure Of Young Children Thesis directed by Professor Kitty Corbett ABSTRACT Reducing secondhand smoke exposure is one of the four priorities for global tobacco prevention and control identified by the World Health Organiziation. Young children are especially vulnerable to health effects related to tobacco smoke exposure, most often exposed as a result of parental smoking practices in their home. Little information is available about the behavioral factors which influence secondhand tobacco smoke exposure in households where children reside. This descriptive study examined household characteristics associated with smoking policies in households using a household production of health framework within a social ecology perspective. Several hypotheses examined characteristics associated with smoking policies in households. They included: 1) no differences in socio-demographic characteristics exist between household with complete smoking bans and those without, 2) children in households with complete smoking bans have less reported health effects related to smoke exposure than children in households with no bans, 3) households with smoking bans have greater knowledge of environmental tobacco smoke exposure, report more negative attitudes toward environmental tobacco smoke exposure, and have less smoke exposure as measured by cotinine levels than households with no smoking bans. A sequential mixed-methods approach was used. Semi-structured interviews were conducted with 20 households, some with smoke exposure and some without. These interviews were used to increase content relevance and construct validity of the survey developed for use in the quantitative phase. Surveys were administered to a cross sectional sample of English and Spanish-speaking subjects, 18 years and older, with children, age newborn through pre-school age residing in the household. Socio demographic factors, knowledge and attitudes of parents, health status of children in the household, smoking behaviors and smoking bans in households were investigated. Reports of bans were validated through measurement of cotinine levels of urine samples from children, newborn to five years in the household. Two hundred twentyfive households completed surveys; 203 households provided a urine sample from an age-appropriate child that was tested for cotinine. Complete smoking bans were reported by 164 (73%) households, and no or partial 111

PAGE 4

bans were reported by 61 households (27%); 97 children(48%) tested negative for smoke exposure through cotinine levels and 107 (52%) tested positive, indicating smoke exposure. Ethnicity ofhouseholds when indicated by language in which the survey was completed, and attitudes toward smoke exposure were household characteristics significantly associated with the presence of smoking bans. No smoking bans correlated positively with cotinine measures of smoke exposure (r=.486). Smoke exposure in households was significantly better explained by the inclusion of smoking bans in a model of smoking behaviors than by other measures. This abstract accurately represents the content of the candidate's thesis. I recommend its publication. Signed IV

PAGE 5

DEDICATION 1bis dissertation is dedicated to my husband, Jim Dunn, and to our children, Collin, Katie, and Kelsie with love and my deepest gratitude to all of you.

PAGE 6

ACKNOWLEDGEMENTS My sincere thanks go to the many people who made this dissertation possible. I would like to thank the members of my committee, Kitty Corbett, Debbi Main, Lauren Clark, Steve Koester, and Michael Zinser. Kitty, who supported me from the beginning, and made the dissertation possible; Debbi Main who provided assistance from afar and always answered my emails; Lauren Clark who always had a positive comment and challenged me to strive forward; Steve Koester and Michael Zinser who encouraged and challenged me as I completed the process. I also thank Carol Vojir for her patience and never ending assistance in completing the statistical portions of this work. And finally, thanks to Chris Pon and Norman Chandler for managing my grant and keeping me abreast of the next deadline. I also wish to thank the many people and organizations that participated in this study: the staff and administration at Sanville Preschool, Adams County Head Start, Adams District 50 Preschool, Sheridan Preschool/Head Start and the staff at the School-Based Health Centers for their assistance in recruitment of subjects and in data collection. Without their enthusiasm and loyal support, this project would not have been possible. I thank the staff at Community Health Services of The Children's Hospital, and my colleagues from the School of Nursing at the University of Colorado Health Sciences Center who provided support and encouragement along the way. Special thanks go to Bonnie Gance-Cleveland, Lynn Gilbert, Geneva Jarvis, Andra Opalinski, and Julie Degenstein for their assistance in data collection and support of this project. Many, many loving thanks go to my family. To my husband, Jim, who walked by my side with unfaltering love, support, and patience, and took on countless extra responsibilities for our family through the years so that I could complete this project. I thank my children, Collin, Katie, and Kelsie, who always had an encouraging word and understood when I couldn't be there all of the time. I thank the members of my class, Cohort 4, for all of your encouragement and timely advice. Finally, to my parents, family and friends from near and far who didn't always understand what I was doing, but provided support and interest anyway. Thanks to all of you for being there through the years and helping me to reach this goal. This dissertation was supported by grant lD-068 from The Colorado Tobacco Research Program, Boulder, Colorado.

PAGE 7

CONTENTS Figures .. ............................................................................ xii Tables ............................................................................... xiv CHAPTER 1. IN"TRODUCTION ................................................................... 1 Overview ............................................................................................... I Hypothesis and Specific Aims .............................................. .4 Overview of Research Methods .............................................. 6 Setting of the Study ............................................................ 9 Description of the Study Population ............................... .......... 11 Structure of Dissertation ..................................................... 12 2. BACKGROUND, THEORY ANDLITERATUREREVIEW .............. l4 Background .................................................................... 14 Theoretical Frrunework ....................................................... 18 Social Ecology ................................................ .. .... .18 Household Production ofHealth .................. ...... ............. ..... 19 Households and Smoking ....................................................... .21 The Household as a Unit of Analysis ........................ ..... 23 Environmental Tobacco Smoke Exposure .................. ................. 25 vii

PAGE 8

Cotinine and Smoke Exposure ..................................... 25 Definition of Terms .. ................................................. 28 Health Behaviors, Indicators, and Outcomes .......................... 28 Smoking Policies in Households .................................. 29 Smoking Bans ........................................................ 30 Review of Literature ......................................................... 34 Epidemiology of Smoke Exposure ................................ 35 Passive Smoke Exposure in Homes: Intra-household Factors ................................................................. 37 Smoking Policies in Households ................................. 40 Demographic Characteristics ...................................... .41 Health Care Utilization and Economic Costs .................... 43 Knowledge of Health Effects of Smoking ........................ 45 Household Behaviors, Attitudes, and Smoke Exposure ........ 46 Social Support and Smoking ....................................... 49 Self-efficacy ......................................................... 50 Interventions for Smoke Exposure ................................. 51 Summary .............................................................. 52 3. METHODOLOGY ............................................................... 54 Overview ...................................................................... 54 Qualitative Methods ......................................................... 56 Vlll

PAGE 9

Issues ofValidity .................................................... 58 Interview Instrument ................................................. 59 Sample Selection for Interviews ................................... 61 Structure of Interview ............................................... 62 Role of the Researcher .............................................. 63 Data Collection ...................................................... 65 Quantitative Methods: Survey Research ................................... 66 Instrument Development ............................................ 67 Measures .............................................................. 73 Knowledge of Smoke Exposure .......................... 76 Attitudes/Beliefs ............................................. 77 Reliability and Validity oflnstrument ............................ 80 Content Validity ............................................. 80 Reliability Analysis ......................................... 81 Factor Analysis of Attitude Measure ...................... 83 Study Population and Setting ....................................... 86 Participant Recruitment ............................................. 87 Data Collection Procedures ......................................... 88 Subject Payment ...................................................... 93 Cotinine Testing ....................................................... 94 Human Subjects Review ...................................................... 97 lX

PAGE 10

Methods for Qualitative Analysis .......................................... 97 Methods for Quantitative Analysis ......................................... 99 4. FINDINGS FROM QUALITATIVE DATA ANALYSIS ................. 105 Description of the Sample ................................................. 106 Observation of Smoking Behaviors .................................... ... 1 08 Results of Qualitative Analysis ............................................ 1 09 Description ofThemes ...................................................... 110 Health Promotion and Health Protection ........................ 113 Implementation of Smoking Rules ............................. 115 Knowledge of Effects of Smoking .............................. 120 Attitudes and Beliefs ............................................... 122 Community Factors ................................................. 125 Conclusions .. ................................................................. 128 5. FINDINGS FROM QUANTITATIVE DATA ANALYSIS ............ 130 Overview ..................................................................... 130 Demographic Characteristics .. ............................................ l32 Ethnicity ............................................................. 135 Demographics by Recruitment Site ........................... ... 140 Household Characteristics and Smoking Bans .......................... 144 Smoking Behaviors and Ethnicity ............................... 152 Discussion of Smoking Measures ............................... 155 X

PAGE 11

Health of Children in the Household ............................. 155 Knowledge and Attitudes/Beliefs and Smoking Bans ......... 158 Smoking Bans and Cotinine Exposure .......................... 163 Ethnicity and Cotinine Measurements .................................... 166 Smoke Exposure in Daycare ............................................... 166 Mulitvariate Analysis-Logistic Regression .............................. 167 Selection of Variables for Logistic Regression Model. ....... 168 Factors Associated With Smoking Bans ........................ 173 Hypothesis Testing ......................................................... 179 6. DISCUSSION AND CONCLUSIONS ...................................... 182 Overview ..................................................................... 182 Discussion of Findings: Hypotheses ..................................... 182 Findings Compared to Other Studies ...................................... 187 Strengths and Limitations of the Study .................................... 196 Future Research Questions ................................................ 203 Conclusions ................................................................... 207 APPENDIX A. CONSENT FORM AND INSTRUMENT FOR SEMI-STRUCTURED INTERVIEWS ................................... 210 B. CONSENT FORMS FOR SURVEY QUESTIONNAIRE SURVEY QUESTIONNAIRE ............................................ 217 C. RECRUITMENT FLYERS FOR QUANTITATIVE PHASE ........ 248 XI

PAGE 12

BIBLIOGRAPHY ............................................................................. 253 xii

PAGE 13

FIGURES Figure 1.1 Household Factors and Tobacco Smoke Exposure ................................. 8 2.1 Social Ecology, Household Production of Health and Environmental Tobacco Smoke Exposure in Households ............................................ 32 Xlll

PAGE 14

TABLES Table 1.1 Study Sites .............................................................................. 10 1.2 Ethnicity of Study Site Population ................................................... 12 3.1 Predictor and Outcome Variables in Survey ........................................ 71 3.2 Survey Questions For Situations in Households When Smoking Is Allowed .................................................................. 75 3.3 Knowledge Assessment ............................................................... 76 3.4 Attitude/Beliefs Assessment ............................ ... ........................... 78 3.5 Reliability Analysis ofKnowledge Scale ........................................... 82 3.6 Factor Analysis: Rotated Component Matrix ...................................... 83 3.7 Comparison ofltem Grouping in Factor Analysis ............................. ..... 84 3.8 Reliability Analysis of Attitude/Belief Scale (13 Items) .......................... 85 3.9 Ethnicity/Racial Characteristics ofPopulation ................................ .... 87 3.10 Independent Categorical Variables ................................................ 101 4.1 Interview Characteristics ofHouseholds by Smoking Status .................. 107 4.2 Interview Characteristics of Households by Ethnicity ................... ......... 108 4.3 Summary of Themes and Codes Identified in Household Interviews on Smoking ................................................. 112 xiv

PAGE 15

4.4 Health Promotion Activities and Health Habits ofFamilies ................... 114 4.5 Reasons For and Against Smoking in the Household ........................... 118 4.6 Knowledge ofEffects ofSmoking ................................................ 120 5.1 Characteristics of the Sample by Household ...................................... 134 5.2 Ethnicity of Subjects .................................................................. 136 5.3 Demographic Characteristics and Ethnicity ....................................... 138 5.4 Sites From Which Subjects Recruited ............................................. .140 5.5 Demographic Characteristics ofPreschool!Head Start Sites and SBHCs ...... 142 5.6 Demographic Characteristics and Household Smoking Bans ................... 144 5. 7 Smoking Behaviors and Smoking Bans in Households ......................... 149 5.8 Correlation of Smoking Measures With Presence of Smoking Bans ........... 151 5.9 Smoking Behaviors and Ethnicity .................................................. 153 5.10 Childhood illness and Smoking Bans in Households ............................ 156 5.11 Correlations oflllness and Smoking Bans in Households ....................... 157 5.12 Attitudes and Beliefs About Smoke Exposure .................................... 159 5.13 Constructs of Attitudes and Beliefs ................................................ 162 5.14 Cotinine Measurement of Passive Smoke Exposure ............................ 164 5.15 Correlations ofReported Smoking Measures and Cotinine Measurements .. 165 5.16 Smoke Exposure in Daycare ........................................................ 167 5.17 Categorical Variables For Regression Models .................................... 169 XV

PAGE 16

5.18 Logistic Regression: Demographic Variables and Smoking Bans ............ 172 5.19 All Variables and Smoking Bans in Households ................................. 175 5.20 Smoking Behaviors and Cotinine Levels .......................................... 178 xvi

PAGE 17

CHAPTER 1 INTRODUCTION Overview Environmental tobacco smoke (ETS) exposure is an important source of morbidity and mortality, being the third leading preventable cause of death in this country (Rudzinski and Sirois, 1994). Unlike most other risk behaviors, tobacco smoking is a health risk for those exposed to ETS who do not smoke directly themselves (Sockrider, 1996). Children are more likely than adults to have adverse health effects from ETS exposure (Ashley and Ferrence, 1998); they suffer from lower respiratory illness, chronic middle ear effusion, pulmonary fimction changes, asthma exacerbations, sudden infant death syndrome, and lung cancer (Cook and Strachan, 1999; Etzel, 1997; Ey et al, 1995; Li, Peat, Xuan, and Berry, 1999; Samet, 1999). Between 8. 7 and 12.4 million American children less than five years of age are exposed to cigarette smoke in their homes (Etzel, 1997), resulting in substantial public health and economic impacts (Ashley and Ferrence, 1998). The household is a primary source of ETS exposure for young children. Little is known about how families in households deal with ETS exposure in spite of the fact that deleterious health effects are well established. Societal, economic, legal and political factors contribute to lower levels of ETS control measures in homes 1

PAGE 18

compared with workplaces and public places (Ashley and F errence, 1998) where policies which restrict and/or prohibit smoking have effectively reduced ETS exposure for non-smokers. There is a growing awareness of the dangers of ETS exposure in households (Borland, Mullins, Trotter, and White, 1999; Farkas, Gilpin, Distefan, and Pierce, 1999; Pizacani, Martin, Stark, Koepsell, Thompson, and Diehr, 2003), but little is known about the factors leading to ETS exposure and its reduction in that environment. The involuntary nature ofETS exposure in children places the burden of protection on those who function in protective capacities, especially in homes. Young children, vulnerable because of physiological and developmental differences, cannot remove themselves from exposure and are dependent on other measures for protection (Ashley and Ferrence, 1998). The examination of events and conditions surrounding household policies related to tobacco use and smoke exposure will increase understanding of how family members make decisions about smoking behaviors and negotiate smoke free environments for their children. When adults act to restrict smoking in their homes, they not only reduce morbidity in their young children, but convey an important message about the dangers of tobacco smoke which may have long term implications in smoking initiation in these children (Biener, Cullen, Zhu, and Hammond, 1997). Reduction of exposure at the household level is necessary for the health and protection of children living there. 2

PAGE 19

Protecting children from the harm of smoke exposure is not only a responsibility of the families with whom children live, but of the community in which they reside, and the health professionals with whom they come in contact. This research resulted from the interest of the researcher in addressing smoke exposure of young children in families of diverse ethnicity, low income, and low educational levels. A greater understanding of the dynamics of the household was needed before specific measures for reducing smoke exposure could be implemented. The gap between knowledge of harmful effects and actual controls in the home indicates a need to examine situational factors that lead to smoke exposure (Goldstein, 1994). In this study, the researcher was in the unique position to intervene with families in maintaining smoke free environments through her role as the primary health care provider of their children. This provided her the opportunity to investigate how households address and deal with smoke exposure and what methods are utilized by family members to reduce exposure. From these, recommendations, strategies, and interventions for reducing exposure can be developed. Examining and describing household health producing and health maintaining behaviors related to environmental tobacco smoke exposure is a necessary prerequisite to analyzing their health effects. 3

PAGE 20

Hypothesis and Specific Aims The purpose of this study was to investigate factors within households that impact smoking practices and environmental tobacco smoke exposure of young children living there. The household, a central unit for examining events and conditions internally and externally, provides a basis for disease prevention and health maintenance interventions (Berman, Kendall, and Bhattacharya, 1994 ). Understanding how household decisions are made regarding potential health risks/problems such as environmental tobacco smoke exposure is the first step in protecting children's health. How adults and family members in households make decisions about smoking behaviors and negotiate smoke free environments for their children provides the basis for development of effective interventions at the individual and/or community level that lead to enhancement of tobacco-free family environments. This research pursued the following aims: 1 Explore household characteristics and relationships between factors associated with reported smoking policies in households. 2. Identify means of implementation and enforcement of smoking policies in households where young children reside. 3. Explore relationships between variables associated with household smoking policies and actual smoke exposure of children as measured by cotinine levels. 4

PAGE 21

Using a social ecology perspective, it is hypothesized that household characteristics will impact smoking policies as evidenced through smoking bans in households. These household characteristics are impacted by factors both within the household and the community in which it is located. It was predicted that cotinine measurements of smoke exposure of young children residing in households with no/partial home smoking bans would be greater than in those households with complete home smoking bans. It was also predicted that children in households with complete smoking bans would have fewer negative health effects than children in households with no/partial bans. Five hypotheses were tested: 1. There are no differences in socio-demographic characteristics between households with complete smoking bans and those without. 2. Children in households with complete smoking bans will have fewer reported health effects related to smoke exposure than children in households with no/partial smoking bans. 3. Households with complete home smoking bans will have greater knowledge of environmental tobacco smoke exposure than households with no/partial home smoking bans. 4. Households with complete smoking bans will report greater negative attitudes toward environmental tobacco smoke exposure than households with no/partial home smoking bans. 5. Households reporting complete home smoking bans will have less 5

PAGE 22

smoke exposure as measured by cotinine levels than households reporting no/partial home smoking bans. Overview of Research Methods In this study, factors impacting smoke exposure in households, smoking policies (smoking bans) as reported by family members, and the relationships between specified variables were investigated using a household production ofhealth framework. To achieve the aims of the study, research employed both qualitative and quantitative methods; results were validated by cotinine measurement of urine samples from children in the household. Semi-structured interviews were conducted with parents in households containing at least one child, newborn to age 5. The purpose of these interviews was to provide information on the social and cultural context and personal meaning of smoking (Erickson and Kaplan, 2000) necessary for the identification and development of factors measured in the survey. The interviews explored health behaviors and how they are implemented in households, smoking policies and practices in the home and car, methods by which smoking policies are implemented and enforced, and factors that families identified as being important in this process. The survey questionnaire was developed to investigate the hypotheses exploring household factors and health effects with smoking policies. The content of questions included: household characteristics associated with smoking policies, 6

PAGE 23

actual household smoking policies (rules), the implementation and enforcement of restrictions in households, health indicators of children in the household, knowledge of effects of tobacco smoke exposure, and attitudes/beliefs about ETS exposure of children. The characteristics identified through the semi-structured interviews were confirmed by literature findings, developed as variables, and measured in the survey. The second hypothesis was tested by measuring actual smoke exposure of urine obtained from a child, age newborn to school age, in a participating household. Measurement of cotinine, a biomarker of nicotine, provided a quantitative measure whereby actual smoke exposure of young children could be measured, validating reported smoking policies and smoking behaviors in households. Identification of the main sources of exposure and quantification of the dose of inhaled smoke are fundamental to the study of passive smoking effects in children (Jarvis, 1999). Surveys provide cumulative information about ETS exposure, but there is a danger of misclassification through underreporting. The reliability of survey assessments is enhanced when combined with biochemical verification such as cotinine analysis (Wewers and Uno, 2002). Together they provide the best assessment of the extent of ETS (Jarvis, 1999; Scherer, Meger-Kossien, Riedel, Renner, and Meger, 1999), more accurate than the use of either of them alone. Only households with children less than school-age were invited to participate because the level of cotinine in body fluids of children in this age groups are a more valid measurement of actual smoke exposure. Younger children are more confined to 7

PAGE 24

their homes in which smoke exposure may occur and do not have the ability to remove themselves from it (Ashley & Ferrence, 1998; Hopper & Kelly, 2000). They also have less opportunity for smoke exposure outside the home environment by other than their parents. The following diagram illustrates the hypothesized relationships in the study: Figure 1.1: Household Factors and Tobacco Smoke Exposure Qualitative Analysis Quantitative Analysis What household characteristics are What household characteristics are associated with smoking policies in associated with smoking bans in households? ____. households? How are smoking policies implemented in households? How do smoking behaviors and smoking bans in households affect cotinine levels in pre-school children? Are household smoking bans an indicator of health in households in which young children reside? Setting of the Study / The population selected for this study includes families with children, 5 years or younger, who reside in Adams, Arapahoe, and Jefferson counties in metropolitan Denver, Colorado. Specific areas within these counties were targeted because of their accessibility to the researcher and their socioeconomic status as measured by income 8

PAGE 25

levels of young families residing there. Because tobacco smoking is associated with lower socioeconomic status (Arborelius et al, 2000) areas with known lower household income were selected. These areas also included populations with ethnic diversity. Four faculty practice sites associated with the University of Colorado Health Sciences Center, School ofNursing, located in Westminster (2), Sheridan (1) and Arvada (I), provide primary health care services to families with children, newborn to eighteen years of age in metropolitan Denver, Colorado. Two of these faculty practice sites (Westminster and Sheridan) are located in areas with a federally qualified underserved health status designation. Services are targeted for families who are Medicaid eligible, CHP+ eligible or who have no health insurance. The school-based health centers in Westminster and Sheridan are jointly administered with staffing provided by Community Health Programs of The Children's Hospital and University of Colorado Health Sciences Center, School of Nursing in Denver, Colorado. Four Head Start sites in Adams County and tWo preschools associated with District 50 and District 14 providing preschool and early childhood services to three and four year old children in Adams County were utilized. One Head Start/preschool in Sheridan School District (Arapahoe County) also participated. 9

PAGE 26

Table 1.1: Study Sites School-based County Pre-schooli.Head Start County Health Centers Sties (SBHCs) Sheridan Arapahoe Sheridan Arapahoe Preschooi!Head Start Westminster (2 Adams Westminster District Adams sites) #50 Pre-School Carin'Clinic Jefferson Adams County Head Adams (Arvada) Start (4 sites) Sanville Preschool Adams The school districts in these locations include families of lower income with 40-60% of families eligible for free/reduced lunch. Many of these families who are Medicaid eligible, CHP+ eligible, or uninsured utilize the school-based health clinics for services. Families whose children attend Head Start met financial criteria for enrollment. The majority of families utilizing services at preschools met financial eligibility requirements necessary to access services. The qualitative portion of the study was conducted using the school-based health center sites. Families from all sites were recruited for participation in the survey and measurement of cotinine levels. 10

PAGE 27

Description of the Study Population The population selected for the study consisted of low-income families with young children, age five and under, who reside in School Districts 14, 50, and 27J of Adams County, Sheridan School District in Arapahoe County, and in Arvada of Jefferson County. Families with children enrolled in four preschool/Head Start facilities and four school-based health centers were eligible as study participants. These sites were selected because of their geographic location, and the willingness of administrators to facilitate participation of families. Two other preschools in Adams County were invited to participate but declined. The school-based health centers were selected because they have families with age appropriate children, are ethnically diverse, and of low income. The researcher had access to these populations because she provided primary health care services in these sites. The Head Start sites and preschools were located in the same school districts as the three school-based health centers. Head Start sites in two other school districts with low income and ethnically diverse populations in Adams County also participated. One school-based health center is located in Jefferson County, providing services to uninsured families who attend three elementary schools and on middle school of lower income in Arvadada Approximately I 021 children were enrolled in preschools/Head Start facilities and 3717 children were enrolled in school-11

PAGE 28

based health centers. Ethnicity/racial characteristics of the study population based on emollment in preschools or school-based clinics are as follows : Table 1.2: Ethnicitv of Study Po:Qulation Total SBHCs Total School Preschools SBHCs And Preschools Districts N % N % N % N % Not-Hispanic, 1459 39 406 40 1865 39 11082 42 White Hispanic 1843 50 515 50 2358 50 13114 50 Black 72 2 30 3 102 2 575 2 Asian 120 3 40 4 160 3 1281 5 Other 223 6 30 3 253 5 383 1 Total 3717 1021 4738 26435 Structure of Dissertation Chapter 1 provides the overview of the need for research in ETS exposure in homes, describes the purpose of the research, states the hypotheses and provides an overview of the research methods, description of the study setting, and the study population. Chapter 2 provides the background for the research including theoretical framework, definition of terms and variables, and a review of the literature. Chapter 3 includes a description of the design with discussion of the qualitative and quantitative methodologies: semi-structured interviews, survey questionnaire, and 12

PAGE 29

testing of urine for cotinine to validate results from survey questionnaire. Chapter 4 discusses qualitative findings and use of these data in development of the survey. Chapter 5 includes quantitative findings, with descriptive summaries, multivariate analysis and conclusions. Chapter 6 further describes conclusions, study limitations and provides recommendations for future research. 13

PAGE 30

CHAPTER2 BACKGROUND, THEORY AND LITERATIJRE REVIEW Background Much has been written about disease risks related to environmental tobacco smoke exposure, and a growing awareness of this hazard has resulted in efforts to reduce exposure in public places (Pizacani et al, 2003). The primary method of protecting nonsmokers has been to restrict smoking through bans and non-smoking policies; these have been shown to effectively limit exposure in public places and thereby reduce health risks of exposure in these locations. Laws limiting ETS exposure in public places do not directly impact exposure in private households. Household smoking restrictions may have effects similar to those observed for workplace and public place restrictions. Not only do these restrictions protect children, but there is mounting evidence that restrictions in homes increase cessation and reduce smoking (Farkas et al, 1999), prevent initiation (Biener et al, 1997), and may be an indicator of the degree of anti-smoking social climate in a community (Farkas et al, 1999). Environmental tobacco smoke is a real, substantial threat to child health. For the vast majority of children, exposure to tobacco smoke is involuntary, arising from smoking, mainly by adults, in places where children live, work, and play. The World 14

PAGE 31

Health Organization estimates that approximately one-half of the world's children breathe air polluted by tobacco smoke, particularly in their homes (WHO, 1999). In the United States, the median prevalence for smoking in 1996 was 23.6% (MMWR, 1997); in 2000 the prevalence was 23.3% (Giovino, 2002); state specific prevalence of in-home exposure of children in 1996 ranged from11. 7% in Utah to 34.2% in Kentucky (MMWR, 1997). The prevalence of current smoking in 2000 was highest among persons involved in childbearing and childrearing age groups, 18-24 and 2544; the Center for Disease Control reported substantial decreases in current smoking prevalence between 1993 and 2000 for all age groups except 18-24 years (MMWR, 2002). In 2000, the prevalence of cigarette smoking among US adults was 28.6% for those with less than 12 years of formal education, 29.5% among high school graduates, 22.6% among persons with some college, and 11.2% among college graduates. Prevalence of current smoking was high among blue-collar workers and service workers, both indicators oflower socioeconomic status (Giovino, 2002). The prevalence of current smoking continues to be higher for those below the poverty level (1999) (31.7% [95%CI = .9]) versus those at or above the poverty level (22.9% [95%CI = 7]) (MMWR, 2002) resulting in a greater burden of negative health effects on children living in these environments. Mannino et al. (1996) using 1991 National Health Interview Survey (NHIS) data, reported that 31.2% of children in United States had daily ETS exposure and 15

PAGE 32

37% had daily or less than daily exposure. This study also reported that the younger children with ETS exposure experienced the strongest negative health effects. Schuster et al. (2001) analyzed NHIS data (1994), and reported that 30% of children live in homes with smoking by residents at least one day per week, 35% live in homes with any smoking, and 34% of children, five years or younger, live in homes with regular smoking. The Environmental Protection Agency estimates that every year in the United States, between 150,000 and 300,000 children under ages one to one-and one-half years suffer respiratory ailments such as pneumonia and bronchitis from breathing second hand smoke (EPA, 1994). According to the Colorado Department of Health and Environment, 193,000 children in Colorado are exposed to ETS in their homes each year (Tobacco Use Prevention and Reduction Plan for Colorado, 2000). Smoking bans and restrictions are currently regarded as a primary means of reducing non-smokers' ETS exposure in the workplace and other public places (Marcus, 1992; Schuster, 2002) while investigation of smoking restrictions in households is just beginning (Gilpin et al, 1999; Pizacani et al, 2003). Home smoking bans are a relatively new approach for dealing with ETS exposure, and requiring parents to conform to smoke-free policies in their own homes may be difficult (Wewers and Uno, 2002). The "sanctity of the family unit" restricts the ability of policy actions to diminish tobacco smoke exposure and cultural values sanction the individual's right to make rules in his/her home (Goldstein, 1994). While a growing body of evidence on opinions and beliefs opposing ETS exposure in 16

PAGE 33

home environments exists (Goldstein, 1994), the commonly held belief that governments or other external agents should not interfere with behavior in private settings has contributed to a lower level of support for ETS control measures in homes compared to other settings (Ashley and Ferrence, 1998). This belief contrasts with other laws and regulations protecting children from physical and sexual abuse in the privacy of their homes, and those requiring school attendance, immunizations, etc. (Ashley and Ferrence, 1998). Further investigation of household factors impacting smoke exposure and evaluation of efforts to reduce ETS exposure are necessary if efforts to provide smokefree home environments for all children are to be successful (Bek, Tomac, et al., 1999; Wewers and Uno, 2002). The emergence of environmental tobacco smoke control in home environments presents unique challenges and as a public health priority raises a host of social, legal, and political issues. The potential for change rests not only on supportive public, professional, and political attitudes with regard to protection of children from harm, but also on the realities of housing, income, education, and child care (Ashley et al, 1998). Understanding the factors impacting ETS in family controlled spaces such as homes and automobiles is a beginning point from which to address these issues. Identifying factors which influence smoke exposure can be the first step in developing interventions appropriate for reducing ETS in the home. Interventions are needed which will be effective in reducing the child's exposure 17

PAGE 34

from all sources (Institute of Medicine, Clearing the Air: Asthma and Indoor Air Exposure, 2000). Theoretical Framework Concepts of social ecology provide the umbrella in which environmental tobacco smoke exposure in a household setting can be understood. Household production of health framework organizes specific macro-level and micro-level factors from social ecology so that they can be measured, and inputs and influences can be accounted for as they impact factors. Finally, household production of health provides a method by which relationships between factors and outcomes can be addressed. Social Ecology Social ecology considers the nature of people's transactions with their physical and sociocultural surroundings (Sallis and Owens, 1997). This approach considers micro-level to macro-level factors operating in a synergistic fashion with individuals, groups, organizations, communities and populations. As such, it honors factors that influence individuals through their social and physical environments, and allows for consideration of health problems such as tobacco smoke exposure and solutions at a variety of levels (Sallis and Owen, 1997; Corbett, 2001 ). People are but one component of the larger behavior-setting system, which restricts the range of 18

PAGE 35

their behavior by promoting and sometimes demanding certain actions and by discouraging or prohibiting others (Sallis and Owen, 1997). The social ecology perspective maintains that multiple levels of influence are important for understanding the problem of environmental tobacco smoke exposure. Smoking behaviors of individuals and households are considered within the context of cultural norms, environmental cues, and infrastructure constraints including costs and restrictive polices (Corbett, 2001). As a result, comprehensive strategies targeting prevention or reduction of environmental tobacco smoke exposure at the level of individuals, groups or social networks, organizations, communities, and populations may be implemented. Household Production of Health While a social ecological perspective provides overall guidance for understanding problems and solutions associated with environmental tobacco smoke exposure, household production of health provides a basis upon which factors directly related to households can be studied as they relate to ETS exposure. Berman, Kendall, and Bhattacharyya (1994) describe use of the household and its application in research and social action. Household production of health is a "dynamic behavioral process through which households combine their knowledge, resources, behavioral norms and patterns with available technologies, services, information, and skills to maintain and promote the health of their members" (Berman, Kendall, and 19

PAGE 36

Bhattacharya, 1994). This approach focuses on the presence and maintenance of health; because the concept of health has multiple determinants, there may be various pathways through which a household can maintain a level of health. Behaviorally chosen determinants have a biological impact on health (DaVanzo and Gertler, 1999), and integration of disciplines allows for a broad investigation of these determinants as well as health changes in the household unit (Berman, Kendall, and Bhattacharya, 1994). Household processes are becoming more critical as determinants of impact as health interventions increasingly rely on behavior changes to produce benefits (Berman et al, 1994). Households are a part of a larger social and economic environment and are best analyzed in that context. Utilizing the household as a central unit of focus allows us to examine both internal and external events and conditions that impact the health of householders (Berman et al, 1994). In the household, individual factors such as knowledge, attitudes, beliefs, and economic considerations (MattilaWiro, 1999) interface with social networks, social support and other factors from the environment to further explain health behaviors. Social ecology provides the context in which all events, both internal and external can be considered in assessing the household production of health. Intra household health behaviors occur or arise from within the household and are well defined in the household production of health model. These include feeding practices, child-care, health seeking behaviors at home, home hygiene and sanitation behavior, 20

PAGE 37

use of preventive and curative services for health. They may be congruent with or differ from external social norms External factors including household income, education, social status arise from social and economic environments in which households find themselves. While these are defined by environments outside of the household, they impact how households view themselves and determine health behaviors. Both intra-household and community factors are influenced by patterns of knowledge, cultural norms, and expectations of efficacy, described by Berman et al. (1994) as inputs to health. All three are important in influencing health behaviors of the household. The complex, dynamic process resulting from the interaction of these ubiquitous factors is not easily studied. Household production of health provides the framework for studying specific factors influential in environmental tobacco smoke exposure and social ecology provides the overall umbrella through the problem, solution, and interventions can be addressed at many levels. Households and Smoking Households are frequently used as the unit of analysis in studies associated with health behaviors as they contain several defining characteristics, relevant to determination of health indicators and health responses. Household is defined as a task-oriented residence associated with physical location (Netting, Wilk, and Arnould, 1984) composed of members living together by mutual consent (Dwyer and 21

PAGE 38

Bruce, 1988), a social environment for child rearing, and a setting for child health interventions (Berman, Kendall, and Bhattacharya, 1994). It is an economic and social unit in which individual and environmental factors shape health outcomes, health promoting behaviors and behavior change (Netting, Wilk, and Arnould, 1984). The household is a primary locale within which daily life takes place as well as the institution for primary social and biological reproduction. It is usually the physical locale and social environment for child rearing and child health interventions. Ideas of social order emerge from households and manifest themselves within the context of intra-household relations. This social order may reflect the larger society or it may offer exceptions from those norms (Berman et al, 1994). Consequently, understanding how households function within the social and cultural environments in which they exist is essential in understanding behaviors related to health problems such as environmental tobacco smoke exposure. The concept of household policies is considered in this project to highlight and target family and household practices and rules based on beliefs, knowledge, and communication about exposure to tobacco smoke. It is at the level of the household that behaviors emerge, directly impacting the health of residents living within. These behaviors do not occur in a vacuum, but are influenced by cultural norms and environmental cues arising from the surrounding environment. Individual factors influence household policies and these are considered as influences on household policies. From the complexity of the household and its surrounding ecological 22

PAGE 39

influences, policies are determined and implemented directly impacting the health of those within. The concept of household was selected because it provides a broader perspective and a more accurate and comprehensive view of smoke exposure as a health risk for children. Individual and family factors are considered under the umbrella of the household as they affect health practices and behaviors. The household production of health framework is the organizing framework through which the relationships between these variables, measured as household and individual factors, can be assessed. The Household as a Unit of Analysis Household has been defined as a place, a mode of social organization, or a cluster of functions (Berman et al, 1994). Inherent in the household are people who live together by consent and who perform certain social and economic functions (Berman et al, 1994). This provides a broader, more comprehensive perspective in addressing factors in a child's life impacting his/her health than does considering just a family unit. However, the defining characteristics of a household are complex and do not lend themselves easily to measurement. Consideration must be given to physical locale, functions within the household, and household patterns in understanding health indicators and health effects of the household. This research focused primarily on the home as a household environment. The home (for which the household is a setting) is the greatest single source of 23

PAGE 40

environmental tobacco smoke exposure of young children and little is known about how smoking policies are determined or implemented in that setting (Ashley and Ferrence, 1998). "Home" and "household" are used interchangeably. Families are two or more individuals who live collectively by consent in a household, including adults and children. Households include the concept of families in a physical locale, with tools for social control (Netting, Wilk, and Arnould, 1984) and social reproduction. While households have been studied extensively from an economic perspective, little is known about composition, functions, or behavioral processes as they relate to environmental tobacco smoke exposure. Little is known about how inputs and influences impact environmental tobacco smoke exposure of young children within the household context. These cannot be studied in isolation; rather, they are considered in context of outside influences impacting the household. The knowledge and range of importance of health effects, beliefs about efficacy of environmental tobacco smoke reduction, and the social and economic factors impacting ETS exposure are explored within the framework. Factors that influence behavior change in social and policy environments and are applicable in the household context need further consideration as well. This research will provide an opportunity to begin investigating possibilities for policy options for families as a basis for behavior change related to environmental tobacco smoke exposure. 24

PAGE 41

Environmental Tobacco Smoke Exposure Environmental tobacco smoke is composed of two types of smoke: side stream smoke and exhaled mainstream smoke. Side stream smoke is emitted from the burning end of a cigarette in between puffs. It contains higher concentrations of chemicals such as ammonia, nicotine, carbon monoxide and carcinogens. Exhaled mainstream smoke consists of smoke which escapes from the burning end during puff-drawing and gases which diffuse during smoking through the cigarette paper (Brownson, Figgs, and Caisley, 2002). Environmental tobacco smoke contains forty three chemicals that are known human or animal carcinogens (Environmental Protection Agency, 1994): eye and respiratory irritants, systemic toxicants, solid particles, and semi-volatile and volatile organic compounds. Significant amounts of nearly thirty volatile organic compounds have been measured, remaining in the air for prolonged periods of time following the smoking of a cigarette (Institute of Medicine. Clearing the Air: Asthma and Indoor Air Exposures, 2000). Exposure to side stream and exhaled mainstream smoke are a threat and result in similar adverse health related affects to active smokers and those non-smokers exposed to ETS (Wewers and Uno, 2002). Cotinine and Smoke Exposure Estimating the public health impact of ETS exposure in children and the extent of that exposure are essential in providing the information necessary for action 25

PAGE 42

to reduce health consequences ofETS (WHO: International Consultation of Environmental Tobacco Smoke (ETS) and Child Health, 1999). There are currently no means by which harmful components of ETS can be directly measured in the organs of interest. Indirect measures of ETS exposure have been developed including self-reports, biological markers and environmental air monitors (Matt et al, 1999). These measures differ considerable in terms of reliability, validity, potential biases, cost, and ease of administration. Cotinine, the major proximate metabolite of nicotine, is the most widely used and considered to be the best biological marker of ETS exposure (Rickert, 1999). It can be detected in blood, saliva, urine, semen, and hair. Its presence indicates that a person has been exposed to nicotine, but does not measure direct exposure to disease causing constituents. The concentration of cotinine is affected by individual differences in uptake, distribution, metabolism, and excretion of nicotine. In breastfeeding infants, the concentration is also influenced by mothers' frequency of breastfeeding, smoking behavior, ETS exposure and use of nicotine replacement therapies (Matt et al, 1999; Haufroid and Lison, 1998). Cotinine is shown to be valid over time, with a half-life of 32-82 hours (Rickett, 1999; Jarvis, 1999) with variations from 16-82 hours (Haufroid and Lison, 1998; Peterson et al, 1997) compared to nicotine that has a half-life of30 minutes to 2 hours. It provides evidence of smoke exposure from several days to a week, but cannot measure cumulative exposure over previous months and years. The half-life of 26

PAGE 43

cotinine is typically longer in infants and young children averaging from 40 hours (18 month) to 65 hours (neonates) (US EPA, 1992). In spite of its limitations, cotinine is recognized as the most sensitive and specific biomarker readily available (Manuel, 1999). Repeated measurements of cotinine have been shown to provide a more accurate descriptions of an infant's ETS exposure but a single measurement may be used as binary marker of passive smoke exposure as well (Woodward and Al_Delaimy, 1999; Peterson et al. 1997). Urinary cotinine excretion is variable across and within individuals, depending on renal function, urinary flow rate, and urinary PH. Urinary results may be expressed as nanograms of cotinine per milligram of creatinine in order to correct for differences in dilution effects. Low levels of creatinine in infants compared to adults may result in cotinine to creatinine ratios that are higher than for adults (Tobacco Monographs, 27; Watts, 1990; Haufroid and Lison, 1998). Racial differences among children may also affect cotinine levels with black children having higher urinary cotinine levels than white children (Knight, Eliopoulos, Kelin, and Greenwald, 1996). The presence of nicotine in some foods such as eggplant, potatoes, tomatoes have raised concern regarding impact on urinary cotinine values but studies thus far indicate negligible interferences from these substances (Haufroid and Lison, 1998; Jarvis, 1994). Using a biomarker such as cotinine as a quantitative measure of ETS exposure validates self-reports ofETS exposure in households because it is specific for tobacco 27

PAGE 44

smoke exposure. Because of the longer urinary half-life of cotinine as compared with nicotine and of the absence of sample contamination during acquisition, it is currently considered the marker of choice. It is a reliable indicator of health outcomes related to smoking practices in households because of the consistent correlations between urinary cotinine and daily tobacco consumption (0.39 to 0.99 and usually greater than 0.75) as reported by Haufroid and Lison, (1998); 0.62 reported by Seifert et al, (2002); 0.50-0.63 for smoking mothers reported by Matt et al, (1999). Definition of Terms Health Behaviors, Indicators, and Outcomes Health behaviors include the actions of individual, household, and groups and the determinants, correlates, and consequences of those activities (Glanz et al, 1997) to improve and enhance the quality oflife. Included are observable, overt actions and mental events and feeling states surrounding these actions which can be reported and measured as (Glanz et al, 1997). In this research, the individual and household activities and the perceptions of these actions will provide measures of health. Indicators reflecting these measures include the parental rating of children in the household as healthy or not healthy, reported numbers of minor illnesses occurring in children in the past year, and the incidence of chronic illnesses and asthma among 28

PAGE 45

children in the household. Inputs to health and influences behind those inputs are also considered as they produce or influence health activities. Health outcomes in households may be quantified in terms of morbidity and mortality, resulting from health behaviors measured over time. For the purposes of this study, health outcomes are not measured directly but quantified through health indicators. Health indicators are those measurable factors demonstrating behaviors associated with prevention and/or reduction of smoke exposure. They include reported smoking bans in households and reported health of children in households. Measurement of cotinine is a health risk indicator as it is used to measure smoke exposure, validating parent reports of smoking bans or no smoking bans. Newborn through pre-school age children live in close proximity to their parents, and do not have the ability to voluntarily remove themselves from the household. Consequently, testing of their body fluids provides the most reliable and valid indicators of smoke exposure in the household. Smoking Policies in Households Smoking policies refer to the rules and practices within the household impacting the environmental tobacco smoke exposure of young children who reside there. These involve rules about whether or not any smoking by anyone is permitted in homes or vehicles. They have been identified as "informal controls" by Goldstein (1994) as opposed to "formal controls" (laws, fiscal measures, and bureaucratic 29

PAGE 46

regulations) implemented in workplaces or public places to decrease ETS exposure. Household smoking rules take many forms and include smoking behaviors allowed in households. Examples of smoking policies include: the last time that someone smoked in the home, number of cigarettes smoked in the home, number of smokers living in the home, locations in home where people are allowed to smoke, or situations in which people are allowed to smoke. The rules that are articulated or practiced related to smoking within a household are considered as they modify health practices and produce health effects within that household. Smoking policies are categorized in three levels in households: no smoking restrictions (no home smoking ban), some smoking but with restrictions (partial home smoking ban), no smoking allowed (complete home smoking ban) (Gilpin et al, 1999; Pizacani et al, 2002; Wewers and Uno, 2002). Smoking policies comprise the health behaviors in the household production of health framework as specified for this research. Smoking Bans For the purposes of this research, smoking bans are selected as the measurable indicator for smoking policies in households. Other smoking behaviors are measured and compared to smoking bans. A complete home smoking ban is defmed as no reported smoking in the household at any time; a partial home smoking ban is defined as smoking with restrictions to either time or place in the home; a no home smoking 30

PAGE 47

ban is defmed as smoking with no restrictions in the home. For purposes of analysis, the partial and no home smoking bans are combined into one category identified as no home smoking ban; complete smoking bans are reported as smoking bans. Socio demographic characteristics (independent variables) are investigated as they relate to smoking bans (dependent variable). In later analysis, smoking bans are the independent variable and cotinine measurements of smoke exposure are the dependent variable. Smoking bans in vehicles include the same categories as home smoking bans and are grouped using the same criteria. Complete vehicle smoking bans allow no smoking in vehicles at any time, partial vehicle smoking bans allow smoking with some restrictions including when children are not in the car or only with the window down, and no vehicle smoking bans allow unrestricted smoking in vehicles regardless of who is in the car or if the windows are down or not. The following schema will be utilized as a framework through which household actors and their relationships to household production of health and health outcomes related to smoke exposure are examined. 31

PAGE 48

Figure 2.1: Social Ecology, Household Production of Health and Environmental Tobacco Smoke Exposure in Households Social Ecology and Household Production of Health HOUSEHOLD Household Smoking Bans ... ETS 1 Related Behaviors Measurement of Children's ETS Exposure Intra-household factors are identified through exploration of household smoking policies, attitudes, actual smoking behaviors in homes, knowledge of smoke exposure with families containing young children. Macro-level or community factors are identified through demographic data. Smoking behaviors are categorized in households and measured as health behaviors. Health indicators related to smoking behaviors are measured through health of children in the home and cotinine analysis of urine of an age appropriate child in the household. The emergence of environmental tobacco smoke control in home environments presents unique challenges and as a public health priority raises a host 32

PAGE 49

of social, legal, and political issues. Reducing or eliminating tobacco smoke exposure in homes rests not only on supportive public, professional, and political attitudes with regard to protection of children from harm, but also on the realities of housing, income, education, and child care (Ashley et al, 1998). Identifying factors contributing to ETS in family controlled spaces such as homes and automobiles is a beginning point from which to support and encourage household smoking policies with known effects. IdentifYing factors which influence smoke exposure can be the first step in developing interventions appropriate for reducing ETS in the home. Interventions are needed which will be effective in reducing the child's exposure from all sources (Institute of Medicine, Clearing the Air: Asthma and Indoor Air Exposure, 2000). In summary, within the context of social ecology, the household production of health as a dynamic behavioral process combines both intra-household factors with external health behaviors (resources, skills) and leads to health outcomes (Berman, Kendall, and Bhattacharya, 1994). In this research, intra-household and community factors, and inputs to these factors are examined using a qualitative approach. These, along with actual health behaviors related to smoke exposure, are measured through a survey. Health behaviors are then compared to health indicators, through cotinine analysis. Understanding the impact of events and conditions internally and externally can reveal the domestic strategies that people employ to deal with factors in their environment to constrain or promote health (Clark, 1998). 33

PAGE 50

Review of Literature Developing a model that includes household factors impacting smoke exposure of young children is a daunting task. The epidemiological literature establishing associations between ETS exposure and health effects focused on predictors ofETS exposure, development of new and improved methods of measurement of exposure, relationships of ETS exposure and measurement with incidence, prevalence and morbidity of illnesses. Many studies were population based utilizing data from national surveys (NHIS, NHANES) with subjects of all ages, while others focused on pre-school or school-aged children. Studies were reported from the US, Great Britain, Scandanavian and Nordic countries, Europe, Africa, New Zealand, and Australia. While becoming more methodologically rigorous over the past decade, study designs include retrospective, cross-sectional, case cohort, longitudinal, and a few randomized controlled clinical trials. Analysis included univariate, bivariate, and multivariate, as well as linear regression, logistic regression, correlation and analysis of variance. Relationships between smoking by intensity and measured exposure through cotinine analysis were found in most studies in which measurement occurred. Behavioral factors were identified less consistently However, knowledge, attitudes, and beliefs, and health protection of children were routinely identified The following review further expands the epidemiological components of smoke exposure, identifies smoking policies and practices in 34

PAGE 51

households, and explores the impact of knowledge, attitudes, and beliefs on behaviors related to smoke exposme. Epidemiology of Smoke Exposme Morbidity and mortality associated with tobacco use was first recognized in the early twentieth century. The effects of tobacco smoke exposme on the nonsmoker was recognized as early as 1928 (Doll, 1998). Medical evidence of the harm done by smoking had been accumulating for two hundred years but was largely ignored until five case control studies were published relating smoking to development of lung cancer in 1950. Studies in the next two decades showed that active smoking was associated with other diseases as well (Doll, 1998). In the early 1970's, mounting evidence linked parental smoking to increased risk for more severe lower respiratory illnesses dming the first years of life. The fust studied effects of ETS exposme in children were increased risks of ear, nose and throat diseases which evidenced themselves as a result of having spent part of a Sunday in a smoke-filled car (Sasco and Vainio, 1999). Colley (1971) found an increased risk of bronchitis and pneumonia in children during their fust year of life iftheir parents smoked. In the late 1970's, reduced lung function in children associated with smoking in the home was reported, and confirmed over the next decade. The Smgeon General's report on Health Effects of Involuntary Smoking (1986) declared a causal relationship between involuntary smoking and lung cancer. US Environmental 35

PAGE 52

Protection Agency (1992) classified tobacco smoke as a known carcinogen. Rubin and Damus (1988) reviewed studies investigating possible associations between passive smoking and health effects in children to explain the wide range of effects reported in the literature. They concluded that studies confirmed the effects of passive smoke on child health but more methodological rigor was needed to delineate the "dose-effect" relationship of the toxin. Additional adverse health effects have now been linked to involuntary exposure of children to tobacco smoke: increased prevalence and exacerbations of asthma (Cook and Strachan 1999; DiFranzia and Lew, 1996; Joad, 2000; Wahlgren et al., 2000), upper respiratory irritation (Cook and Strachan, 1999; Joad, 2000), decrease in lung function (Charlton, 1994;Cook and Strachan, 1999; Sasco and Vainio, 1999), middle ear disease (Ey et al., 1995; DiFranzia and Lew, 1996; Gaffuey and Lynch, 2000), Sudden Infant Death Syndrome (Charlton, 1994; Joad, 2000; Samet, 1999), childhood cancer (Sasco and Vainio, 1999) and lung cancer (Charlton, 1994; Cook and Strachan, 1999; DiFranzia and Lew, 1996; Doll, 1998; Ey et al., 1995; Li et al., 1999; Norman et al., 2000; Strachan and Cook, 1998). These reviews include both systematic, quantitative meta-analyses (Strachan & Cook, 1998; Ey et al., 1995) and narrative reviews as developed by EPA (1992) and WHO (1999). Overall, Cook and Strachan (1999) identify a consistent picture with odds ratios for respiratory illness and symptoms and middle ear disease between 1.2 and 1.6 for either parent smoking. Odds ratios in children age 0-2 years (1.55 [CI 1.16-2.08]) 36

PAGE 53

were higher than in pre-school children who in turn were higher than school age children. Cook and Strachan (1998) further suggest a dose-response relationship as evidenced by studies where odds ratios for asthma prevalence are higher when both parents smoke (OR 1.5[CI 1.29-1.73]) than when just mother smokes (OR 1.36 [CI 1.20-1.55]) or just when father smokes (OR 1.07 [CI 0.92-1.24]). There is no evidence identifying at what, if any, level ofETS exposure for a child could be risk free (Institute of Medicine, 2000). The health effects of smoke exposure on young children have been well established; other factors that strongly influence actual smoke exposure of young children have been explored. Community factors include demographic phenomena, economic costs, social support and cultural norms resulting in expected normative behaviors. Intra-household factors reviewed are smoking policies within households and in vehicles, actual smoking practices in households, and the knowledge of environmental tobacco smoke exposure. Other influences (inputs) are beliefs, attitudes toward ETS exposure impacting these factors. Passive Smoke Exposure in Homes: Intra-Household Factors Children's vulnerability to ETS and the difficulty that they have in protecting themselves from the threat imposed by adults places the burden of reducing or eliminating ETS exposure on adults in protective capacities (Ashley and Ferrence, 1998). Most ETS exposure of young children occurs at home with parents who 37

PAGE 54

smoke or allow smoking in their home. Maternal and paternal smoking habits affect ETS exposure of children. (Cook et al,1994; Willers, Axmon, Feyerabend, Nielsen, Skarping and Skerfving, 2000). Dose-response relationships that were associated with the amount smoked by both parents as measured by cotinine levels in children were reported by Cook et al, (1994). Mothers were less likely to smoke than fathers, but when they did, the effect on their children's cotinine concentrations was greater. Mother's smoking status was the most important predictor of urinary cotinine concentration (Cook et al, 1994; Jordaan et al, 1999: Willers et al, 2000). A significant relationship between cotinine levels of children and the total reported amount of tobacco smoked indoors by parents and others was also found. Jordaan et al, (1999) found maternal smoking to account for 21.8% of variation in urinary cotinine studies of school age children while male parents or other household smokers accounted for 12.7% of variation. Bahceciler, Barlan, Nuhoglu, and Basaran (1999) evaluated the effects of parental smoking modification on cotinine levels of children. Although they had a very small group of subjects (n=77), they found that children whose parents reported smoking indoors had significantly higher cotinine levels than those in whose homes there was no smoking (p<.001). Children whose parents reported smoking on the balcony also had significantly higher cotinine levels (p<.002) than those in homes with no smoking. Parental reports of exposure and no exposure were consistent with urinary cotinine levels in 84% and 82% of children, respectively. They concluded that there was a strong relationship between parental 38

PAGE 55

reports of smoking indoors and urinary cotinine levels but between parent reports of smoking outdoors and cotinine levels. Emmons et al. (1994) compared levels of exposure of adult non-smokers who lived with a smoker and non-smokers who lived with non-smokers when smokers quit smoking. Passive nicotine monitor measurements showed significant reduction in ETS exposure among non-smokers who lived with smokers who quit. Borland (1999) reported cotinine concentrations in children to be affected by frequency with which parents smoke in the same room as the child. He also found that presence of open ventilation and smoking only in restricted home areas reduced cotinine concentrations in the child. Smoking by household members other than parents and smoking by visitors resulted in increases in children's cotinine levels but the magnitude of effect was small compared to parents. Community factors including exposure from outside of the home have also received some attention in the literature. Jordaan et al, (1999) quantified community exposure through measuring exposure of children at school and found that it accounted for 3.3% of variance in cotinine levels. Ownby, Johnson, and Peterson (2000) investigated passive smoke exposure of infants (birth to 2 years of age) from parents and other sources of exposure i.e. daycare workers, visitors in the home, and residents of the home other than child's parents in a longitudinal study conducted over a two year period. Data were analyzed to determine relative contributions from different groups of smokers to urinary cotinine concentrations in the infants. 39

PAGE 56

Frequency of smoking by those other than parents in the household was a significant contributor to urinary cotinine concentrations of infants. A highly significant correlation existed between maternal smoking and quantity of cotinine in an infant's urine; furthermore, smoking by other adults was also significantly correlated with urinary cotinine measurements. Smoking Policies in Households Home smoking restrictions and household bans are a new and important area of research as these are increasingly recognized as the most effective steps that parents can take to reduce ETS exposure of children (Gilpin et al, 1999). Gilpin et al (1999) analyzed Tobacco Surveys (1996) in California looking at smoking status and behavior, household smoking restrictions, and other social variables. Sixty-four percent of households surveyed had a total or partial smoking ban. The presence of non-smokers in the household was a major determinant of whether or not the home was smoke free. Smokers with children in the household were more likely to have smoke-free households than smokers with no children, and the younger the age of the youngest child, the more likely the home was to be smoke-free. Households with both a child and an adult were 5.7 times more likely to be smoke-free than households with neither. A belief in the harmfulness of second hand smoke was also related to smoke-free homes. The results of this study suggest that tobacco control policies promote smoke-free homes A similar study in Oregon (1997) investigated 40

PAGE 57

households according to type and degree of smoking restrictions and explored whether smoking restrictions are associated with decreased environmental tobacco smoke exposure using a cross-sectional telephone survey (Pizacani et al, 2002). This study also found that the presence of children in the home (OR=4.6) and awareness of the harm ofETS (OR= 12.8) were closely associated with full bans on smoking in the home. In contrast, 50% of households with a smoker and children did not have a full ban on indoor smoking. Okah, Choi, Okuyemi, and Ahluwalia, (2002) found that home smoking restriction was associated with presence of children (P<. 0001 ), and a non-smoking partner in the home (P=.002). Restrictions were not associated with age, gender, race, education number of best friends who smoke or perceived harm from smoking which have been previously identified as impacting smoking in homes. This suggests that inner city smokers are concerned about the effects of ETS and take steps to limit exposure in their children. Furthermore, significantly more steps are taken to limit ETS exposure when there is a nonsmoking partner in the home. Demographic Characteristics Smoke exposure in households is related to socioeconomic status of families, level of education, income and family structure, age of children living in home. Low income families may encounter more difficulty providing smoke free environments for their children as they are more likely to smoke, associate with people who are also 41

PAGE 58

------T-,----------------------------------------smokers, live in small housing units which may have limited access to the outdoors, shared rooms, and ventilation systems (Ashley, 1999). Eriksen and Bruusgaard (1995) found, in a cross-sectional study in Norway, that parents were less likely to smoke if they were more than 35 years of age, had a child less than one year of age, had a spouse, and had a "long" education. Smoking parents also smoked less if they had a spouse/co-habitee, had a child under one year of age or had few children. Arborelius, Hallberg, and Hakansson (2000) reported that the Swedish Medical Birth Registry showed a significantly higher proportion of women with less education and from lower socioeconomic groups that were smokers. Lund, Skrondal, Vertio, and Helgason (1998) likewise found in a population based study in Nordic countries that parents with lower socioeconomic status and single parents expose their children to more ETS, in spite of the fact that they are just as likely to report having ''tried" to change their smoking behavior because of their children. Whitlock et al. (1998) found an inverse relationship between socioeconomic status and smoke exposure. He assessed ETS exposure through a self-report survey among adults and measured socioeconomic status through educational level, occupational status, and median household income. Two measures were reported-the number of hours per week spent near someone who is smoking and prevalence of regular exposure to ETS. Both were inversely associated with all three indicators of socioeconomic status (p
PAGE 59

that socioeconomic indicators explained 4.8% of variation in cotinine levels of a population study including children age 6-11 years in South Africa. Interaction affects were found between smoking variables (mother's smoking) and socioeconomic variables (number of people living in the home) (F=4.24, P=.015). They also found that the number of people who lived in the house but not density (number of individuals/room) was a significant predictor of cotinine levels. Mannino (2001) analyzed NHANES III data (1988-1994) in an attempt to identify predictors of cotinine levels in young children age 4-16 years. He found that demographic factors such as age, race/ethnicity, poverty status, and region of the US predict cotinine levels in children with the strongest predictor being the reported number of cigarettes smoked in the home daily. Teen-agers without smoke exposure at home had higher cotinine levels than did younger children who had no smoke exposure at home, suggesting that they were exposed to sources outside the home. Health Care Utilization and Economic Costs Closely associated with the relationship of passive smoke exposure and health affects in children, is increased medical care utilization by children and greater related costs such as prescription drugs for respiratory complaints, more school absences, and increased health care costs (Archives ofFamily Medicine, 1994; Stoddard and Gray, 1997; Wewers and Uno, 2002). Passive smoke exposure is associated with 19% of all expenditures for childhood respiratory conditions 43

PAGE 60

(Stoddard and Gray, 1997). DiFranizia and Lew (1996) estimated that 15%-25% of all hospitalizations of infants and children with lower respiratory illnesses were associated with ETS exposure. Active smoking is known to vary by socioeconomic and educational status with greater representation among low income and less educated (Ashley, 1999) suggesting that children living in these households have greater risk for adverse effects related to smoke exposure. Lam, Leung, and Lai-Ming (2001) examined the association between ETS patterns and doctor consultations and hospitalizations of infant participants in a longitudinal study in Hong Kong. Data were collected from families at visits when the infant was 3 months, 9 months, and 18 months. Using multivariate logistic regression analyses in a population based study, they found a higher level of doctor consultation visits for respiratory and febrile illnesses (P<.OOI) in infants who had been exposed to ETS in utero and high hospital admission rates among infants exposed to ETS either before or after birth. They also demonstrated a clear dose response gradient between the total number of smokers at home and increased hospitalizations for respiratory and febrile illnesses (P=.0003) and any illness (P <.001). Economic ramifications of increased utilization were not addressed in this study; however, increased utilization has economic consequences for the health care system as a whole (Lam et al, 200 I). 44

PAGE 61

Knowledge of Health Effects of Smoking Knowledge of health risks of smoking and passive smoke exposure have increased substantially over the past several decades. Population data collected from high-risk urban population in St. Louis and Kansas City, Missouri (Brownson et al., 1992) show that a majority of population were aware of health hazards of smoking (78.2%) and 87.2% knew that passive smoke exposure was hazardous to young children. The belief in harmful effects was inversely related to age and positively correlated with educational level. Current smokers were significantly less likely than never smokers to acknowledge the effects of passive smoking on non-smokers health (OR= 0.5) and were less annoyed by passive smoke. Knowledge of health effects of smoking and passive smoking were lower among older age groups, women, respondents with less education and current smokers. In a randomized clinical trial, Sorum and Bruusgaard (1996) tested the effects of an anti-smoking information program with smoking parents of young children during well-child visits. Information (minimal intervention program) was given to one group of parents by a health visitor and the control group received no information unless they asked. No differences were found in smoking behaviors of the two groups at a one-month follow-up after the interventions. There was possibility of contamination in this study and a large attrition rate that may have impacted results. Several studies have investigated the relationship between knowledge and behaviors that protect children from ETS. Goldstein (1994) found that although 45

PAGE 62

households had knowledge of harmful effects (90%) ofETS, only (24%) had actual controls (24%) in place, restricting smoke exposure. Household Behaviors, Attitudes and Smoke Exposure Results of studies in which attitudes, beliefs, and practices in homes with small children have been investigated show that awareness of the dangers of passive smoke of children is increasing (Al-Delaimy et al, 1999; Ashley et al, 1998; Eriksen and Bruusgaard, 1995; Goldstein, 1994; Norman et al, 1999). These same studies show that from 24% (Goldstein, 1994) to 76% (Norman et al, 1999) of households surveyed have some smoking restrictions in the home. In spite of the increasing awareness of dangers of passive smoke exposure, Ashley et al. (1998), in analyzing population based surveys in Canada, reported that 34% of homes with non-smokers and 20% of homes with daily smokers were reported to be smoke free. Of these, non smokers who thought that parents should not smoke when spending time with children increased from 62.6% in 1992 to 78.0% in 1996. Smokers answering the same questions increased from 16. 7% (1992) to 42.6% (1996). Results of a telephone survey in New Zealand examining attitudes about protection of children from passive smoking (Al-Delaimy et al. (1998) found similar discrepancies. Both smokers and non-smokers indicated that smoking at home and in private cars was almost as unacceptable as smoking in public places with children around Fewer than two thirds of the smokers reported that they refrained from or reduced their smoking in 46

PAGE 63

presence of children. The high prevalence of smoking in the presence of children in spite of knowledge and changes in attitudes indicates the continued need for investigation of factors influencing smoking behaviors resulting in ETS exposure. Fearnow et al. (1998) examined parent activism (how much parent discourages, talks about, and monitors child smoking) and parental permissiveness about children smoking at home in homes containing teen-age children. Data were obtained from a cohort-sequential study of cigarette smoking from adolescence to adulthood. Study variables included smoking status, educational attainment of parents, parental values concerning child nonsmoking, health beliefs about smoking, parents' belief about addictiveness of cigarette smoking and stress. Four outcome variables in which parents were asked how they would deal with a child's smoking were measured on 5 point scale; three of these were combined to measure "parental activism". Another measure, "parent permissiveness", was measured as a single item outcome. Correlations between study variables and parental beliefs about health consequences of smoking were associated with parental activism but not permissiveness. Variables were examined through three regression models including health beliefs, perceptions of addiction, and environmental stress. Health dangers of smoking were more strongly correlated with smoking activism at high levels of parental education. Health beliefs about smoking and parents' values on child's non smoking were only marginally significant. Relationships between parent values and actions were stronger for parents who had negative health beliefs about smoking than 47

PAGE 64

--------------for parents who had more positive health beliefs about smoking. Marginally significant interactions between parents' smoking status and their perceived personal health in predicting parent permissiveness. Smokers who perceived themselves as healthy were more permissive of child smoking; ex-smokers who perceived themselves as healthier were less permissive of their child's smoking. Personal health status was unrelated to parent permissiveness for non-smokers. Arborelius, Hallberg, and Hakansson (2000) investigated methods of preventing smoke exposure in small children focusing on effectiveness of interventions to determine what measures were most effective. They reviewed studies in which varying interventions were used with parents who smoked and evaluated outcomes of these interventions. Demonstrable effects were found for interventions geared to behavioral strategies with parents, patient's beliefs about smoking and effects on children, counseling efforts for stress reduction, strengthening parental self efficacy. No effects were found for interventions focusing on providing factual information about dangers of ETS exposure or interventions in which pediatricians provided feedback regarding cotinine levels of children exposed to ETS. They concluded that interventions effectively focus on provision of a smoke-free environment for children and not on helping parents stop smoking. 48

PAGE 65

Social Support and Smoking The discrepancy between the awareness of the dangers oftobacco smoke exposure and the smoking behaviors of adults in the presence of children suggests that other factors must be addressed if ETS exposure of children is to be reduced. The prescriptions and proscriptions of parents, friends, and peers have shown to be related to the use of tobacco. Goldstein (1994) investigated smoking rules within homes through "informal smoking controls" which he defines as the rules about smoking that individuals impose on one another in their everyday lives. He found that informal smoking norms were directly related to negative attitudes toward smoking and smokers, and to believing that ETS is harmful. He also found that informal smoking controls were less likely to be found as the number of friends who smoke increased and when the partner of the respondent smoked. Formal controls (smoking restrictions and/or smoking bans in public places including work sites) have been emphasized in ETS exposure much more than informal controls. Less attention has been paid to the factors that impact how informal controls are developed and articulated. The sanctity of the family unit restricts the ability of policy actions to diminish tobacco exposure in the home; the political will to monitor, enforce, or change behavior related to tobacco smoke exposure is also lacking. Other efforts are necessary if providing smoke free home environments for all children are to continue and increase (Bek, Tomac, et al, 1999). 49

PAGE 66

Erikson and Bruusgaard (1995) investigated social support as a variable in smoking prevalence among parents with young children. A longitudinal study in which smoking behaviors of parents of small children were assessed at age of 6 weeks, 2 years, and 4 years by self-report of parents. Parents were categorized into levels of social support (low, medium and high) and associations between levels of support and smoking behaviors were investigated. Daily smoking was not related to level of social support. Associations were found between high social support and smoking less than ten cigarettes per day in parents with several children. Smoking ten or more cigarettes per day was associated with medium and low social support. A significant interaction was found between the level of social support and the number of children in the family. Smoking parents were also less inclined to smoke indoors if they had high social support suggesting the importance of strong and available social support in reducing passive smoke exposure. Self-efficacy In addition to knowledge and attitudes, self-efficacy has been identified more recently as an important factor related to the prevention of passive smoking (Crone et al, 2001). Mothers lack confidence in their ability to ask others to refrain from smoking around children. Arborelius et al, (2000) report that in Sweden interventions geared to behavioral strategies resulted in demonstrable effects in reducing passive smoke exposure of children. Specifically, a methodology aimed at strengthening self50

PAGE 67

efficacy of parents demonstrated that 66% of parents involved in the study by self report were successful in reducing smoke exposure (Aborelius et al, 2000). Oakley (1993), in a qualitative study investigating social support in pregnant women, found that in addition to low income, unemployment, and single marital status, smoking was associated with stress and crises involving family illness, relationship problems, financial difficulties, and violence. All women expressed awareness of the health effects of smoking regardless of their own behavior and Oakley concluded that health promotion strategies need to do more than reinforce the moral message to stop smoking for the sake of"someone else" (the baby). Women in these situations were involved in what Oakley described as the paradox of health promoting work that may be health-damaging to them; providing an environment conducive to the health of their children came at the expense of their own health. Interventions for Smoke Exposure Studies of attempts to reduce passive smoke exposure of children report varying success rates. Greenburg et al. (1994) provided an education intervention based on social learning theory delivered over a 6-month period to mothers in their homes. Intervention effects were studied for smoking mothers and non-smoking mothers. Differences in self-reports of smoke exposure between smoking and non smoking mothers were reported by not supported by similar differences in cotinine to-creatinine ratios of children. Hovell et al (2000) provided counseling for smoking 51

PAGE 68

cessation to smoking mothers in a controlled clinical trial in which one group of mothers received seven individual counseling sessions during a three month period and a control group received brief advice to quit smoking and not to expose their children to ETS. Mothers' self-reports of tobacco use and child's exposure to ETS, cotinine levels of urine samples from children, saliva cotinine concentrations from mothers, nicotine monitoring to validate mothers' self-reports were done at baseline, three months, and twelve months. Exposure declined more significantly in counseled group than in control group from baseline to three months (P=O.Oll) and at twelve months (P=0.0002). Significant differences remained at twelve months but neither group showed significant change over time, suggesting that counseling effect was maintained but no later improvement occurred. Urine cotinine concentrations were significantly different for the two groups (P=0 0008) with a slight decrease in urine from children in families counseled over twelve months and a substantial increase in cotinine in urine of children in control families. Both of these studies confirm the efficacy of parental smoking related counseling to reduce children's exposure to environmental tobacco smoke. Summary For the purpose of this research, factors in households were explored and their relationship to presence of household smoking bans and measurement of actual ETS exposure were investigated. The household factors identified from the research 52

PAGE 69

which warrant further investigation include smoking behaviors, demographic factors, knowledge of health effects ofETS exposure, health promoting behaviors related to smoke exposure in households, social support, and attitudes and beliefs regarding health protective behaviors and ETS exposure. In the household production of health framework, these include both intra-household factors and community factors and are investigated equally in the study. Very little research has occurred investigating environmental tobacco smoke exposure as an issue of health within the household context. The household production of health (HHPH) as a conceptual framework has been little utilized in designing programs which focus on presence and maintenance of health related to environmental tobacco smoke exposure. Rather, research in public health has focused on social and behavioral sciences addressing only issues set by the development of technology (Berman, Kendall, and Bhattacharya, 1994). 53

PAGE 70

CHAPTER3 METHODOLOGY Overview This research used qualitative methods in the form of semi -structured interviews and observations, and quantitative methods in the form of a survey questionnaire and measurement of cotinine as an indicator of smoke exposure. The qualitative investigation provided description and understanding, thus informing instrumentation used in the survey questionnaire. Sequential mixed methods, including triangulation, were used to study behavioral issues related to tobacco smoke exposure. Qualitative and quantitative procedures operate from different assumptions and seek to answer questions in different ways (Miles and Huberman, 1994). They elicit different but complimentary kinds of information, which are important to the understanding of the issues associated with smoke exposure in households (Erikson and Kaplan, 2000). Qualitative description is a prerequisite of good quantitative research, particularly in areas that have had little previous research (Pope and Mays, 1995). Parents have been encouraged to ban smoking in homes to reduce environmental tobacco smoke exposure (Wakefield et al, 2000). However, it is 54

PAGE 71

unclear whether restrictions and bans by themselves offer protection from ETS exposure. Little information is available on how families negotiate smoke free environments for their children. Qualitative methods investigated how families address smoking issues in their households with children present, including description and understanding of behaviors and attitudes regarding smoke exposure in households. From these data, information emerged which identified and confirmed factors to be investigated further through the survey questionnaire. Quantitative methods in a descriptive, cross-sectional design examined factors in households associated with smoking rules that affect environmental tobacco smoke exposure of children, age newborn to five years. The design provided for quantitative follow-up of discoveries (Crabtree and Miller, I 992) of the semi-structured interviews through a questionnaire investigating factors and smoking behaviors/smoke exposure in households. Reports of smoke exposure in the household were validated through measurement of cotinine, a biomarker for nicotine. The design allowed further investigation of a model examining health factors associated with the presence of home smoking restrictions, using the household primarily as a unit of analysis. In such a context, smoke exposure in a child's environment could be more accurately measured. As factors were identified and quantified, this design allowed for examination of possible relationships between factors and smoking policies in households. 55

PAGE 72

Qualitative Methods Qualitative methods in this study provided description and increased understanding of smoking behaviors in households to explicate ways that people in particular settings come to understand, account for, take action, and manage their day-to-day situations (Miles and Huberman, 1994). The identification of ideas, values, and beliefs and how people draw on them to make sense of situations, actions, and processes in their lives were explored through semi-structured interviews (Williams, 1998). According to Pope and Mays (1995), qualitative techniques provide description and understanding of situations or behavior preceding more quantitative investigation, are essential in triangulation for validation purposes, and provide a means for exploring complex phenomena not amenable to quantitative research. The inclusion of more qualitative strategies in smoking research provides information on the social and cultural context and personal meaning of smoking necessary for the development of prevention and cessation strategies (Erickson and Kaplan, 2000). Semi-structured interviews were conducted to explore behaviors, attitudes, and experiences regarding smoking in households. The interviews attempted to elicit contextual meaning of smoking from respondents as a prerequisite for the identification of variables and testing relationships between them (Erickson and Kaplan, 2000). The use of semi-structured interviews combined the advantages of obtaining the responses needed through a structured interview approach and the 56

PAGE 73

breadth and richness afforded through a more human-to-human relationship of unstructured interviewing (Denzin and Lincoln, 1998). The interviews attempted to increase understanding about how families really "behave" related to smoke exposure in their homes and to describe their experiences and associated meanings regarding smoke exposure in the context of their household. Household smoking rules or policies, implementation and enforcement of these policies, and factors which families identified as being important in this process were explored. Dialogue with families provided opportunity for rich description and detail, strategic comparison across cases, and to initiate new lines ofthinking (Miles and Huberman, 1994). In addition to the semi-structured interviews, observation was used as an adjunct method of data collection. This involved systematic, detailed observations of behavior and talk (Mays and Pope, 1995), and watching and recording the subjects' behaviors and interactions. The concerns with validity and reliability with the observational techniques able to be employed in this study limited its usefulness as a data collection method. There was no member validation of findings; consistent trends were sought in each situation to ensure reliability. The original intent was to conduct the activities in the "natural" home setting to facilitate observation, but only three households agreed to be interviewed at home. Consequently, observation provided limited information about the working and functioning of individuals within households related to smoke exposure. 57

PAGE 74

Issues ofValidity Qualitative approaches require specific aims and a clear purpose to guide data collection and systematic analysis of the data. Included are explicit sampling strategies, systematic analysis of data, and a commitment to examining counter explanations (Green and Britten, 1998, Pope and Mays, 2000). Procedures to ensure validity and reliability of qualitative fmdings have been the subject of much controversy. Appropriate criteria and how they should be assessed have been topics of much debate. Lincoln and Guba (1985) proposed criteria for establishing measures of validity and reliability in qualitative research. These include: credibility, confirmability, dependability, and transferability. Credibility is associated with internal validity, confirmability with objectivity, dependability with appropriateness of science behind the method, and transferability with generalizability (Lincoln and Guba, 1985). In a similar fashion, Pope and Mays (2000) recommend that relevance and validity be considered using the following criteria: 1. Triangulation involving the use of two or more methods of data collection. 2. Respondent validation in which researcher's findings are reviewed and compared with research subjects and experts in the field. 3. Clear exposition of methods of data collection and analysis. 4. Reflexivity-sensitivity and assumptions of researcher in data collection. 5. Attention to negative cases--deviant case analysis. 58

PAGE 75

6. Transferability of findings to other settings. Consistent with these recommendations, this analysis meets the criteria for triangulation, expert validation, and clarification of researcher bias. Observations, in addition to interviews, were used to overcome the discrepancy between what people say and what they actually do; they circumvent the biases inherent in the accounts people give of their actions caused by factors such as the wish to present themselves in a "good light" (Mays and Pope, 1995). Member checks were conducted with two nurse colleagues and one PhD expert researcher with whom the researcher worked during data collection. Respondent validation with research subjects was not possible because of the confidential nature of the interviews. Data analysis involved referring back to original transcripts and reviewing coding and themes with conversations of the participants, using Atlas/ti software. The researcher also carefully reviewed differences in responses of households that allowed smoking and those who did not. Interview Instrument Questions composing the semi-structured interview (See Appendix A) were based on the household production of health framework, focusing on intra-household and community factors. Specific content was organized using the HHPH framework including intra-household factors, inputs to health, and community factors. Once developed, the questions were reviewed by an expert researcher for validity and 59

PAGE 76

accuracy. After revisions, the finalized version of the interview was conducted with one household before actual data collection began to test the usability of the guide for questions. Questions in the interview were sequenced to facilitate and maintain rapport and good feelings between the interviewer and the respondent (Jones, 1996) to elicit as much description about smoking in households as possible. A modified funnel sequence (Jones, 1996) was used, progressing from general to more specific questions, which were related to questions before them. The interviews began with general questions focusing on health activities in household. More specific inquiries followed about promoting health of children and protection from harm. Next, there were questions on health protective behaviors in the household, and then on sensitive topics based on the assumption that sufficient time for gaining trust and establishment of rapport had occurred (Fontana and Frey, 1998). Smoking behaviors and practices in the home, attitudes and/or beliefs related to smoke exposure, health protection and smoke exposure, rules for smoking in the home, and methods of implementation comprised the more sensitive topics to be included (See Appendix A). In addition to open-ended questions of the semi-structured interview, household observations were made using a checklist (See Appendix A). The size of dwelling as measured by the number of rooms, number of people living in the home, number of smokers living in the home, and evidence of smoking behaviors in home (ashtrays, cigarette paraphernalia, smell of cigarette smoke, etc) were recorded. 60

PAGE 77

These data were obtained at the end of the interview. In instances where observation in the home was not possible, the researcher asked questions to obtain information. A contact summary form was used to collect demographic data, write short field notes, and identify preliminary themes of the interview. Also included were field notes of visual observations of responses and activities occurring during the interview for all interviews. Respondent behaviors and visual expressions were recorded as field notes immediately following the interviews. Sample Selection for Interviews Through systematic sampling, twenty families in households were selected using the following inclusion criteria: 1. At least one child under school-age (age 5) lived in the home. 2. Children in the household were registered to receive services at school-based health centers associated with the School of Nursing Faculty Practice at University of Colorado Health Sciences Center. 3. Participant of household being interviewed could understand, converse, and write in English. Families were asked to participate in the study either by telephone or when they utilized the school-based health center for services. The researcher explained the purpose of the study and invited participation, carefully explaining that it made no difference for research purposes if anyone in the home smoked. As the interviews 61

PAGE 78

progressed, the researcher selected families to insure that households with smoking members and households with no smoking members were included in the sample. Structure of the Interview An adult member, usually a parent, was the spokesperson for each of the households that participated. After completion of the written consent, the researcher verbally confirmed with the person being interviewed that he/she agreed to being taped and that the researcher might make other notes as they talked together. The respondent was also informed that he or she could stop the interview at any time. The interview began with basic general questions, proceeding to more specific questions, and subjects were encouraged to share experiences related to their responses. Probes such as direct questions, repetition of questions, and silence were used as needed to encourage respondents to continue, amplify, or clarify answers (Jones, 1996). The nature of these interviews was such that the person being interviewed became a spokesperson for the household, in keeping with the study design in which the household was established as the unit of analysis. The researcher directed questions toward activities and behaviors of the household; however, it was not always possible to ensure that the responses reflected the characteristics of the household as a unit. In households where there were two parents, both parents were invited to participate in the interviews 62

PAGE 79

The interviews included but were not limited to the topics of discussion outlined in the semi-structured interview instrument. Questions were asked in similar order with potential variations if the family member being interviewed addressed questions before the interviewer asked them. All questions included in the instrument were asked; respondents could choose not to answer if they so desired. Because families with young children were selected for these interviews, children were usually present during the interview, sometimes interrupting the dialogue between the researcher and spokesperson. Although these interruptions interfered with the interview process, they provided opportunity for making notes and observations about the interview. Interviews lasted from twenty minutes to one hour. Role of the Researcher The researcher explained in detail differences in her role as a researcher and as a care provider for the participating families so that subjects could share "truths" regarding their households without being evaluated or judged. Because many of the families were familiar with the researcher as a care provider, establishment of rapport was not problematic. However, the familiarity and comfort level of the family with the researcher as a care provider led to potential bias in the interviews. To address this, the interviewer explained her role as a researcher as being one of hearing and understanding how families deal with smoking in their homes. The researcher prefaced each interview with a statement indicating that how each family deals with 63

PAGE 80

smoking is a household decision. She reiterated several times that, as a researcher, it made no difference to her whether or not there was smoking in the home, and that she wanted to talk both with families who allow smoking and those who do not. It was necessary to repeat and/or interpret that objective role in interviews. The researcher affirmed how families managed their health, encouraged expression of both positive and negative perceptions, attitudes, and feelings, and asked for further explanations on various answers. Subjects frequently asked the researcher for her opinion regarding the questions. The researcher responded by stating that the purpose of the interview was not to provide her opinion about smoking issues but to discover how families deal with smoking "in real life." For some interviews it was necessary to reiterate the importance of the family behaviors repeatedly until they became comfortable sharing their information. If the person being interviewed asked for the opinion of the researcher, she responded by stating that she was more interested in hearing the opinions of the person being interviewed. Each interview ended with the question, "Is there anything else that would be helpful for me to know about you and your family?" On several occasions, the family asked for information about smoking cessation or the dangers of environmental tobacco smoke exposure after the interview was complete and the tape recorder turned off. Several families asked, "So what does smoke exposure really do 64

PAGE 81

to children?" The researcher provided information as requested and in several instances provided resources on smoking cessation after the interview was completed. Data Collection After obtaining informed consent from the subjects including permission to audiotape, the researcher audio taped the interviews; they were later transcribed and analyzed. The researcher completed all transcription to insure that the data were accurately transcribed; many of the tapes contained background noise making understanding the dialogue difficult at times. Numbers were assigned to study participants for transcription purposes. Names mentioned in the taped interviews were omitted in the transcription, identified instead by individual's role in the household (i.e. husband, child etc.). Once transcribed, data were stored in a computer, protected by password. A form used for tracking purposes including participant name, contact information, and number of the interview was attached to the informed consent and kept with research files. This was completed prior to the interview and utilized when necessary to contact households if changes in time and date of interview were necessary. Subjects were given one copy ofthe informed consent and the other copy was retained in a locked filing cabinet. 65

PAGE 82

Quantitative Methods: Survey Research A survey questionnaire was developed to examine the household factors identified in literature and through the semi-structured interviews with a larger sample. Self-reports obtained through the survey were validated by cotinine measurement of urine samples from a child in the household. No names were placed on surveys to encourage truthful responses to questions which might be considered sensitive and which household members might answer as they perceived the "correct answer" to be. Each survey was numbered so that cotinine measurement of urine could be compared with survey results. The survey number was recorded on the specimen cup for urine and on the consent form completed prior to collection of urine so that accuracy of records could be maintained and results could be shared with parents if they requested. Consent forms and surveys were stored separately so that surveys and cotinine results remained confidential. The study instrument investigated individual and household demographics, smoking behaviors in households, home smoking restrictions, knowledge of harms and health effects of smoke exposure, and attitudes/beliefs about smoke exposure. Variables examined in the survey included behavioral norms and patterns in households as measured by individual and household smoking behaviors, health status of children in household, household smoking bans, attitudes/beliefs related to smoke exposure, knowledge of smoke exposure, and demographics. 66

PAGE 83

Outcome variables were selected based on research questions; smoking bans, used to measure smoking policies of households, were categorized by selfreport, and actual smoke exposure of household was measured by cotinine level in urine samples. Household smoking bans as a predictor variable were compared with cotinine level of urine. Predictor variables included both individual and household measures. Parental reports of exposure of older children (over 6 years) dependent on memory and physical parameters have demonstrated acceptable reliability and validity in surveys (Fried et al, 1995). The same has not been documented for young children and infants because of the perceived social desirability not to expose infants to ETS and limitations of the reporter in documenting accurately the duration, proximity, and frequency of exposure (Matt et al, 1999). In this study, memorybased reports by parents were validated by measurement of cotinine, a biomarker of nicotine, providing a quantitative measure of smoke exposure in their children. Instrument Development Development of the survey instrument followed procedures recommended by Aday (1996) and Czaja and Blair (1996) and included survey design, preliminary planning, and pre-testing. A continual process of development, testing, and revision resulted in the final survey instrument (See Appendix B). 67

PAGE 84

Items in the survey were selected based on household production of health framework within a human ecology context. They were divided into three main groups based on definitions of household production of health framework (HHPH): 1. Intra-household factors included actual smoking behaviors, both individual and household. Individual measures include smoking status of respondent. Household smoking practices include number of smokers living in household, number of cigarettes smoked in home in past week, last time smoke exposure occurred in household, situations and/or locations in which smoking is permitted in the household, exceptions to smoking policies in household. 2. Inputs to household policies include knowledge of respondent, attitudes/beliefs of respondent, and factors identified in HHPH that influence smoking behaviors. 3. Macro-level factors included race/ethnicity variables, household size as measured by number of residents in home and number of rooms in household, household income, gender of respondent, educational level of parents, health status of children in household. In addition to the framework of the household production of health, individual survey questions were informed by results of the semi-structured interviews. Sixty-seven questions included multiple choice, fill-in-blank, check all that apply, "yes" and "no" options, true/false, and identification of attitudes using a 5point Likert scale. Included in the 5-point Likert scale was a neutral category to 68

PAGE 85

allow respondents to indicate a more neutral attitude toward the four components of smoke exposure. Aday (1996) suggests that scales with five to seven points are more valid and reliable than those containing fewer categories. The survey incorporated items from previous interviews with known validity and reliability; new questions were included based on literature review and results of qualitative analysis. Smoking behaviors in households and home smoking restrictions questions were published previously by the National Health and Nutritional Examination Survey and by the Centers for Disease Control (MMWR, 1999). A self-administered questionnaire was selected for this study to facilitate disclosure of information which respondents might otherwise be hesitant to provide due to the somewhat sensitive nature of the topic and their desire to present themselves in the best way possible. Because information was collected from subjects, some of whom read English and some of whom only read Spanish, the survey was developed at a fifth grade reading level using the SMOG readability tool (US Department of Health and Human Services, 1989) to facilitate completion by subjects with little or no assistance from the researcher. The survey was finalized in English and transformed into teleform format so that it could be scanned for data entry. The Spanish survey could not be transformed because of the lack of accents and tilde in the computer program. Care was taken to ensure that each survey item had comparable responses in Spanish and English to minimize bias and facilitate statistical description and analysis. A cover letter informed the subjects that by 69

PAGE 86

completing the survey, they were providing permission to participate in the study. This further facilitated self-administration of the survey. Measures based on similar items used in previous research (Norman et al, 1999) relating to household factors and smoking practices were incorporated into the self-administered questionnaire. Some measures were assessed as household measures; others could only be measured at the individual level. The framework for the research established household as the unit of analysis, and this was used whenever possible in reporting results and fmdings. When characteristics could only be measured as individual factors, they were reported as such in the summarization and analysis. Items were grouped into variables, either predictor or outcome. Outcome variables include self-reported smoking policies in households: no smoking allowed (complete ban), smoking allowed in some places or at some times (partial ban), no smoking ever allowed (complete ban). Cotinine measure of smoke exposure in household is used to validate identified smoking policies in households. 70

PAGE 87

Table 3.1 Predictor and Outcome Variables in Survey Level of Domain Variable Measures Proposed Measures Measurement Outcome Household Household Complete ban, partial Reported Parent smoking policy Measure ban, no ban. measurement Household Child Measure Cotinine measurement Quantitative smoke exposure of urine Child measurement Predictors Demographic Individual Age of respondent Parent self report Measure Gender of respondent Household -Educational level of measure parents -Race/ethnicity -Household income -Number of residents -Type dwelling -Number of rooms in dwelling -Perceived health of children living in household Predictors Smoking Individual -Smoking status of Parent self report Behaviors Measure respondent Household -Number of smokers Parent self report measure living in household -No. of cigarettes smoked in house in past week -Last time in which smoking occurred in household -No. of rooms in which smoking allowed -Situations in which smoking allowed in household 71

PAGE 88

Table 3.1 Predictor and Outcome Variables in Survey (Cont.) Domain Variable Predictors Smoking Behaviors Attitudes/Beliefs Knowledge Level of Measures Individual Measure Household measure Individual Measure Individual Measure 72 Proposed Measures -Smoking status of respondent -Number of smokers living in household -No. of cigarettes smoked in house in past week -Last time in which smoking occurred in household -No ofrooms in which smoking allowed -Situations in which smoking allowed in household -Attitudes toward smoke exposure -Attitude toward health protection of children -Attitudes toward legal control on smoke exposure -Attitudes toward smoke exposure of children Health effects of smoke exposure Harms/dangers of smoke exposure Measurement Parent self report Parent self report Parent self report Likert scale Parent self report True/false

PAGE 89

Measures Several measures of smoking behaviors in households were included in the survey, all of which have been documented in previous research. These included the number of smokers living in the household (Borland, 1999), frequency of passive smoke exposure in household as measured by number of cigarettes smoked in household in past week, (Al-Delaimy et al., 1999), indoor locations of smoke exposure by rooms (Goldstein, 1994), smoking status of respondent as a current smoker (smoked within the past thirty days), ex -smoker (smoked, but not in past thirty days) and never smoker. Smoke exposure in the household is measured by the last time that someone smoked in the home from within the past 24 hours to more than a month ago (Crone et al., 2001 ). Smoking status of the parent and the number of children over age 10 who smoked in the home were also measured. To insure that cotinine measurements of urine were accurate and reflective of smoke exposure based on half-life of cotinine in young children, one month was selected as the cut-off point for current exposure. Past smoke exposure measured smoke exposure between one month and one year ago. No smoke exposure represented smoke exposure of more than one year ago. "Past smoke exposure" and "no smoke exposure" were reduced to one category of "no smoke exposure" for multivariate analysis. 73

PAGE 90

The number of cigarettes smoked in the house in the past week assessed a second measure of smoke exposure. The categories were reduced to none, two packs or less per day, and more than two packs per day for summarization and analysis. Smoking rules in household, as identified in the National Health Information Survey (1994 ), comprised the outcome indicator of smoking behaviors in the household and included three categories: no smoking allowed anywhere or anytime, smoking sometimes or in some places in the household, and smoking allowed without restriction anywhere and anytime in the household. These data were reduced into two categories for data analysis: complete smoking ban defined as "no smoking occurs in the household", and no smoking ban defined as "smoking is allowed in the home with or without restrictions". This categorization of smoking in the household for purposes of analysis is consistent with other studies (Ashley et al., 1998; Crone et al., 2001; Norman et al., 1999; Okah et al., 2002). The strength of the relationship between the smoking ban variable and cotinine level of urine sample was also investigated as a validation measure. Five questions comprised a measure of situations in households when smoking was allowed, including presence of children and permission for relatives and friends to smoke in home. 74

PAGE 91

Table 3.2: Survey Questions For Situations in Households When Smoking Is Allowed Smoking is allowed indoors when children are at Yes or No home. Smoking is allowed indoors when children are in the room. Yes or No Smoking is allowed indoors when children are asleep. Yes or No Some relatives are allowed to smoke in my home. Yes or No Some friends are allowed to smoke in my home. Yes or No Higher scores indicated more situations in the home in which smoking was allowed, reflecting a higher level of smoke exposure. The five questions were evaluated for reliability, resulting in a Cronbach's alpha of .8948. Corrected item-tototal correlation of individual items was .6988 or greater. Both of these measures suggest that the five questions provide an acceptable measure of household situations resulting in smoke exposure of children. Smoking practices of friends were investigated through two items: if friends were allowed to smoke in home and if friends allowed smoking in their own homes. Several previous studies (Goldstein, 1994; Erikson & Bruusgard, 1995) emphasized the importance of effects of friends on smoking practices in households. Although this did not arise as a major theme in the qualitative portion of the study, it was investigated as a part of situations in households in which smoking is allowed. 75

PAGE 92

Knowledge of Smoke Exposure. Inputs to smoking behaviors in HHPH were measured by knowledge and attitudes/beliefs reported by households. Knowledge of ETS exposure was measured using a twelve-item true/false tool. Two concepts of knowledge were measured-knowledge of health effects ofETS exposure (Brownson et al., 1992) and knowledge ofharms/dangers of smoke exposure (Fearnow, Chassin, and Presson, 1998; Kegler and Malcoe, 2002). The items identifying health effects were derived from analysis of the semi-structured interviews in which illnesses, commonly associated with smoke exposure, were identified by parents. Table 3.3: Knowledge Assessment CODE ITEM HARMTF If a child is healthy, tobacco smoke will not harm them. SCIENTF The scientific evidence doesn't really prove tobacco smoke is harmful. ASTHTF Indoor tobacco smoke makes children's asthma worse. COLDSTF Children who are often around indoor tobacco smoke have more colds and coughs. CANCERTF People who are exposed to tobacco smoke when they are children are more likely to get cancer as adults. EARTF Children who are often around indoor tobacco smoke have more earaches. SICKTF Smoke exposure hurts children only if they are already sick. 76 CORRECT CATEGORY ANSWER False Harms False Harms True Health Effects True Health Effects True Health Effects True Health Effects False Health Effects

PAGE 93

Table 3.3: Knowledge Assessment (Cont.) CODE ITEM CORRECT CATEGORY ANSWER CHRONTF Smoke exposure hurts children only if they have a chronic illness like asthma. False Health Effects BABYTF A little smoke exposure will not harm a baby. False Health Effects CHILDTF A little smoke exposure will not harm a child. False Health Effects ETSTF Smoke exposure harms children whether they are sick or healthy. True Harms DANGER The dangers of tobacco smoke have been exaggerated False Harms Attitudes/ Beliefs. A fourteen-item scale assessed attitude/beliefs of households using a Likert scale composed of five points. Reponses ranged from "strongly agree" to "strongly disagree" and were categorized as to whether they represented negative attitudes, neutral or positive attitudes toward smoke exposure. Higher scores indicated more negative attitudes toward smoke exposure. Four measures of attitudes/beliefs were assessed: 1. Health Protection of Children against ETS exposure. Participants completed five items assessing attitudes and responsibilities in protecting children from ETS (Fearnow, Chassin and Presson, 1998). 2. Parent beliefs about smoke exposure related to selves or children. Three 77

PAGE 94

items assessed parent response to general statements indicating attitudes toward smoke exposure. 3. Beliefs about effects of smoke exposure on children in household. (Ashley et al., 1998; Al-Delaimy et al., 1999). Three items assessed parent responses to smoke exposure of children in households. 4. Beliefs about legality protection of children from smoke exposure in households. Two items assessed parent beliefs about legal restrictions on household smoke exposure. Table 3.4: Attitude/Beliefs Assessment Code Item Category HARMFUL I believe that smoke exposure is Parent beliefs about smoke exposure harmful to children. of children. HATE I hate it when I see adults smoking Parent beliefs about smoke exposure around children of children. MAD It makes me mad when people smoke Parent beliefs about smoke exposure indoors around children. of children. DEAL It's not a big deal if adults smoke Parent beliefs about smoke exposure around children. OKAY It's okay for people to smoke around Parent beliefs about smoke exposure children as long as they don't smoke around my kids. NOPROB I don't mind when people smoke Parent beliefs about smoke exposure around me. NEVER Children should never be exposed to Health protection of children against environmental tobacco smoke. smoke exposure 78

PAGE 95

Table 3.4: Attitudes/Beliefs Assessment (Cont.) Code Item Category UPSET I get mad or upset when I see someone Parent beliefs about smoke exposure smoking close to a baby. of children. PROTECT A parent should protect their child from Health protection of children against smoke exposure. smoke exposure RIGHTS Parents have the right to decide Beliefs about laws protecting whether or not they will smoke around children from smoke exposure. their children. LAWS We should have laws against smoke Beliefs about laws protecting exposure in home just like in public children from smoke exposure. places and work places. ILLEGAL Smoking indoors should be illegal Beliefs about laws protecting where children live. children from smoke exposure. TEACH It is my job to teach my children about Health protection of children against dangers and health effects of tobacco smoke exposure smoke. Community or macro-level factors were assessed by demographic characteristics of households including age/ gender/marital status of respondent, number of adults/children living in household, educational level of mother and father, household income, type of housing, number of rooms in home, and ethnicity/race of household. "Health of children" living in the household was assessed by three measures: 1. Number of minor acute illnesses in past year of each child living in 79

PAGE 96

household. 2. Presence of chronic illnesses in children in household. 3. Presence of children with asthma in the household. 4. Perceived health of children in household (healthy or unhealthy). The number of minor acute illnesses in the household in the past year was measured as a continuous variable and the other three were measured as categorical variables. Households were categorized as to whether they contained children with chronic illnesses or children with asthma Households who identified all children as being healthy were categorized as healthy and those who identified at least one child as not healthy were categorized as "not healthy". Extensive fmdings in the literature demonstrate relationships between children's health and smoke exposure (Cook & Strachan, 1999; DiFranzia & Lew, 1996; Etzel, 1997; Gergen et al, 1998; Mannino et al., 2001; Pershagen, 1999). "Exceptions to smoking rules" were investigated by asking parents whether they made exceptions for smoking in the home. Weather-related exceptions, relatives visiting, discomfort in telling someone not to smoke in house were identified in semi structured interviews and included as items in the survey. Reliability and Validity of the Instrument Content Validity. Questions adopted from the NHIS survey included smoking categories, number of cigarettes smoked in the home and demographics. The 80

PAGE 97

remaining survey questions (in English) were reviewed for content and clarity by two PhD level researchers, two masters' level nurse practitioners, and two patient service coordinators at clinic sites, and revised based on recommendations made by reviewers. The Spanish version of the survey was reviewed by one bilingual PhD researcher and two bilingual patient service coordinators who had Spanish as a first language and were employed in school-based clinics. Changes in grammar and sentence structure to make the survey more understandable were done based on their recommendations. The surveys were pilot-tested by twenty families in English and ten families in Spanish. Interviews were conducted with families who completed the surveys, problem questions were identified and changes were made to clarify items. The English and Spanish versions of the survey were approved by human subjects committee (COMIRB) before they were administered. Reliability Analysis. Reliability analysis was used to determine the ability of the true/false items to measure the constructs they were intended to measure. A Kuder Richardson 20 (KR 20) was used to determine the alpha coefficient because of the dichotomous nature of the true/false items. Reliability analysis revealed a Kuder Richardson 20 (KR 20) measure of .6003. Of the two measures tested, the corrected item to-total correlation suggested that harms/danger questions were better measures of knowledge than the health effects questions. 81

PAGE 98

Table 3.5: Reliability Analysis of Knowledge Scale Mean of Correct Corrected Item KR20 Answers to Total Correlation Health Effects (6 items) 5.39 (90%) 6 of 6 items <.4 .4175 Harms/Dangers ( 6 items) 5.34 (89%) 3 of 6 items <.4 .5746 Total Knowledge (12 items) 10.73 (89%) 8 of 12 items <.4 .6003 Harms/Dangers (5 items) 4.42 (88%) 2 of 5 items <.4 .6315 Health Effects (2 items) 1.82 (91%) 0 of2 items <.4 .6618 Revised Scale (7 items) 6.24 (89%) 2 of 7 items <.4 .7174 Removing questions with low corrected item-to-total correlation (less than .4000) in the original reliability analysis improved the KR 20 (See Table 3.6). Removing four items measuring health effects and one item measuring harm/dangers resulted in improvement ofKR 20 (alpha coefficient) to .7174. All questions were kept as a measure of overall knowledge and utilized in later analysis. Reducing the number of questions increased reliability of the scale but reduced correlations with dependent variables. Consequently, all items were kept for analysis in which level of knowledge was a predictor of health indicators related to smoke exposure. 82

PAGE 99

Factor Analysis of Attitude Measure. Factor analysis was performed on the 14 items measuring attitude/beliefs. The total number of subjects provided a ratio of 10-15 subjects for each attitude/belief variable measured. A Kaiser-Meyer-Olkin measure of sampling adequacy was .815, making factor analysis an appropriate measure of construct validity for this scale. The 14 items were contained in 4 theoretical constructs (See Table 3.7). The reliability analysis of the 14-item scale was .7906 (Cronbach's alpha). The question pertaining to the rights of parents regarding smoking (RIGHTS) had a low item-to-total correlation, and when the item was removed, the 13-item scale had Cronbach's alpha of .8012. Although the difference was small, the item also had a low reliability coefficient and was poorly written, so it was removed and the thirteen-item scale was used for further analysis. Table 3.6: Factor Analysis: Rotated Component Matrix ITEM FACTOR 1 FACTOR2 FACTOR3 FACTOR4 COMMUNALITIES JOB .755 .655 TEACH .704 .518 PROTECT .698 .541 HARMFUL .676 .480 HATE .619 .677 ILLEGAL .936 .894 LAWS .914 .881 MAD .552 .622 NEVER .685 .637 UPSET .902 .820 NOPROB .615 .631 OKAY .882 .676 DEAL .761 .653 83

PAGE 100

Four components were extracted from the 13-item scale with Eigenvalues of greater than 1, accounting for 67% of cumulative variance in 5 iterations. These components were consistent with the four theoretical constructs upon which the original scale was developed although several items factored differently into groups as a result of the analysis (See Table 3.6). Four groupings of variables remained and these groupings were maintained for further analysis. Table 3.7: Comparison ofltem Grouping in Factor Analysis Theoretical Constructs Item In Original Item in Group After Group Factor Analysis (14 items) (13 items) Beliefs of parents regarding health NEVER* PROTECT* protection of children. PROTECT JOB JOB TEACH TEACH HARMFUL HATE Attitudes of parents regarding smoke DEAL DEAL exposure. OKAY OKAY NOPROBLEMS NOPROBLEMS Attitudes of parents regarding smoke HARMFUL NEVER exposure oftheir children. HATE UPSET MAD MAD UPSET Attitudes of parents regarding use of LAWS LAWS legal means to protect children from RIGHTS ILLEGAL smoke exposure. ILLEGAL *See Table 3.4 (page 78, 79) for explanation of codes. The final step involved computing reliability analysis of the 13-item scale using the groups of variables, which emerged from the factor analysis. The four 84

PAGE 101

components representing theoretical constructs listed below indicate item-to-total correlations of less than 0.4 and alpha coefficients which are adequate to confirm the reliability of the scale as a measure of negative attitudes toward smoke exposure. Table 3.8: Reliability Analysis of Attitude/Belief Scale (13 Items) Corrected Item Cronbach's Constructs Items to Total Alpha Correlation N=211 N=21l Beliefs about Health Protection PROTECT* 0 items <0.4 .8383 (5 items) JOB TEACH HARMFUL HATE Beliefs About Legal control of Smoke LAWS 0 item< 0.4 .9178 Exposure in Home (2 items) ILLEGAL Parents Attitudes About Smoke DEAL 0 item< 0.4 .6390 Exposure (3 items) OKAY NOPROBLEMS Attitudes about Smoke Exposure of NEVER 0 items <0.4 .6799 children (3 items) UPSET MAD Overall Negative Attitudes/Beliefs Toward Smoke Exposure (13 items) 3 items <0.4 .8012 See Table 3.4 (page 78, 79) for explanation of codes. Evaluation of the items using a priori (theory-based) and mathematical (factor analysis) methods led to the conclusion to use all twelve items of the knowledge assessment and 13 of the attitude/belief items to measure knowledge of and negative attitudes toward tobacco smoke exposure. Knowledge items were evaluated using 85

PAGE 102

reliability analysis because of the dichotomous (true/false) answers, a less powerful method. Factor analysis of the attitudes/belief scale demonstrated consistency between the theoretical constructs and the mathematical groupings of the attitude/belief items. An adequate item-to-total correlation and adequate alpha coefficients of the four groups of variables and of the total measure of negative attitudes support the use of these items as a measure of negative attitudes toward smoke exposure among parents who completed the survey. Study Population and Setting The population selected for the study consisted of families of low income with young children, age five and under who reside in School Districts 14, 50, and 27J of Adams County, Sheridan School District in Arapahoe County, and in Arvada of Jefferson County. A low income, ethnically diverse population was selected because the prevalence of smoking is higher and this population carries a greater burden of health effects because of poverty. Because approximately 30% of the population does not read or write in English, surveys were translated into Spanish and administered to Spanish-speaking families. Data summary and analysis considered ethnicity as a variable. Families with children enrolled in 4 pre-school/Head Start facilities and 4 school-based health centers were eligible as study participants. Approximately 1021 86

PAGE 103

children were enrolled in preschool/Head Start facilities and 3717 children were enrolled in school-based health centers. Table 3.9: Ethnicity/Racial Characteristics ofPopulation Number Children White Hispanic Black Asian Other Users enrolled Preschools 1021 406 (40%) 515 (50%) 30 (3%) 40 (4%) 30 (3%) SBHCs 3717 1459 (39%) 1843 (50%) 72 (2%) 120 (3%) 223(6%) Total 4738 1865 (39%) 2358 (50%) 102(2%) 160 (3%) 253 (5%) There were no data available on the number of Hispanic families who did not have English as a primary language in these settings. Participant Recruitment Parents, age 18-50 years, in households with children, age newborn to school age (five years), were invited to participate in this study. Only families with preschool children were included in the study. They were asked to complete a survey and assist an age-appropriate child to obtain a Urine sample to test for smoke exposure. Two hundred twelve households completed the survey and a child from the household provided a urine sample for cotinine testing. An additional 14 families completed the survey without obtaining a urine sample. 87

PAGE 104

Survey data were collected at four Head Start sites in Adams County, one Head Start site in Arapahoe County, two preschools in Adams County, and four school-based health center clinics associated with the faculty practice at the University of Colorado Health Sciences Center, School of Nursing. These sites were selected because of accessibility to the researcher, demographics of the communities including ethnic diversity and low income. Surveys were administered to a convenience sample of households containing parents (age 18-50) and children (age newborn through age 5) in participating preschools and school-based clinics. Some families recruited in preschool/Head Start sites also used school-based health centers for services; they completed the survey at only one site, usually at the preschool/Head Start site. Consent forms obtained at school-based health centers were compared with those obtained at preschools to insure that no one household completed surveys at both sites. Data Collection Procedures Pre-school and Head Start sites were selected because of location and demographics for administration of surveys and collection of urine samples for cotinine testing. The researcher obtained permission from administrators of each preschool, Head Start, and clinic to collect data. This was done at a face-to-face meeting whenever possible to address questions or concerns related to the study. Dates were established and mutually agreed upon for data collection. Copies of the 88

PAGE 105

study proposal and COMIRB (Colorado Multiple Institutional Review Board) protocols were distributed to research sites upon their request. All administrators who were contacted regarding the study requested surveys in both Spanish and English. Of the 226 surveys completed, 140 ( 62%) were completed in English, and 86 (38%) were completed in Spanish. Of212 surveys involving cotinine analysis, 130 (61%) were completed in English and 82 (39%) were completed in Spanish. Data were collected over a three-month period during the fall of 2002. This time was selected because young children spend less time out of doors and more time indoors than in summer making them slightly but equally more susceptible to exposure to tobacco smoke in their homes. Surveys were anonymous and confidential. A cover letter (See Appendix B) with the survey provided the following information: purpose and importance of the study confidentiality of results completion of the survey was the parent's permission to participate purpose of cotinine testing special permission needed to obtain urine specimens from young children plans for reimbursement after completion of survey and obtaining urine specimen. 89

PAGE 106

When possible, data collection was scheduled simultaneously with parent/teacher conferences at Head Start sites and preschools to provide the greatest opportunity for participation; this occurred at four sites. Because transportation is provided for children attending Head Start programs, the simultaneous scheduling increased access to parents who do not transport their children but attended parent/teacher conferences. At the three sites where this was not possible, the researcher and assistant spent from one to two days inviting parents to participate when they brought their children to school or when they picked them up. Parents at all sites were notified by flyers one week prior as to when data collection would occur, so they could come to the site on the date of data collection or call the researcher to indicate interest in participation. Flyers, in Spanish and English, included the date of data collection and a telephone number through which families could contact the researcher. Approximately I 000 flyers were distributed with specific information for each site (See Appendix C). At preschool and Head Start sites, the researcher was located in a visible location, close to a bathroom for data collection. Parents were approached as they entered the building and invited to participate. If interested, they were given a numbered survey and a short explanation of the study including time required, informed consent and collection of a urine sample, and reimbursement for participation. A Spanish-speaking research assistant was available at each site to provide study information, informed consent, and assistance in Spanish to those 90

PAGE 107

families who did not speak English. This person was directly supervised by the researcher and never staffed a site alone. The survey was checked for completeness, and the consent form for obtaining urine was completed, witnessed and a copy given to the parent. Urine samples in the amount of 30 ml were obtained from the child with the parent's consent and cooperation. They were tested immediately for cotinine, a biomarker of nicotine. In most cases, a urine sample was obtained from the one age appropriate child in the household. In situations where there was more than one age appropriate child, urine was usually obtained from the child who accompanied the parent, or from whom it was easiest to obtain the urine. This was acceptable because of the use of the cotinine screening as validation in this study. Specimen cups, numbered to correspond with completed surveys, and instructions for obtaining urine were provided for children who were potty-trained. For children who were not potty trained, two to three cotton balls were placed in the diaper and when saturated were removed and squeezed into a specimen cup for testing. Urine samples collected through standard method and with cotton balls are both reliable methods for obtaining appropriate specimens. Matt et al, (1999) compared results of urine testing using the cotton ball method of collection on one group of children and standard collection method with the other group of children on three different occasions. Using cotton balls to obtain samples did not systematically affect the level of cotinine measured nor did it vary with repeated samples. In Matt's 91

PAGE 108

study, no significant differences were found in cotinine main effects when comparing the two groups over time (p=. 40). Parents were allowed to obtain a urine sample from the child at home if they were not able to bring the child to the site. They were asked to sign permission for the researcher to contact them and provide a telephone number for follow-up in case they were not able to return with the urine. Instructions were given to obtain the urine within an hour before bringing it for testing to insure freshness of the sample necessary for a valid test. One hundred seventy-six (83%) specimens were obtained on site and 36 (17%) were brought to the researcher at a later agreed upon time. Fourteen subjects completed surveys but did not return urine specimens. Two to three telephone attempts or contacts were made to these households to obtain urine specimens. The researcher utilized principles of universal precautions in handling the urine specimens. Urine specimens were tested using the dipstick method immediately upon receipt and results were shared with parents upon request. Results of cotinine analysis of the urine, date of analysis, and the age of the child from which sample was obtained were recorded on the survey for data entry. The urine and specimen cups were then discarded using appropriate universal precautions for handling of body fluids. A letter with more complete explanation of the study was made available for families and staff at various sites. Surveys were administered and urine obtained by 92

PAGE 109

Cotinine Testing Measurement of environmental tobacco smoke exposure of children was accomplished through testing urine samples for cotinine. Because the cotinine testing in this study was used for validation purposes, a screening test was used instead of a more expensive, time consuming, albeit accurate, complete analysis. NicAlert, the cotinine screening test selected, uses a colorimetric method using either saliva or urine. It is inexpensive, easy to store and administer (Nymox, 2001). The test is an immunoassay, which utilizes monoclonal antibody, coated gold particles and a series of avidity "traps" that allow for quantification. Other substances in urine such as glucose, albumin, and hemoglobin do not affect the colorimetric method of analysis. The presence or absence of exposure to nicotine is indicated through measuring cotinine and 3-hydroxycotinine (metabolite of nicotine) levels in urine. It detects cotinine at seven different levels, thus providing a range of exposure, not a definitive measurement. NicAlert has a reported 87% sensitivity and nearly 100% specificity when urine is used as the body fluid. These values are slightly lower than more accurate immunological or chromatography methods of analysis. Cotinine, a sensitive and specific biomarker for nicotine, provides an integrated measure of the total amount of nicotine absorbed from all sources of smoke exposure over previous two to three days. It is considered to be the "gold standard" for assessing questionnaire validity as a surrogate marker for multiplicity of other smoke components (Jarvis, 1999; Ricket, 1999) but is not without limitations. Its 94

PAGE 110

specificity is limited in that it cannot distinguish relative contributions to measured concentrations of smoking by different individuals in different settings (Jarvis, 1999) and can only index current exposure. Cotinine is shown to be valid over time, with a half-life of32-82 hours (Rickett, 1999; Jarvis, 1999), and provides evidence for smoke exposure from several days to a week. It cannot measure cumulative exposure over previous months and years. The half-life of cotinine is typically longer in infants and young children, averaging from 40 hours (18 month-old children) to 65 hours (neonates) (US EPA, 1992). Urinary cotinine excretion is variable across and within individuals, depending on renal function, urinary flow rate, and urinary PH. Urinary results may be expressed as nanograms of cotinine per milligram of creatinine in order to correct for differences in dilution effects. Low levels of creatinine in infants compared to adults may result in cotinine to creatinine ratios that fall into the range reported for active smokers (Tobacco Monographs, 27; Watts, 1990). Repeated measurements of cotinine have been shown to provide a more accurate descriptions of an infant's ETS exposure (Woodward and Al_Delaimy, 1999; Peterson et al. 1997) but a single measurement for an infant may be adequate as a simple indicator variable of passive smoke exposure (Peterson et al. 1997). Urine was selected as the body fluid of choice because of the ease of specimen collection in children up to school age. Obtaining urine samples is more convenient and less intrusive, requires less voluntary cooperation than other body fluids, and is 95

PAGE 111

considered the most appropriate for studies involving infants and young children (Scherer et al, 1999; Jordaan, Ehrlich, and Potter, 1999). The concentration of cotinine in urine provides a representative measure of exposure found in other biological fluids and correlates with serum levels (Jarvis, 1999; Matt et al, 2000). Cotinine measures in this study were utilized to verify self-reports of smoking behaviors in homes with only one reading per subject. Strong positive correlations between urinary cotinine levels in children and the smoking habits of their parents in previous studies suggest that this is an appropriate use of the cotinine analysis (Fried et al, 1995; Hauford and Lison, 1998; Seifert et al, 2002; Willers et al, 2000). The level of 30 ng/ml was selected as the maximum point for which there was no ETS exposure. This level is consistent with cut-off levels in other studies when cotinine is measured using a screening procedure (Haufroid and Lison, 1998; Seifert et al, 2002). The reduced accuracy of the screening test, providing a range of values instead of actual point values of exposure, necessitates the inclusion of a range to ensure maximum sensitivity and specificity of the test. In spite of its limitations, cotinine is recognized as the most sensitive and specific biomarker readily available (Manuel, 1999). Furthermore, since parental reports of exposure of infants and young children have not yet demonstrated reliability and validity (Matt et al, 1999), cotinine testing can help reduce bias and error in reporting with these subjects. 96

PAGE 112

Human Subjects Review Human Subjects' Committee approval was obtained from the University of Colorado at Denver and from Colorado Multiple Institutional Review Board (COMIRB) for the qualitative phase. Written consent was obtained from all subjects at the beginning of the interviews. Participant names were kept confidential as each interview was assigned a number. The cross list of names and interview numbers was kept in a locked cabinet accessible only to the researcher. This list was destroyed after all transcription was completed to further insure confidentiality. The Colorado Multiple Institutional Review Board (COMIRB) reviewed and approved the quantitative portion of the study as an extension of the qualitative phase. All written materials used in the second phase of the study were translated into Spanish by a certified Spanish translator and approved by COMIRB prior to use. A Spanish-speaking research assistant was present at all data collection sites to provide information on the study and informed consent for those families who needed Spanish translation. Written consent was obtained from parents from whose children urine was obtained for cotinine testing. Methods for Qualitative Analysis The purpose of this analysis is to understand and inform the quantitative results of the study. The analysis was done using the framework approach outlined by Pope and Mays (2000) to include the following steps: 97

PAGE 113

1. Familiarization in which immersion in the raw data results in listing of key ideas and recurrent themes. 2. Identification of thematic framework in which all key issues, concepts, and themes by which data can be referenced. 3. Indexing of all data through numerical coding with text descriptors. 4. Developing charts which arrange and rearrange data in the appropriate thematic framework. 5. Using the charts to define concepts, map the range and nature of phenomena, create typologies and find associations between themes to explain findings. Data were organized using Atlas/ti software. Codes were predetermined based on research questions; new codes and themes were added as they emerged. Observations were also coded and counted. Researcher bias was addressed to increase validity of findings. The researcher assumed that all respondents were honest in their discussion of smoking behaviors in their households. It was also assumed that the setting in which the interview took place did not interfere with information obtained via the interview. The researcher assumed that family members reported accurately the points originally intended for observation if the interview was conducted in the home. 98

PAGE 114

Methods for Quantitative Analysis All data were entered into an SPSS data file and were reviewed for accuracy. Data obtained from English surveys were scanned into the SPSS file using the teleform system. All questions on original surveys were proof read by the researcher and were compared with the SPSS file with another research assistant to ensure accurate scanning and entry. Data from Spanish forms were entered into the database by the researcher and proof read for accuracy. Outliers were identified and removed from data set or recategorized into categorical variables. Frequencies on all items entered into the data set revealed less than 5% missing variables in all but 12 items, which had less than 10% missing data. Cases with missing data were excluded from analyses on a case-by-case basis. Inconsistent responses were evaluated within the context of responses to the remainder of the survey. Decisions regarding the best course of action for inconsistent data were made in consultation with the dissertation committee and statistician. It was determined through examination and through tests for skewness and kurtosis that variables did not have normal distribution. Data analysis was done with consideration of this limitation. After the data set was finalized and corrected, data analyses were conducted sequentially, beginning with description of variables. Subject characteristics were examined by comparing socio-demographic variables using the following groupings: site from which sample was recruited (preschool/Head Start and SBHCs), ethnicity of 99

PAGE 115

the household, and households with complete home smoking bans versus households with no home smoking bans. Chi-Square tests were used to compare categorical variables. T -tests were used to compare continuous variables between households with complete smoking bans and households with no/partial smoking bans. All chi square tests and t-tests were two tailed and alpha was set at 0.05. Because of the number ofsociodemographic variables and the number of categories within each variable, the alpha for each socio-demographic variable was corrected to .001 using a Bonferroni adjustment to reduce the chances of Type I error. Spearman and Pearson correlation coefficients were calculated to evaluate associations between continuous variables. Kendall's Tau h correlation coefficients were used when variables did not meet assumptions for Spearman or Pearson correlations (homoscedasticity and normal distribution). Items in the survey were grouped and analyzed as predictor (independent) or outcome (dependent) variables based on study hypotheses. 100

PAGE 116

Table 3.10: Independent Categorical Variables Variable Demographics Number of residents in household Age of respondent Gender of respondent Marital status of respondent Household income Educational level ofMothers Educational level ofFathers Ethnicity of Household Knowledge of Respondent Attitudes/Beliefs of Respondent Health of Children in Household Identified Health of Children Minor Acute Illnesses in past year Children with Chronic Illness Children with Asthma Presence of Complete Home Smoking Ban Last time someone smoked in home Number of smokers living in home. 101 Rationale for Inclusion Crowding associated with increased smoke exposure in households where smoking occurs. Standard demographic variable Standard demographic variable Risk factor for smoke exposure Risk factor for smoke exposure Risk factor for smoke exposure Risk factor for smoke exposure Household production of health Household production of health Household production of health Associated health risk Associated health risk Associated health risk Associated health risk Measure of smoke exposure Measure of current smoke exposure. Measure of smoke exposure.

PAGE 117

Table 3.10: Independent Categorical Variables (Cont.) Variable Rationale for Inclusion Smoking status of respondent Measure of smoke exposure. Situations in which smoking allowed in home. Measure of smoke exposure. Demographic items, level of knowledge of smoke exposure, and attitudes/beliefs in the survey were grouped and analyzed as predictor variables and entered into a logistic regression model to determine which variables affect the probability of reported smoking bans. Four measures of the health of children in household were entered into the model after the socio-demographic variables to determine if they contributed to prediction of no smoking bans. Hosmer and Lemeshow goodness of fit test and model chi-square tests were done to assess model adequacy. Odds ratios for likelihood of a reported smoking ban in the household were calculated. The odds ratio measure has a clearer interpretation than that of the chi square and can be applied to multidimensional cross tabulations (Wang, Eddy, & Fitzhugh, 1995). The dependent variable of interest, reported smoking ban in the household, was coded as 0 =total smoking ban and 1 =no ban. Model building was conducted through direct logistic regression in which all socio-demographic variables were entered simultaneously followed by health indicator variables. This method was 102

PAGE 118

selected because no order of importance of socio-demographic independent variables was identified in the hypotheses. Health indicators, demonstrated previously as related to smoke exposure, were entered after socio-demographic variables to determine if their presence made a difference in smoking bans. These analyses were conducted using the total sample of households (n=225). Bivariate correlation analysis was used to examine the relationship between smoking bans and actual cotinine exposure for those households from which cotinine measurements were obtained. Measurement of cotinine levels validate reports of smoking behaviors in households and reduce potential bias and error found in memory-based reports ofETS exposure in young children (Matt et al, 1998). The outcome variable was dichotomized into smoke exposure as measured by cotinine levels of 2 or greater and no smoke exposure as measured by cotinine levels of 0 or 1. Pearson correlations were computed to assess relationships between reports of smoking bans and ETS exposure through urine cotinine measures. The predictor variable, complete or no home smoking ban was used as the measure of smoking behaviors in the household. Other smoking behavior measures investigated in the survey correlated significantly with household smoking bans. Logistic regression was used to identify relationships between possible predictors and outcome variables using odds ratios. These were conducted on the sample, which provided urine specimens to test for cotinine exposure. 103

PAGE 119

The strength of associations was expressed by odds ratios with significance level of .05. All analyses were conducted using SPSS for Windows (version 11.0). A Bonferroni adjustment was made for socio-demographic independent variables containing several categories to reduce the chance of an inflated Type I error. 104

PAGE 120

CHAPTER4 FINDINGS FROM QUALITATNE DATA ANALYSIS The qualitative phase of the research occurred over a five-month period in which semi-structured interviews were conducted with adults over 18 years of age from households in which a child less than 5 years of age resided. They were recruited from faculty practice clinical sites ofUniversity of Colorado Health Sciences Center, School ofNursing. The qualitative phase met two of the study's specific aims: 1. Explore household characteristics and relationships between factors associated with reported smoking policies in households. 2. Identify means of implementation and enforcement of smoking policies in households where young children reside. Twenty families participated in the interviews. Three families refused to participate; two families indicated that they were not interested; one family agreed to be interviewed but did not keep the interview appointment. Of the three who refused to participate, two were introduced to the study by telephone; the other was recruited in person by the researcher No data were available on those who refused to participate. 105

PAGE 121

Participating families had the option of being interviewed in their homes or at an agreed upon location. Seventeen of the twenty interviews were conducted in a clinical setting at the request of the families. A private, quiet location in clinics was utilized for the interviews to insure subject confidentiality. Three interviews were done in homes. After a verbal agreement to participate in the interviews, informed written consent was obtained (See Appendix A). Description of the Sample Demographic data obtained through semi-structured interviews and observations were summarized and reported (See Table 4.1). Ten (50%) of the households from which a spokesperson was interviewed had smokers living in the home and 10 had no smokers living within. Of the 10 households with smokers living in the home, only 3 (15%) reported that smoking was allowed in the home. Six of the adults interviewed (30%) identified themselves as smokers and 14 (70%) were not. Seventy percent of household spokespersons identified themselves as White, not Hispanic and 30% identified themselves as Hispanic. Each household in which smoking was allowed had an average of 2 adults and 2.3 children residing within; households in which no smoking was allowed had 2.2 adults and 2.9 children. The average age of children in smoking households was 4.5 years and in non-smoking households was 5.3 years. All families were oflow income, based on their fmancial eligibility to receive services at the School-based Health Centers as established by 106

PAGE 122

federal poverty guidelines (2002). Most families lived in their own homes (70%); the remaining 30% lived in apartments, duplexes, or mobile homes. Table 4.1: Interview Characteristics ofHouseholds by Smoking Status Household Smoking Status Smoking No Smoking Allowed Allowed Total N=20 N=3 N=17 Ethnicity White, not Hispanic 3 10 13 (65%) Hispanic 0 7 7 (35%) Smoking Status of Smoker 3 3 6 (30%) Informant Non-smoker 0 14 14 (70%) Average Number Adults 2.0 2.2 2.2 Living in Home Children 2.3 2.9 2.8 No. of Smokers 0 0 10 10 (50%) Living in Home 1 2 6 8 (40%) By Household 2 0 1 1 (5%) 3 1 0 1 (5%) Average Age of Children 4.5 yrs. 5.3 yrs 5.16 yrs. Living in Home Type ofDwelling House 2 12 14 Apartment 1 2 3 Duplex 0 2 2 Mobile Home 0 1 1 107

PAGE 123

Table 4.2: Interview Characteristics of Households by Ethnicitv Ethnicity of Household White, Not Hispanic Hispanic Total N=l3 N=7 N=20 Smoking Status of Smoking Allowed 3 (23%) 0 3 (15%) Household No Smoking Allowed 10 (77%) 7 (100%) 17 (85%) Smoking Status of Smoker 5 (38%) 1 (14%) 6 (30%) Informant Non-Smoker 8 (62%) 6 (86%) 14 (70%) No. of Smokers 0 6 (46%) 4 (57%) 10 (50%) Living in Home by 1 6 (46%) 2 (29%) 8 (40%) Household 2 0 1 (14%) 1 (5%) 3 1 (8%) 0 1 (5%) Average No. Living Adults 1.9 2.3 2.2 in Home Children 2.5 3.4 2.9 Average Age of 4.6 6.2 5.2 Children Living in Home Type of House 8 6 14 (70%) Dwelling Apartment 2 1 3 (15%) Duplex 2 0 2 (10%) Mobile Home 1 0 1 ( 5%) Observation of Smoking Behaviors There was no smoking by respondents during the interviews held in homes. One household allowed smoking in the home and the respondent was a current smoker, one household allowed no smoking and the respondent self-identified as a current smoker, and one household did not allow smoking and had no smokers living 108

PAGE 124

in it. In two of the three homes, there was no evidence of smoking. In the third home, cigarettes were lying on the table and ashtrays were located in the living room. No tobacco smoke smell was noted by the researcher in any of the homes. Results of Qualitative Analysis Semi-structured interviews were transcribed and entered into Atlas/ti software for summarization and categorization. The interviews were read and reread to become familiar with the content. A deductive approach (Pope, Ziebland, and Mays, 2000) was used to examine key ideas and themes based on the household production of health (HHPH). The transcriptions were reviewed to identify intra-household factors upon which the interview questions were based: knowledge of health effects of tobacco smoke exposure, health promoting behaviors of households as they related to smoke exposure, harm prevention activities related to smoke exposure, and presence of smoking policies. Questions were asked about each of these factors, and answers were coded and then organized by text descriptors. These descriptors were further examined and categorized by intra-household factors, inputs to household factors, or community factors. Attitudes/beliefs about smoke exposure were investigated as inputs to smoking behaviors of households. Influences were referenced based on factors influencing household smoking policies as identified by the interviewee. Trends and relationships were reviewed, and concepts to be further tested in the survey were 109

PAGE 125

identified and examined. Meanings associated smoke exposure for households were explored through past experience of respondents and perceptions of their responsibility for protecting the health of children in the household. Out of the attitude and belief questions came positive and negative feelings, and emotions regarding experiences with smoke exposure in the lives and households of the respondents. These factors were investigated to inform the development of questions which could be asked on a larger scale in the survey. Macro-level or community factors were investigated in the interviews with the use of more open-ended questions. The investigation of how families implemented rules and policies about smoking in their households revealed factors arising from within the household and from social networks and community. Factors from outside the household impacting how families determined smoking policies in their homes and the associated meanings were examined as they related to behaviors within the household that encouraged health and health related activities. Description ofThemes Coding began using the pre-determined themes around which interview questions were developed: activities identified by the respondent which were health promoting in the household, activities which prevented harm to children, knowledge of health effects of smoking, knowledge of harms of smoking, reports of actual smoking behaviors in the home, presence of smoking rules, attitudes/beliefs toward 110

PAGE 126

smoking, exceptions for allowing smoking in the home, and locations where smoking is allowed in the home. Specific codes were selected based on terms and/or ideas consistently articulated by parents as they discussed the questions. New concepts emerged as the coding was completed. The use of the exact words or similar words derived from the text were coded and then clustered together. First, activities of households were investigated, followed by exploration of implementation of rules and policies about smoking. The influence of "inputs" and community factors on those activities was explored with associated meanings. Once data were reviewed and coded, 8 themes were further examined and deconstructed into 13 categories to better describe the content. They included: health protection, health habits of household, locations for smoking, smoking harmful to children, illnesses associated with smoking, smoking and pregnancy, smoking and media, reasons for smoking, reasons for not smoking, exceptions for smoking, rules for enforcement of smoking policies, attitudes toward smoking, and respect. The categories were entered into the focused network function of the Atlas/ti software so that associations between actual quotes and categories could be further examined and summarized. 111

PAGE 127

Table 4.3: Summm: of Themes and Codes Identified in Household Interviews on Smoking. Original Themes Developed as Codes Number of Number of Themes for Result of Respondents Quotes Coding Summarization/ Analysis Who Identified Associated With Theme Each Theme Health Health Protection: Promoting Efforts by parents to 13 19 67 Activities of protect children from Household health risks Health Habits of Household: 16 19 68 Activities identified by parents to keep children healthy Locations in Home for 10 18 66 Smoking Knowledge Harms Other Than Health 4 8 16 About Associated With ETS Harms of Exposure Smoking Illnesses Associated With 22 18 51 Smoking (Knowledge) Smoking and Pregnancy 2 7 9 Attitudes Attitudes Toward Smoking-Pro and Con 7 12 34 Smoking Reasons For Smoking as 18 12 58 Behaviors in Identified by Smokers Home 112

PAGE 128

Health Promotion and Health Protection A question on health habits of households was included as a means of eliciting family perceptions of health in their household. Ideas that families had to keep the family healthy and activities in which they engaged to be healthy were encouraged. Their responses could be divided into two categories; one focused on the ideas that they had and the other focused on activities in which they engaged. 113

PAGE 129

Table 4.4: Health Promotion Activities and Health Habits of Families Codes Designated as No. Codes Designated as No. Health Promotion Activities Times Health Habits Times Listed Listed Education Nutrition Teach children 4 Good nutrition I4 Commitment to kids 2 Drink water Safe daycare 1 Vitamins 3 Health Care Three meals I Responsible for health care I6 Prevent Illness Catch them before they get sick. I Health insurance 2 Take to doctor Alternative medicine I Physicals 6 Baby on oxygen 2 Healthy Kids 8 Breathing treatments 1 Keep them away from smoke 4 Preventing Smoke Exposure No smoking allowed 23 Cleanliness Keeping them clean 4 No ashtrays 2 Cream on skin No one smokes 7 Clean house 2 Hope for smoke free environment 1 Other Good parenting 3 Prevent smoking 6 Safety 5 Exercise 2 This was followed by the question about protection from harm, "If someone were to do something to harm your child, what would you do? Give me some examples of that." Parents had a difficult time answering this question when it was 114

PAGE 130

asked generally. They did not understand it and provided a vague answer such as, "I wouldn't let anything happen to my child." When asked a more specific question related to health, they responded by answering, "I keep them away from whatever it is that would hurt them." Perceptions which families shared indicated that meanings associated with smoke and smoke exposure were considered a part of general health of family but not of health habits of household. In three interviews, helping children determine whether or not they would smoke based on parental teaching was discussed. One parent said, "My older daughter, I just let her know that it's not right and she should not do that. She comes up with, well, I'm not taking drugs, Mom." (Interview 05). Another parent, "I let him judge it on his own. He is pretty good at judging." (Interview 6). In two interviews, keeping away from smoke was identified as a health habit in the family. One of these involved a household in which there was smoke exposure originating from another apartment attached to the home. Implementation of Smoking Rules Some respondents did not consider smoking rules or policies important because they did not think smoking was a problem. That is, they assumed that no one would smoke in their home. One parent said, "Nobody we know smokes so we are fortunate, we don't have to deal with that." .... Like I said-we don't associate with anyone who smokes. I've never had to deal with telling anyone-"Don't smoke in 115

PAGE 131

my house." (Interview 6). Another said, ''No, I don't have any friends who smoke. Or family." (Interview 15). Another parent indicated that it is not a problem because everyone knows how they feel about smoking. She said, "Because everybody knows that we are non-smokers and that they (children) have asthma And they'll usually go outside ..... or they'll ask first. But I really haven't had anybody ask." (Interview 4). One parent had established clear rules about smoking activities in the home and felt that they were important because of a child with asthma in the home. He said, "She (grandmother living in the home who smokes) follows the rules. I mean-they all respect the rules. I tell them if they live here, they have to live by my rules." (Interview 7). Another parent with friends who smoke said, "We don't have ashtrays so if they want to smoke they can go outside." (Interview 19). Another respondent expressed difficulty in enforcing rules with people in authority and said, .... We have only one friend who smokes and he smokes outside. The landlord-it was so hard to tell her. You would tell her and she would do it anyway. I'm like--I always say that my husband doesn't like it. It's an authoritive thing and he's very assertive about it. (Interview 1 0). Families who allow smoking also identified rules for smoking behaviors. One parent who allowed smoking in the bathroom did not consider it a part of the house. Her comments suggest that smoking continues to occur in the household. As far as smoking goes, we used to smoke in the house until she was born and then everybody knew that you went to the bathroom and you didn't do it in the house anymore. Sometimes you have to nag-somebody might actually walk in with one (cigarette) But people usually respect us and that's cool." (Interview 3). 116

PAGE 132

Locations for smoking in the home clustered into locations both within the household and outside the household in which smoking was allowed. These included bathroom, living room, kitchen, garage, basement and with an open window. Outside locations included a general category called smoking outside smoking in the workplace, and in a vehicle. Two respondents who are current smokers acknowledged that they smoked in the car. (Interviews 2 &11). One other respondent indicated that smoking occurred in the car with the windows down (Interview 8). Seven indicated that there was no smoking in the car (Interviews 3, 7, 9, 14, 15, 18, 20). Articulation of and implementation of smoking rules were closely associated with smoking status of respondent. Those households who did not have any smoking residents or who did not associate with smokers gave little thought to smoking rules or implementation. Those who associated with people who smoked or smoked themselves addressed smoking in the household by verbalizing their wishes to associates or by taking actions without verbalizing which indicated their wishes. Reports of smoking behaviors in the home were categorized into reasons for and for not smoking in the household and positive/negative attitudes toward smoking. 117

PAGE 133

Table 4.5: Reasons For and Against Smoking in the Household Reasons for Not No. of Smoking Times Listed Dirty 2 Bad smoke smell 16 Finances expensive Food off the table Health of kids No choice for children History of exposure in family of origin 5 1 4 9 Reasons for Smoking No. of Times Listed Vice Smoke when stressed Smoke when fighting with spouse Relieves stress Addiction Social activity Feels good Occupier of time Mind game Time away Distraction Finances Prevent eating Agitated 1 2 3 5 7 5 1 2 2 2 Persons interviewed expressed strong feelings about the pros and cons of smoking in households. Nine non-smokers expressed negative remarks about the smell of smoke. "The smell is bad." (Interview 2). "The house stinks when people smoke (Interview 3) "The smell of it is the worst, I can't stand it." (Interview 5). 118

PAGE 134

"It's disgusting." (Interview 4). Respondents who were smokers expressed guilt about their smoking. One parent said, "I know that it is not good for the baby. I really feel guilty when he gets sick because I know that I made him sick." (Interview 11 ). Another parent, "I am sorry that I am a junkie, you know but anytime I am around a non-smoker I do my darndest to keep smoke away from them." (Interview 3). Smoking rules in the household included identification of smoking policies and exceptions made for smoking rules. Nine respondents identified that they have smoking rules. Two codes reflected smoking rules, one for smoking rules and the other for exceptions. Three reasons for exceptions were noted: cold weather, "hard to tell her not to smoke" (Interview 1 0), and "grandparent has right to smoke in home, cannot tell him what to do." (Interview 1 ). Six interviewees stated that although they would be willing to make exceptions for family members to smoke in the home, it was not necessary. Family members "know the rules" and smoke outside without being requested to (Interviews 4, 7, 1 0). Two other respondents stated that they no longer allow children to visit grandparents overnight because of smoke exposure and grandparents' ''unwillingness" to stop smoking. 119

PAGE 135

Knowledge of Effects of Smoking Knowledge of smoking was categorized into knowledge of harms of smoking and knowledge of health effects of smoking. Respondents provided information on negative health effects (See Table 4.6) of smoking. Table 4.6: Knowledge of Effects of Smoking Health Effects of Smoking on Children Damage Sick Blame illness on smoking Specific Health Effects of Smoke Exposure Eyes irritated Colds Runny nose Congestion Respiratory problems Breathing problems Coughing or sneezing Asthma SIDS Kills lungs Ear infection Allergies Burns Bad migraines Skin Cancer Knowledge of harmful effects of smoking in pregnancy was discussed in seven interviews as a result of the questions asked about effects of smoke exposure. Respondents identified behavior change in smoking habits during the pregnancy, either decreasing smoking or stopping for the extent of the pregnancy. One respondent decreased smoking while pregnant, and three respondents quit during 120

PAGE 136

pregnancy and began smoking again after the pregnancy. Two respondents quit during the pregnancy and never smoked again. Although pregnancy had an impact on smoking behaviors of mothers, this theme was not pursued further because this study focused on effects on children already born. Parents identified nineteen different illnesses or conditions in children and/or adults resulting from smoke exposure. The most commonly mentioned illnesses associated with smoke exposure were used in the knowledge questions in the survey: asthma (9 times), cancer (3 times), ear infections (6 times), respiratory/coughs/colds ( 6 times). SIDS was mentioned 6 times and allergy was identified twice. Other illnesses or conditions mentioned one time each included runny nose, bums, eyes irritated, bad migraines, breathing problems, skin problems, and kills lungs. "Keeping them away from smoking" was identified five times. Parents identified four actions to protect their children from harm related to smoking: "no smoking allowed at all" (13 interviews), "no smoking inside the house," (6 interviews) "prevent smoking" (3 interviews), and "no one smokes". (5 interviews). Damage to children, damage of second hand smoke, danger, and harm to children were identified as a general outcomes of smoke exposure, not directly related to health. Parents' knowledge regarding smoke exposure was consistently high. All had some information on the negative health effects of smoke exposure. Some could articulate clearly the negative health effects related to smoke exposure; others had less information but demonstrated a definite commitment to maintaining a smoke free 121

PAGE 137

environment. Those parents who had knowledge and a previous negative experience with passive smoke exposure expressed the most negative feelings about smoke exposure for themselves or their children. The knowledge demonstrated by these families is consistent with what Goldstein (1994) and Brownson (1992) found among smoking parents. A noted difference in these studies was that even with knowledge of health effects, smoking behaviors in homes did not change or decrease. Attitudes and Beliefs Discussion about smoking being allowed in the home was emotional for most respondents. Eighteen codes reflected reasons for smoking at home and eight reflected reasons for not smoking. Twelve interviews contained content on positives of smoking in the home and thirteen contained content on negatives of smoking in the home. This contrasts with three households that self-reported that smoking was allowed and seventeen households who reported that it was not allowed. Negative aspects of smoking were identified in all interviews. Non-smoking respondents used terms such as "dirty," "disgusting," "smoke is bad," and "offensive to others." Respondents who were current smokers identified "offending the rights of others," "paying attention to those around you (not smoking)," and being "aware" as important. Parents who spoke negatively about smoking raised their voices while talking or expressed disgust and anger about those who smoke. 122

PAGE 138

When parents were asked how they managed smoking behaviors and activities in their household, they gave the following reasons for not allowing smoking in household: cost/finances, dirty and disgusting, smoke smell, smoke is bad, health concerns such as respiratory problems, ear infections, and family history of smoke exposure (9 quotes). Family history of smoke exposure was an important factor in several families' decisions not to allow smoking. Parents identified their own parents' smoking behaviors as influencing their decision not to smoke (Interviews 4, 7, 9,15, and 19). "Well, we both grew up with parents who smoked. And didn't care if we were in the car with the windows rolled up or not." (Interview 4). One parent stated, We hated smelling like smoke. Our clothes were constantly smelling like smoke-it didn't matter if they were washed. You know .... being in the car all closed up with somebody smoking. It was just not good. (Interview 4). The same parent said later, Yes, my mom and my dad (smoked). For a long time. So I didn't like it. Actually when I was a kid---everybody did. All of the adults around me did. You know. So I didn't want to do it. (Interview 9). Another parent stated, "My parents both smoked when I was growing up and I didn't like it. And maybe so that I was absolutely convinced that nobody would smoke in our home." (Interview 19). When asked for reasons why smoking is allowed in their households, parents 123

PAGE 139

responded: "It's an addiction." (Interviews 8 & 20). "I do when I get agitated." (Interview 3) "It's a distraction." (Interview 3) "It feels good." (Interview 1 & 3). "It keeps me from eating." (Interview 3). "It relieves stress." (Interviews 2, 8, & 11 ). "I do when my husband and I fight so I can take a break" (from the fight). (Interview 14) "It's a social activity". (Interviews 7, 11,& 14). One parent stated, "Whether or not we have access to cigarettes is kind of like a good measure of how we are doing financially. If I don't have my nicotine, that's just showing me that I don't have money either." (Interview 3). Four parents who are current smokers identified feeling guilty about the effects of smoke on their children. "I have a hard enough time explaining it to my 7 year-old. It's a very bad habit and Mommy's got to quit and you know-and every day you do it". (Interview 1). "I am sorry I am a junkie you know but anytime I am around a non-smoker I do my damdest to keep smoke away from them ...... and if it bothers them too bad, I'll put it out." (Interview 3). Another parent stated, ''Now if my kids develop a cough or something like that-I' ll tell you-the first thing that I think of is my cigarette habit .......... If he develops a cough, the first thing that I say is that I am not going to smoke around you for a little while." (Interview 8). And regarding a small baby in the home, the same 124

PAGE 140

parent said, "I know that it's not good for the baby. I really feel guilty when he gets sick because I know that I made him sick." (Interview 8). Community Factors Demographic characteristics and more general questions investigated community factors impacting smoking behaviors in homes. The questions designed to obtain this information began with investigation of meanings associated with smoking behaviors. The effects of media as an influence on smoking attitudes and behaviors was identified. In 2 interviews, the pros and cons of media influences were identified as important influences in initiation of smoking in children. Changes in norms around smoking behaviors were also mentioned. Although not further addressed in the survey, the effects of media on household behaviors toward environmental tobacco smoke exposure were identified in four interviews. When I was in high school, you were in if you smoked. And now it's so different. .... Kids are getting to the age where it's not okay to smoke .. .its fmally having some effect on these kids. They're (cigarette companies) not allowed to do all of the advertising. The Joe camel ... and I can't tell you how many young kids smoked Camels. I would guess that 50% of high school kids smoke Camels. Definitely has an impact. (Interview 1 ). Another parent said, Yea, they talk about it a lot. Probably from school. The commercials nowadays. I suppose that is a result of the lawsuits. The kids are exposed to anti-smoking propaganda a lot. (Interview 3). 125

PAGE 141

The concept of respect emerged as a major theme in the interviews. Within the IUIPH framework, this theme reflected both intra-household and community factors. Within the household, it was important because of its effect on actual smoking practices in the household. Eightyfive percent (85%) of interviewees discussed its importance in maintaining a smoke free environment. Friends and/or relatives were considered respectful if they were compliant with wishes of parents in the household and did not smoke in the home or smoked outside. Parents defined respect as it related to behaviors of friends and family when they are in the home. One parent said, My husband's father-! probably would let him (smoke) but he respects us and doesn't. You know he might light up and then say, 'Wait a second, you guys don't smoke in here'. And we might not have really told him anything. (Interview 1 ). Another parent said in talking about a grandparent, "She might ask me, but she really respects my rules" (Interview 7). Another parent stated, My husband's mother smokes. She never smoked in the house while he was growing up and she doesn't smoke in front of us now. She knows that we don't like smoking and she goes outside if she is at our house. She respects us and what we want about smoking. (Interview 12). Those who were smokers identified respectful behaviors toward non-smokers as being important. One stated, .... anytime I am around a non-smoker, I do my darndest to keep the smoke away from them. And if it bothers them too bad, I'll put it out. But I am more aware--like I say I don't like second hand smoke. (Interview 3). 126

PAGE 142

The concept of respect is most closely associated with inputs to household factors as identified in the HHPH framework. As with attitudes and beliefs, respect for household rules and/or policies results in behaviors affecting a smoke free environment. While it most directly affects smoking behaviors in households, its basis lies in the social context and social networks surrounding the household. Within such a context, it can also be associated with macro level factors identified in HHPH framework. The concept of respect has not been identified previously in the literature as an important consideration in environmental tobacco smoke exposure in the household. Findings related to the role of social support in smoking restriction or cessation have been mixed as reported by Fisher, (1997). Poland et al. (1999) explored the interactions between non-smokers and smokers in public places. Factors identified by non-smokers in their willingness to protect themselves against smoke exposure included being bothered by the smoke, their own perceived rights to be free of smoke exposure, length of exposure time, perception of how likely the smoker would comply, desire to avoid confrontation, and relationship (if any) to the smoker. Some also felt more comfortable by using the health of their children to legitimize their claims to a smoke free space. Rights in a public place may be addressed differently than rights in a private household. "Respect" as a factor in maintaining 127

PAGE 143

smoke-free households is a consideration worthy of further investigation. Its role as a factor in the implementation of smoke free households deserves further study. Parents who had knowledge and a previous negative experience with passive smoke exposure expressed the most negative feelings about smoke exposure for themselves or their children. Some could articulate clearly the negative health effects related to smoke exposure; others had less information but demonstrated a definite commitment to maintaining a smoke free environment. Those parents who had knowledge and strong, negative attitudes toward smoking were most emphatic in their opposition to smoke exposure. Conclusions Households with children overwhelmingly identify a desire to do what is in the best interests of their children; from this they identify activities that keep children healthy. Parents want to be the best parents that they can be and their perceptions of what is in the best interests of their children take on various forms and activities. Strong emotions exist surrounding smoke exposure in households. Parents make decisions for their household based on their experiences with smoke exposure as children or on their current smoking status Ex-smokers expressed stronger negative statements about smoking than did current smokers or never-smokers. Emerging concepts from this study are deserving of further study especially as new strategies are developed for reducing smoke exposure in young children. Future 128

PAGE 144

explorations of parental attitudes should further explore the concept of respect in shaping household behaviors around smoke exposure. The influence and respect from others in maintaining a smoke free environment is deserving of further study as it relates to smoking policies in households. As a result, several questions were added to the survey to obtain basic information about smoking practices of friends and relatives who live outside of the home. The effects of media on smoking policies were not further investigated. Harms of smoking in the home included knowledge of illnesses and harms other than health associated with smoke exposure. Smoking and pregnancy also emerged as a theme associated with harms. 129

PAGE 145

CHAPTERS FINDINGS FROM QUANTITATIVE DATA ANALYSIS Overview Survey data analyses were conducted sequentially, beginning with description of variables as articulated in the hypotheses. Socio-demographic characteristics and smoking behaviors of households were assessed and differences based on ethnicity, sample recruitment sites, and presence of home smoking bans were examined. Households with smoking bans were compared to households without smoking bans in demographics, health of children in the household, knowledge of ETS exposure, attitudes related to ETS exposure of children, and actual smoking behaviors. Selected independent variables, identified in the hypotheses, were entered into a logistic regression model to identify predictors of home smoking bans. Bivariate correlations confirmed relationships between smoking policies and cotinine levels in urine, validating self-reports by household members. Logistic regression was used to examine how smoking behavior variables in the household relate to the level of cotinine in urine samples. Two hundred twenty-six (226) surveys were completed; 212 households (94%) completed surveys and provided urine for cotinine testing from a child in the 130'

PAGE 146

household. One household survey was excluded from all analyses because of missing data on smoking bans, resulting in 225 eligible surveys for analysis. One hundred forty surveys (62%) were completed in English and 85 (38%) in Spanish. One hundred sixty-four households (73%) reported a complete smoking ban and 61 households (27%) reported a no home smoking ban. Of the 211 eligible subjects who completed cotinine testing, 8 were excluded from further analysis; one household contained a child who was breastfed with reported smoke exposure and 7 households reported children exposed to tobacco smoke in daycare but not in the home. The household with the breastfed child was excluded from cotinine analysis due to the inability to quantify the effects of breastfeeding on cotinine measurements. Cotinine concentration in breastfeeding infants is influenced by mothers' frequency of breastfeeding, smoking behaviors, and ETS exposure (Matt et. al, 1999) Subjects with cotinine measurements who were exposed to smoke in daycare and not at home were removed from the sample for analysis to avoid possible confounding. Two hundred three (203) eligible households completed surveys and provided urine for cotinine testing for analysis. Ninety-seven ( 48%) were negative for smoke exposure and 1 07 (52%) were positive for exposure. In the preschool/Head Start population, 169 households out of 1051 eligible households completed surveys (16%). Fifty-six households enrolled in SBHC clinics completed surveys. The total number of households in SBHCs with children 0-5 years was not available as data are recorded only by ages of children receiving 131

PAGE 147

services and by households containing children, 0-18 years. No data were available on those who refused to participate in the survey. In the pre-school/ Head Start study population 49% were Hispanic, 39% were not Hispanic White, 4% were African American, and 5% were Asian In the SBHC study population, 53% were Hispanic, 46% were not Hispanic White, 2% were African American and 4% were Asian. There were more Hispanic (62%) and fewer African American (2%) and fewer Asian (1 %) subjects than in the study population. Demographic Characteristics The racial characterization of subjects was as follows: 24% White, not Hispanic, 62% Hispanic, 2% African-American, 1% Asian, 10% mixed, and .4% other (See Table 5.2). Subjects were divided into Hispanic and not Hispanic White categories for investigation of ethnicity variables. One hundred thirty-nine surveys (62%) were completed in English and 86 (32%) were completed in Spanish. Fifty-six households (25%) were recruited from school-based health centers and 169 (75%) from preschool/Head Start sites (See Table 5.1 ). Each household contained a mean of 2.6 adults (SD 1.218, range 1-7) and 2.6 children (SD 1.258, range1-7) resulting in a mean of5.2 residents per household (SD 1.858, range 2-11). Ten percent ofthose completing the surveys were fathers, 90% were mothers One hundred sixty-eight (75%) households reported parents and/or stepparents married or cohabiting and 57 (25%) households reported a single or divorced parent. The mean age of respondents 132

PAGE 148

was 29 years (range18-50 years, SD 6.184). In 92 households (41%), mothers had less than a high school education and in 85 households (43%), fathers had less than a high school education. One hundred forty households (62%) lived in single-family dwellings (house or mobile home) and 85 (38%) lived in multi-family dwellings (apartment, duplex, or condominium). One hundred thirty-three (64%) households had annual household incomes of$30,000 or less; 38 (18%) had incomes between $30,001 and $50,000 and 36 (17%) were greater than $50,000 annually. Eighty-four (84%) percent of households were eligible for low cost insurance or Medicaid, categorized at 125% of poverty or greater (Federal guidelines, 2002), based on household income and the number living in the household. Compared to the demographics of the population from which the sample was taken, Hispanic families were slightly over represented in the study (62% in the study compared to 50% of study population). 133

PAGE 149

Table 5.1 Characteristics by Household Category Characteristic N Percentage Language of English 139 62 Surveys Completed Spanish 86 38 N=225 Recruitment Site Preschool/Head Start 169 75 N=225 School-Based Health Center (SBHC) 56 25 Adults Living in 1-2 156 75 Home 3 or more 69 25 N=225 Children Living in 1-3 169 75 Home By Household Morethan3 56 25 N=225 Age of Children 0-1.9 75 18 Living in Home 2-4.9 195 47 N=413 5-12.9 121 29 13-18 22 5 Age of Respondent 18-24.9 58 26 N=222 25-30.9 86 39 31-39.9 62 28 40-50 16 7 Gender of Male 56 25 Respondent Female 169 75 N=225 Marital Status Married/Living N=225 Together 168 75 Single/Divorced 57 25 Education-Mother
PAGE 150

Table 5.1 Characteristics of (Cont.} Category Characteristic N Percentage Education-Father
PAGE 151

Hispanic White. The mixed/other category (n=31) included all subjects reporting more than two racial categories in the household or two racial categories, when one was Hispanic or not Hispanic White, and the secon,d was Asian, African-American, or other. Because of the small numbers (n=ll), households who could not be categorized as either not Hispanic White or Hispanic were dropped from the analyses involving ethnicity. Table 5.2: Ethnicitv of Subjects Subjects Not Hispanic AfricanWhite Hispanic American Asian Mixed Other Total N % N % N % N % N % N % N Total Subjects 55 24 139 62 5 2 3 1 22 10 1 .4 225 Total Subjects for Cotinine 51 25 127 62 5 2 3 1 16 8 1 .5 203 Analysis Not Hispanic White Hispanic Other/Mixed Total N % N % N % Total Subjects 55 24 139 62 31 14 225 Total Subjects for Cotinine 51 25 127 62 25 13 203 Analysis Not Hispanic White Hispanic Total 62 29 152 71 214 Total Subjects Total Subjects 59 30 137 70 196 for Cotinine Analysis 136

PAGE 152

Ethnic groups differed significantly in household income (p<.OOl); more Hispanic households had incomes less than $30,000 and less Hispanic households had incomes between $30,000-$50,000 and over $50,000 than the non Hispanic White group. Seventy-four percent of Hispanic households had an annual income of less than $30,000 compared to 45% of not Hispanic White households. The number of adults living in the home (p<.OOl), educational levels of mothers/fathers (p<.OOl), and language spoken in the home (p<.OOl) differed significantly (See Table 5.3). There were significant differences in levels of education between mothers and fathers. More Hispanic mothers and fathers (57% and 62% respectively) had less than high school education (p<.OOl) than mothers and fathers in not Hispanic White group. More not Hispanic White mothers (31 %) and fathers (24%) had completed college (p<.OOl) than mothers and fathers in Hispanic households (4% and 3%). 137

PAGE 153

Table 5.3: Demogra};!hic Characteristics and Ethnicity Demographic Category Not Hispanic Total Significance Hispanic White N=l51 N=213 N=62 Value d f. p No. Living in Home By Adults 2.1/hshld 2.8/hshld 2.6/hshld -4.753 190 000 Household Children 2.4/hshld 2.6/hshld 2.5/hshld -1.135 212 .187 N=213 Total 4.5/hshld 5.4/hshld 5.1/hshld -3.786 157 .000 Age of 30.35 yrs. 28 09yrs. 28.74 yrs. 2.508 208 .013 Respondent (SD 6.62) (SD 5 66) N=210 N % N % N % Value d f. p Marital Married! Status Living N=214 Together 52 84 121 80 173 81 .517 1 .472 Single/ divorced 10 16 31 20 41 19 Mother
PAGE 154

Table 5.3: Demogra:ghic Characteristics and Ethnicitt {Cont.) Demographic Category Not Hispanic Total Significance Hispanic Whi t e N=151 N=213 N=62 No. % No. % No. % Value d f. p $50,000 17 28 15 11 32 16 Language English Spoken In Only 59 95 42 28 101 47 Home English/ 80.744 3 .000 N=188 Span. 3 5 34 23 37 17 Spanish Only 0 34 75 50 35 Type of Single Dwelling family 45 73 88 58 133 62 4 038 1 044 N=212 Multi17 27 64 42 81 38 family Number of 3-5 rooms 7 12 40 27 47 23 rooms in 6-9 rooms 39 64 76 52 115 55 6.130 2 047 dwelling 10-13 15 25 31 21 46 22 N=208 rooms a= .001 (Bonferroni adjustment) 139

PAGE 155

Significant differences exist between Hispanic and not Hispanic White groups in income (p<.001) and educational level (<.001). Previous studies by Lund et al. (1998), and Jordaan et al. (1999) report relationships between income and educational level and smoke exposure whereby smoke exposure is greater in households with lower income and lesser education. Demographics by Recruitment Site Seventy-five percent of subjects (n=169) were recruited from preschool/Head Start sites and 25% (n=56) from SBHCs (See Table 5.4). Table 5.4: Sites From Which Subjects Recruited Site All Surveys Surveys for N=225 Cotinine Analysis N=203 N % N % School-based Health Centers Baker 6 3 5 2 Gregory Hill 9 4 8 4 Sheridan Health Services 29 13 26 13 Carin' Clinic 12 5 11 5 Total 56 25 50 25 Preschool/Head Start District 50 58 26 48 24 Sanville Preschool 26 12 22 11 Adams County Head Start 72 32 70 34 Sheridan Head Start 13 6 13 6 Total 169 75 153 75 140

PAGE 156

Significantly more households from preschool/Head Start sites reported owning their residence (p<.OOI) and living in a single family dwelling (p=.OOI) than did families from SBHCs. Differences in levels of income (p=.035) were no longer significant when adjusted for type I error (a.=.OOI). More households recruited from SBHCs had children who were breastfed than households from preschool/Head Start sites (x2::J3.713, d.f.=l. p<.OOI). This was consistent with ages of children tested for cotinine (t= -6.297 d.f.=60.813, p<.OOI) as children in households recruited from school-based health centers were younger than the children from pre-school/Head Start (M=2.15 vs.3.68). There was no significant difference in presence of home smoking bans d.f. = 1, p=.682) or cotinine exposure between the two groups (F-.128, d.f. = 1, p=.720). In this analysis, subjects from preschool sites and from Head Start sites were comparable. 141

PAGE 157

Table 5.5: Demographic Characteristics ofPreschool/Head Start Sites andSBHCs SBHCs Preschool 4 sites 4 sites Total Demographic Category N=56 N=169 N=225 Significance N % N % N % Value d. f. Language of English 31 56 108 61 139 86 Surveys Spanish 25 44 64 36 89 62 .818 1 Completed N=218 Adults No. Living in 1-2 34 60 122 72 156 75 Home By 3 or more 22 40 47 28 69 25 2.11 1 Household Children 1-2 41 74 139 82 169 75 3 or more 15 26 30 18 56 25 2.15 1 Total Hshld. 56 169 225 2.18 2 Age of 0-4.9 54 60 158 60 212 60 4.357 4 Children 5-12 9 30 33 91 34 121 34 5.133 3 Living in 13-18 6 7 16 6 22 6 2.674 2 Home N=355 Mother
PAGE 158

Table 5.5: Comnarison ofDemogranhic Characteristics of Preschool/Head Start Sites and SBHCs (Cont.) SBHCs Preschool Demographic Category 4 sites 4 sites Total Significance N=56 N=l69 N % N % N % Value d. f. p Father < High school 23 23 62 43 85 44 Education High school .873 3 .832 N=196 grad 11 12 30 25 46 24 Some college 14 12 35 21 44 22 College grad or grad. Study 5 6 16 14 21 11 Household $30,000 or less 42 79 91 51 133 64 Income $30,0016.71 2 .035 N=207 50,000 5 9 33 21 38 18 $50,001 or more 6 11 30 20 36 17 Language English Only 24 42 82 49 106 47 Spoken In Eng./Spanish 23 40 53 32 76 34 Home Spanish Only 8 14 30 18 38 17 4.22 4 .376 N=225 English/Other 2 4 3 2 5 2 Racial Hispanic 40 70 100 59 139 62 Categories Not Hispanic N=225 White 12 21 43 25 55 24 2.29 2 .317 Mixed 5 9 26 15 31 14 Number of 3-5 rooms 20 35 31 19 51 23 rooms in 6-9 rooms 26 46 93 57 119 54 6 .51 2 .038 dwelling 10-13 rooms 11 19 39 24 50 23 N=220 *Type of dwelling: Single family = House, Multi-family = Apartment, Duplex, Condominium. a =.00 l using Bonferroni adjustment 143

PAGE 159

Household Characteristics and Smoking Bans Out of225 households reporting, 7 households (3%) identified no home smoking bans (unrestricted smoking) and 54 (24%) reported a partial home smoking ban, for a total of 61 households (27%) reporting partial or no home smoking bans. One hundred sixty-four (73%) reported complete home smoking bans. Table 5.6: Demographic Characteristics and Household Smoking Bans Variable Subject Home Smoking No/Partial Characteristics Ban Home N=164 Smoking Ban Significance N=6l Mean Mean TTest d f. Sig. Living in Adults 2.62 2.49 .686 222 .493 Home SD 1.218 SD 1.299 Children 2.64 2.41 1.243 223 .202 SD 1.258 SD 1.174 Total 5.26 4.90 1.278 222 .203 SD 1.952 SD 1.567 Age of 29.08 28.98 .091 219 .928 Responde SD6.220 SD 6.177 nt Number of 7.54 7 90 -995 217 .321 rooms in SD2.399 SD 2.334 Dwelling Smoking No/Partial Ban Smoking Ban Total N % N % N % Value d.f. p Age of 0-4.9 154 73 58 57 212 60 7.182 4 .127 children 5-12.9 91 75 30 25 121 34 1.781 3 .619 N=355 13-18 12 55 10 45 22 6 .623 2 .732 a= .00 1 Bonferroni adjustment 144

PAGE 160

Table 5.6: DemograRhic Characteristics and Household Smoking Bans (Cont.) Variable Subject Smoking No/Partial Characteristics Ban Smoking Total Ban N % N % N % Value d. p Gender of Male 12 55 10 45 22 10 Respondent Female 152 75 51 25 203 90 4.152 .042 N=225 Marital Married/living Status together 136 75 46 25 182 81 N=225 Single/divorced 28 65 15 35 43 19 1.625 1 .202 Household $30,000 or Income less 91 69 41 31 132 64 N=206 $30,0012.971 2 .226 50,000 28 74 10 26 38 18 >$50,000 30 83 6 17 36 18 Racial Not Hispanic Categories White 33 60 22 40 55 24 N=225 Hispanic 112 81 27 19 139 62 10.89 2 .004 Other/Mixed 19 61 12 39 31 14 Ethnicity Not Hispanic N=214 White 35 56 27 44 62 29 11.94 1 .001 Hispanic 121 78 31 53 152 71 Language English 87 62 53 38 140 62 Surveys Spanish 77 91 8 9 85 38 21.656 1 .000 Completed N=225 Language English Only 65 61 41 39 106 47 Spoken in Spanish only 68 91 7 9 75 33 Home English/Spanish 26 68 12 32 38 13 24.106 1 .000 N=224 English/other 5 3 0 5 2 Type of Single 101 72 39 28 140 62 Dwelling Multifamily 63 74 22 26 85 38 .104 1 .747 N=225 145

PAGE 161

Table 5.6: DemograQhic Characteristics and Household Smoking Bans (Cont.) Variable Subject Smoking No/Partial Characteristics Ban Smoking Total Ban N % N % N % Value d. f. p Mother
PAGE 162

5.6). Of Hispanic households completing surveys, 80% had smoking bans and 20% did not; in not Hispanic White households, 57% had smoking bans and 43% did not. Six measures of smoking behaviors were investigated as methods of identifying smoke exposure in the household. These measures included: smoking status of respondent, last time someone smoked in household, number of smokers living in home, number of cigarettes smoked in home in past week, situations in which smoking was allowed in the household and presence of smoking bans. The dependent variable, home smoking ban, was dichotomous including a complete home smoking ban and a partial or no home smoking ban. One hundred sixty-four (164) current smokers lived in 107 households (M=l.54, SD=.974 Range=5). One hundred sixteen households (52%) reported no smokers living in the home. Twenty-three percent of households identified mothers who were smoking members, 29% identified fathers, 9% identified grandparents, and 8% identified friends. Nine percent of all households reported that both mother and father smoked. Sixty households (27%) included children age 10 and over; four of these households (7%) reported children who smoke (2% of all households surveyed). Sixty-one respondents (28%) identified themselves as current smokers, 32 (15%) as ex-smokers, and 125 (57%) as never smokers. Households with complete home smoking bans had 1.38 smokers/household and households with partial/no home smoking bans had 1.67 smokers per household. In 209 households reporting 147

PAGE 163

when someone last smoked in the home, 74 (35%) reported smoking in home within past month, 13 7 ( 65%) reported no smoking in household for one month or more. In assessing specific situations in which smoking was allowed, 61 families scored from 1 to 5 on the scale (M=2.7, SD. 1.626) indicating smoke exposure in specific situations and the other 164 families reported 0 indicating that there were 0 situations in which they allowed smoking in the home. Thirty-one households (14%) reported that they allowed smoking with a child in the home, 22 (10%) reported that smoking was allowed at home when children were in the room, and 30 (13%) allowed smoking when children were asleep in the home. Forty-seven families (21 %) allowed relatives to smoke in the home and 35 (16%) allowed friends to smoke in the home with children present. Twenty-eight percent (28%) of households indicated that their friends allowed smoking in their own homes. In families reporting a complete smoking ban, 3 families acknowledged that they allowed smoking in the home when children were present, 10 households allowed relatives and 4 households allowed friends to smoke in their homes. The results of this question are of interest compared to the qualitative fmdings where families did not report difficulties in maintaining a smoke free environment when relatives smoked. Of the households who reported situations in the household in which smoking occurred, 54 (24%) allowed friends and relatives to smoke in the home and 38 (17%) allowed smoking with children present. Significantly more households allowed 148

PAGE 164

friends and relatives to smoke in the home than allowed smoking in situations where children were present in the home (p=<.01). Of the one hundred sixty-three households reporting complete smoking bans, twenty-two households (13%) acknowledged making exceptions for smoking in their homes; 141 households (87%) reported that they never made exceptions. Even though only 22 households reported exceptions for smoking in their homes, 113 households identified specific times when they made exceptions for smoking. Fortyone (36%) identified cold weather/rain as a reason, 33 (29%) made exceptions for relatives and 39 (35%) did not feel comfortable telling someone not to smoke in their home. Table 5.7: Smoking Behaviors and Smoking Bans in Households Measu r ement Characteristic Complete Smoking Ban N % Smoking Current Status of Smoker 25 41 Respondent Ex/never 133 85 smoker N=218 Total 158 72 Last Time Within Someone Past Smoked in Month 24 32 Home More Than One N=209 Month 124 92 Ago Total 148 71 Partial/No Smoking Total Significance Ban N % N % Value d. f. 36 59 61 28 42.115 2 24 15 157 72 60 28 218 50 68 74 35 81.654 1 11 8 135 65 61 29 209 149 p .000 .000

PAGE 165

Table 5.7: Smoking Behaviors and Smoking Bans in Households (Cont.) Measure Characteristic Complete Partial/No Smoking Ban Smoking Total Significance Ban No. % No. % No. % Value d. f. p Smoking None 143 91 14 9 157 72 Situation Score 1-5 15 25 46 75 61 28 97.37 .000 Scores N=218 158 73 69 27 218 No. None 97 92 8 7 105 49 Cigarettes 2 pks or Smoke in less 52 61 34 39 86 40 51.832 2 .000 Home in More than Past Week 2pks 6 26 17 74 23 11 N=214 Total 155 72 59 28 214 Number of .51 1.29 T value d. f. Sig. smokers SD .847 SD 1.043 -5.134 85 .000 living in home None 103 89 13 11 116 52 p N=222 1-3 per 28.018 2 .00 household 57 58 42 42 99 45 0 4-6 per household 4 57 3 43 7 3 Total 164 74 58 26 222 Households with complete smoking bans and households with no/partial smoking bans showed significant differences in all measures of smoking behavior. All measures of smoking behaviors correlated significantly with presence of smoking bans and with each other. 150

PAGE 166

Table 5.8: Correlation of Smoking Measures With Presence of Smoking Bans 1 2 3 4 5 6 N=206 N=222 N=213 N=215 N=218 N=222 1. Last Time Someone .474* .565* .473* .552* .625* Smoked in Home 2. No. Smokers Living .474* .765* .598* .269* .339* in Home 3. No. Cigarettes .565* .765* .542* .352* .492* Smoked in Home in Past Week 4. Smoking Status of .473* .598* .542* .387* .440* Respondent 5. Smoking Situations .552* .269* .352* .387* .694* Score of Households 6. Complete Smoking .625* .339** .492* .440* .694* Ban *p<.Ol. **p<.05. Fifty-four (54) households reported allowing smoking in one or more rooms in the house. Of these, 29 (54%) allowed smoking in only one room, 14 (26%) households reported allowing smoking in two rooms, 3 ( 6%) households reported allowing smoking in three rooms, 2 (4%) households reported allowing smoking in four rooms, 6 (11%) households reported smoking in 5 rooms. Of all households reporting (n=203), 25 (12%) reported that smoking was allowed in kitchen, 22 (11 %) in the family/living room, 20 (1 0%) in the bedroom, 19 (9%) in the basement, and 16 (8%) in the bathroom. 151

PAGE 167

Smoking Behaviors and Ethnicity Significant differences between not Hispanic White and Hispanic groups were found in all smoking measures except the last time that someone smoked in the home and exceptions for smoking when a complete ban was reported (See Table 5.8). The Hispanic group reported a complete home smoking ban significantly more than not Hispanic white group (p=.001, 80% versus 57%). More Hispanic households reported no smokers living in the home (p=.047) than the not Hispanic white group (57% versus 40% ), while the not Hispanic White group had a larger number of 1-3 smokers reported living in the home (58% versus 40%). The Hispanic group had significantly more respondents(80% versus 55%) who were ex or never smokers (p<.OO 1 ). The category of "number of cigarettes smoked in the home in the past week"' was reported as "none" more by Hispanics than by the not Hispanic white group (p<.001). Hispanic families had a statistically significant lower average score on situations in which smoking is allowed in home than not Hispanic White families (M=2.43 vs 3.36, p= .025). However, when scores were categorized, there were no significant differences between groups in the numbers of households reporting that they allowed smoking in no situations (Score = 0) and those who did allow smoking in specified situations Score= 1-5). There were no significant differences between groups in the smoking measure, ''the last time that someone smoked in the home" categorized as more or less than one month ago (p=.128). 152

PAGE 168

No significant differences exist between ethnic groups for making exceptions for people smoking in the home, nor for reasons for making exceptions. Table 5.9: Smoking Behaviors and Ethnici:ty Smoking Characteristic White, Not Hispanic Total Significance Behavior Hispanic N % N % N % Value d. f. p Home Complete 35 22 121 78 156 73 11.95 .001 Smoking No/partial 27 47 31 53 58 27 BanN=214 No. None 25 23 85 77 110 52 Smokers 1-3 36 38 59 62 95 45 6.13 2 .047 Living in 4-6 1 17 5 83 6 3 Home N=222 Smoking Ex/Never Status of Smoker 34 23 116 77 150 73 13.77 .000 Respondent Current N=207 smoker 28 49 29 51 57 27 No. None 23 23 77 77 100 49 cigarettes 2 pcks. or smoked in less 20 25 61 75 81 40 28.98 2 .000 past week Morethan2 N=204 pcks. 18 78 5 22 23 11 Number of 0 25 23 85 77 110 52 smokers 1-3 59 62 36 38 95 45 6.133 2 .047 living in Morethan3 1 17 5 83 6 3 home N=211 153

PAGE 169

Table 5.9: Smoking Behaviors and Ethnicity (Cont.) Smoking Characteristic Not Hispanic Total Significance Behavior Hispanic White N % N % N % Value d. f. p Smoking 0 40 26 109 74 149 72 2.445 1 .118 Situation 1-5 22 38 36 62 58 28 Score By Category N=207 Smoking M SD M SD T value Situation Score 1.15 1.863 .56 .56 2.284 83.4 .025 N=207 Exceptions Yes 24 36 43 64 67 31 To Smoking No 38 26 108 74 146 69 2.135 I .144 Ban N=I63 Reasons for Exceptions Weather No 28 25 82 75 110 75 1.290 .256 N=147 Yes 13 35 24 65 37 25 Relatives/ No 29 24 91 76 120 80 Friends Yes 11 37 19 63 30 20 1.918 1 .166 N=150 Not No 34 30 78 70 112 75 Comfortable Yes 7 18 31 82 38 25 2.035 .154 Telling N=150 154

PAGE 170

Discussion of Smoking Measmes Almost half of all households ( 49%) reported smokers living in the household. Twenty-seven percent (27%) of subjects reported that smoking was allowed in the home with or without restrictions, 26% reported smoking had occurred within the home in the past week and 35% reported smoking in the past month. On another measme, 51% of households reported that some cigarettes had been smoked in the home in the past week. Measmements of smoking behavior were consistent with previous studies in which smoking behaviors were studied. The use of situations in which smoking occms in homes has not been widely used as a measme of smoking behavior. Health of Children in the Household Three measmes of reported health of children in household were assessed: the number of minor acute illnesses in the past year of each child in the household, number of children with chronic illnesses, and number of children with asthma. A fomth measme, health of each child in the household, identified 188 (95%) out of 199 reporting households with all healthy children. Out of 474 children in 199 households, 438 (92%) were categorized as healthy and 36 (8%) were not healthy. Thirteen households ( 6%) reported no illnesses among children in the past year; 116 155

PAGE 171

households (54%) reported 1-4 illnesses in the past year; 58 households (27%) reported 5-8 illnesses in the past year; 30 households (14%) reported 10 or more illnesses in the past year. Twenty-seven households (12%) reported at least one child with a chronic illness and 31 households (15%) reported at least one child in the household with asthma. There were no significant differences between households with complete smoking ban and households with no/partial smoking in the reported health of children, report of minor acute illnesses, presence of chronic illnesses, or presence of asthma (See Table 5.1 0). Table 5.10: Childhood Illness and Smoking Bans in Households Health of Children/Household Not healthy Healthy Total Reported Chronic Illnesses of children By Household No illnesses At least one illness Total Children with Asthma By Household No Asthma Asthma Total Complete Partial/no Chi-Square Smoking Smoking Total Significance Ban Ban N % N % Value d. f. 7 64 4 36 11 142 76 46 34 188 .782 149 50 199 134 74 47 26 181 2.550 1 16 59 11 41 27 208 136 75 46 25 182 2.413 19 62 12 38 31 213 156 p .377 .llO .120

PAGE 172

There were no significant correlations between health measures of children and smoking bans (See Table 5.11 ). The presence of a chronic illness and asthma correlated positively, as would be expected as asthma is a chronic illness. Table 5.11: Correlations oflllness and Smoking Bans in Households 2 3 4 5 1. All children's health .116 -.121* -.108 -.063 2. Minor acute illness in .116 .127 .084 .078 household in past year 3. Presence of chronic illness -.121 .127 .451** .111 4. Children with asthma in -.108 .084 .451** .106 household 5. No/ partial Smoking Bans -.063 .078 .111 .106 **Significant at .01. *Significant at .05. No significant differences were found between the Hispanic, not Hispanic White groups for reporting children as healthy or unhealthy (x2::003, d.f. 1, p=.958). More minor acute illnesses in the past year were reported by not Hispanic White households (M=5.8, SD 4.709) than Hispanic households (M=4.49, SD 3.982). This was a significant difference (t-value = 2.046, d. 204, p= .042). The number of households reporting chronic illness (x2=3.704, d.f. 1, p= .058) approached significance and reporting of asthma as a chronic illness (xz=12.415 d.f. 1, p<.001) was significantly higher among not Hispanic Whites than among Hispanic groups. It 157

PAGE 173

was not possible to determine if these differences occurred because not Hispanic White families have better access and therefore, seek care and have a higher rate of identification of illnesses and/or asthma as a result. Knowledge and Attitudes/Beliefs and Smoking Bans The twelve item subscale measured knowledge of environmental tobacco smoke exposure by measuring knowedge of harms of exposure, and knowledge of health effects of exposure. The mean score from 225 surveys was 10.58 (range 3-12). Significant differences in knowledge scores between households with complete home smoking bans and no/partial bans were found. (t-value =2.642, d.f.=223, p=.009, 95% CI .163-1.120 of difference). A low but statistically significant negative correlation (r= -.179) existed between the level of knowledge and a no/partial home smoking ban. Significant differences were found between knowledge scores of Hispanics (M=10.43, SD 1.736) and not Hispanic White (M=11.06, SD 1.158) respondents (t-value=2.615, d.f. = 167.59 p=.002, 95% CI .64-.204). The thirteen-item scale of attitudes and beliefs about smoke exposure measured four concepts: beliefs of parents regarding health protection of children against smoke exposure, attitudes of parents regarding smoke exposure, attitudes of parents regarding smoke exposure of their children, attitudes of parents regarding use of legal means to protect children from smoke exposure. Questions were assigned 158

PAGE 174

values from 1 to 5; !indicated a positive attitude, 3 indicated a neutral attitude and 5 indicated a negative attitude toward smoke exposure. Each survey had a possible point range from 13 to 65, with total scores above 39 indicating increasing negative attitudes toward smoke exposure, and scores below 39 indicating more positive attitudes toward smoke exposure. The mean score per survey was 55.88 (range 29-68, SD 7.850) and per question was 4.31 (SD .99). Table 5.12: Attitudes and Beliefs About Smoke Exposure Code Item Strongly Agree Neutral Disagree Strongly Agree Disagree BELffiFSOFPARENTSABOUT Mean Score (SD) PROTECTION OF CHILDREN Negative Positive attitude AGAINST SMOKE EXPOSURE 5 4 3 2 1 HARMFUL I believe that N=224 smoke 193 28 2 0 1 exposure is 4.84 harmful to 86% 13% 0.9% 0.4% (.455) children. HATE I hate it N=223 when I see 151 42 21 5 4 adults 4.49 smoking 68% 19% 9% 2% 2% (.890) around children. A parent PROTECT should N=224 protect their 182 28 10 3 4.72 child from 81% 13% 5% 0.4% 1% (.694) smoke exposure. 159

PAGE 175

Table 5.12: Attitudes and Beliefs About Smoke ExQosure (Cont.} Code Item Strongly Agree Neutral Disagree Strongly Mean Agree Disagree Score JOB It is my job to keep N=224 my children away 187 27 6 3 4.78 from tobacco 84% 12% 3% 1% 0.4% (.606) smoke. TEACH It is my job to N=224 teach my children about dangers and 200 20 2 1 1 4.86 health effects of 89% 9% 1% 0.4% 0.4% (.467) tobacco smoke. ATTITUDES OF PARENTS TOWARD Positive ---------------------7 Negative attitude SMOKE EXPOSURE 2 3 4 5 DEAL It's not a big deal if 4.19 N=223 adults smoke 16 18 16 31 142 (1.280) around children. 7% 8% 7% 14% 63% It's okay for people OKAY to smoke around N=222 children as long as 19 20 28 44 Ill 3.94 they don't smoke 9% 9% 13% 20% 50% (1.327) around my kids. I don't mind when NOPROB people smoke 28 40 27 41 85 3.52 N=221 around me. 13% 18% 12% 19% 39% (1.467) 160

PAGE 176

Table 5.12: Attitudes and Beliefs About Smoke Exposure (Cont.) Code Item Strongly Agree Neutral Disagree Strongly Mean Agree Disagree Score ATTITUDES OF PARENTS TOWARD SMOKE EXPOSURE OF Negative Positive attitude CHILDREN 5 4 3 2 MAD It makes me mad N=221 when people 156 42 15 4 4 4.55 smoke indoors (.844) around children. 71% 19% 9% 2% 2% Children should NEVER never be exposed 150 30 29 4 11 4.36 N=224 to environmental 67% 13% 13% 2% 5% (1.190) tobacco smoke. ATTITUDES ABOUT LEGAL PROTECTION AGAINST SMOKE Positive attitude EXPOSURE 5 4 3 2 1 We should have LAWS laws against smoke N=221 exposure in home 89 40 40 19 33 3 60 just like in public 40% 18% 18% 9% 15% (1.457) places and work places. Smoking indoors ILLEGAL should be illegal N=224 where children 93 38 43 14 36 3.62 live. 42% 17% 19% 6% 16% (1.472) 161

PAGE 177

Table 5.13: Constructs of Attitudes and Beliefs Theoretical Constructs Items Included Mean Score (SD) BELffiFSOFPARENTSABOUT HARMFUL PROTECTION OF CHILDREN AGAINST HATE 4.74 SMOKE EXPOSURE PROTECT (.622) JOB TEACH ATTITUDESOFPARENTSTOWARD DEAL 3.88 SMOKE EXPOSURE OKAY (1.359) NOPROB ATTITUDES OF PARENTS TOWARD UPSET SMOKE EXPOSURE OF CHILDREN NEVER 4.51 MAD (.958) ATTITUDES ABOUT LEGAL LAWS 3.61 PROTECTION AGAINST SMOKE ILLEGAL (1.46) EXPOSURE Parents expressed stronger attitudes about protection of children against smoke exposure and stronger negative attitudes toward smoke exposure of children than they did about legal protection against smoke exposure and attitudes about smoke exposure. Significant differences in scores of attitudes were found between households with smoking bans (M=58.12) and households with partial/no smoking bans (M=49.18) (t value 6.852, d.f. 84.6, p<.001). Attitudes scores also correlated negatively and significantly with no/partial smoking ban in the household (r= -.365, p =.01). In other words, households expressing more negative attitudes toward smoke exposure were less likely to have no/partial smoking ban. 162

PAGE 178

The Hispanic group (M=57 .34, SD 6.925) reported significantly more negative attitudes based on their scores on the attitude scale than not Hispanic Whites (M=52.68, SD 8.587), (t value=-3.979, d.f. 94.9, p=<.001). The 13-item attitudes/belief scale was utilized as a single scale for logistic regression purposes as a measure of negative attitudes toward smoke exposure of children. Negative attitudes as measured by the attitude/belief scale showed a moderate and statistically significant negative correlation with no/partial home smoking ban (r = -.468) i.e. the more negative attitude toward smoke exposure, the less likely the household was not to have a complete smoking ban. Smoking Bans and Cotinine Exposure Urine samples from 203 subjects, age 2 weeks through age 5 were measured and analyzed for cotinine exposure. Five-year-olds were included only if they were still in pre-school. Ninety-seven ( 48%) tested negative for cotinine exposure and 106 (52%) tested positive. The average age of all children tested was 3.34, and there were no significant differences in ages of children who were negative for exposure (M=3.26) and children who were positive for exposure (M=3.42, p=.381). There was a significant difference in measurements of cotinine in families who had home smoking bans and those who did not (.x2 = 46.16, d.f. = 1, p<.OO 1 ). Significantly fewer households with complete smoking bans tested positive for cotinine exposure than did 163

PAGE 179

households with no/partial smoking bans. Partial/no home smoking bans moderately correlated with levels of cotinine (Kendall's tau_b=.453, p<.001). Table 5.14: Cotinine Measurement of Passive Smoke Exposure No exposure Exposure Negative Positive Cotinine Total cotinine Complete Home Smoking Ban 92 (62%) 56 (38%) 148 (73%) Partial/No Home Smoking Ban 5 (9%) 50 (91%) 55 (27%) Total 97 (48%) 106 (52%) 203 The relationship between reported smoking bans and cotinine exposure was used as a measure to validate self-report of smoking behaviors of households, to reduce potential bias and errors inherent in self-reporting. Parental reports of smoking behaviors are subject to sources of error and bias associated with self-reports especially because of the social desirability not to expose infants and young children to ETS (Matt et al, 1999). In addition to home smoking bans, the last time someone smoked in home, smoking status of respondent, the number of cigarettes smoked in home in past week, and the smoking situation score significantly correlated with cotinine measurements. Conversely, the number of smokers living in the home did not. 164

PAGE 180

Table 5.15: Correlations ofRegorted Smoking Measures and Cotinine Measurements 1 2 3 4 5 6 7 N=203 N=187 N=197 N=202 N=96 N=203 N=203 1. Smoking Ban .628** .488** .454** .187 .701 ** .472** 2. Last time person smoked in home .628** .488** .524** .199* .554** .465** 3 Smoking status .488** .446** .559** .330** .408** .352** 4. Cigarettes Smoked in Home .454** .556** .559** .048 .342** .397** in Past Week 5. Numberof .187 .199* .301 ** .029 .139 .191 smokers in Household 6. Smoking .701** .554** .408** .342** .139 .407** Situations in Home 7 Cotinine level .455** .492** .352** .397** .150 .407** **Significant at .01 level. *Significance at .05 level. Ethnicitv and Cotinine Measurements The difference in cotinine exposure between Hispanic and not Hispanic White households was significant (x2::3.851, d.f. 1, p=.050). Less Hispanic households had measurable cotinine exposure than not Hispanic White households. This is consistent with the significant differences in reported smoking behaviors of the two groups. One smoking measure, the last time that someone smoked in the home (within the 165

PAGE 181

past month versus more than a month ago), showed no significant differences between ethnic groups; all others did. Smoke Exposure in Daycare Smoke exposure in day care was investigated to prevent confounding of measurements of household smoke exposure. Of 57 households who reported children in daycare, 12 households (21 %) reported that children were exposed to smoke there. Of the 12, 7 reported no smoke exposure at home. These 7 cases were removed from data analysis with cotinine measurements to insure that cotinine analysis provided a measure of household exposure only, not confounded by known exposure outside of the household. Cotinine levels of 6 of 7 children from households with smoking bans were negative, regardless of reported smoke exposure in daycare. Smoking behaviors were consistent among the 5 households reporting smoke exposure in daycare and at home. All 5 reported smokers living in home, cigarettes smoked in home in past week, and no/partial smoking ban. Of 7 households with no reported exposure at home, 6 reported smoking bans in the household. Cotinine measurements for 5 families with smoke exposure at home were in the "3" range indicating positive exposure. One household with a smoking ban, no smokers living in the home, and exposure in daycare had a positive cotinine measurement in "2" range. Cotinine measurements for households with home smoking bans and who reported smoke exposure in daycare were negative. This 166

PAGE 182

fmding raises the question of whether or not there was accurate reporting of smoke exposure in daycare. If parents have smoking bans in home with no reported smoke exposure, the question of why they might allow a child to be placed in a daycare environment with smoke exposure must be considered. Table 5.16: Smoke Ex:gosure in Daycare Households with Children in Daycare N=57 Percentage Households with Children Exposed to Smoke in Daycare I2 21 Smoking allowed Yes 5 42 in household No 7 58 Households with smokers living in home 5 42 Cigarettes smoked in home None 6 55 in past week N=ll Less than 2 pcks. 3 27 2-3 pcks I 9 more than 3 packs I 9 Smoking status N=II Current smoker 3 27 Ex/Never smoker 8 73 Cotinine measurement No exposure 7 58 Exposure 5 42 Multivariate AnalysisLogistic Regression Logistic regression was used to investigate the relationship between the presence of home smoking bans and socio-demographic variables; also the 167

PAGE 183

relationship between reported measures of health in children and home smoking bans. Chi-square tests of independence and bivariate correlations tested relationships between the predictor variables and home smoking ban as an outcome. Predictor variables were selected apriori or if they were associated with a p value <.05 in the bivariate analysis. Then odds ratios were conducted for the likelihood of having partial/no home smoking bans. Home smoking ban was coded as complete smoking ban = 0 and partial/no smoking ban = I. Socio-demographic variables, knowledge of smoke exposure, negative attitudes and beliefs regarding smoke exposure, and perceived health of children were examined as they related to home smoking bans. Smoking behaviors were examined as they predicted cotinine measurement in the second model to determine which smoking measures are most predictive of smoke exposure in the household Selection of Variables for Logistic Regression Model Socio-demographic characteristics were selected based their on statistical significance (p = <.OI) with smoking bans, or if these factors had been identified in the literature previously as affecting smoke exposure. Mother and father's level of education (Mannino et al, 200I), household income (Norman et. al, I999), ethnicity (Norman et. al, 1999), language in which survey was completed, number of residents in home (Mannino et al, 200 I), age of respondent parent (Mannino et al, 2001; Norman et. al, 1999), marital status, and the type of dwelling (either single family or I68

PAGE 184

muliti-family) were included as demographic variables Ethnicity and language in which survey was completed were the only statistically significant variables (p<.OOl) identified in this study Knowledge of smoke exposure of the respondent, and attitudes/beliefs of respondent related to smoke exposure were also included based on statistically significant correlations with complete home smoking bans. Household income, and measures of ethnicity ( ethnicity, language in which survey was completed) and educational level could have been proxy measures for the same variable. However, significant differences between households with smoking bans and without bans were found only for ethnicity and language in which surveys were completed. To determine whether these two measured a similar concept of ethnicity, both were entered into the model. All categorical socio-demographic variables were categorized based on the expected predictability before they were entered into the model. Remaining variables were entered as continuous (knowledge, attitude, number of total illnesses in past year, number of children with chronic illnesses in household). Table 5.17: Categorical Variables For Regression Models Variable Categories Frequency Parameter Asthma No asthma 84 .000 Asthma 15 1.000 All children's health Not healthy 3 1.000 Healthy 96 .000 169

PAGE 185

Table 5.17: Categorical Variables For Regression Models (Cont.) Variable Categories Frequency Parameter Marital status Single/divorced 9 1.000 Married/Living with 90 .000 Single/ Multi-family Single family dwelling 61 .000 Housing Multi-family dwelling 38 1.000 Income Level 74 1.000 Above $30,00 25 .000 Mother education High school or below 65 1.000 More than high school 34 .000 Father education High school or below 74 1.000 More than high school 25 .000 Gender Male 13 1.000 Female 91 .000 Ethnicity Hispanic 72 .000 Not Hispanic White 27 1.000 Language in Which Spanish 44 .000 Survey Completed English 55 1.000 All socio-demographic variables were forced into the model together, followed by health characteristics. Because health characteristics were identified in the theoretical model as resulting from community factors (socio-demographic factors) and inputs (knowledge and attitudes), the measures of health status were entered into the model after the socio-demographic and input factors. The health of children living in the home is related to lack of smoke exposure (Ashley & Ferrence, 170

PAGE 186

1998; Cook & Strachan, 1999; Gaffney & Lynch, 2000; Joad, 2000; Mannino et al, 2001) but did not correlate with presence of home smoking bans in this study. Four measures of health were entered into the model as measures of health status following the socio-demographic factors and inputs. The goodness of fit of the model was evaluated using chi-square, and the percent of correct classifications. Socio-demographic variables and knowledge and attitudes (12 variables) were entered simultaneously into the model investigating predictability of no/partial smoking bans (n=124). The initial-2 Log Likelilhood was 116.019 and ended at 65.890. The model predicted 86.9% of correct estimates, that is, it identified smoking bans in households correctly 86.9% of the time. The model chi-square (x2=50.128, d.f.=12, p<.OOI) assessing goodness-of-fit, was significant, indicating that the independent variables are not related to the log odds of the dependent variable. The Hosmer and Lemeshow Test was not significant (x2=6.550, d.f.=8, p=.586), confirming goodness of fit of the model. 171

PAGE 187

Table 5.18: Logistic Regression: Demogra.Qhic Variables and Smoking Bans Variable Estemiated Standard Wald d.f. Sig. Exp (B) 95% C.I. Coefficient Error (OR) (B) (SE) Lower Upper FAMLIV .235 .215 1.193 .275 1.264 .830 1.927 Residents in household DOLD -.048 .067 .505 .477 .953 .836 1.087 Age of Respondent MARCAT -.186 1.162 .026 1 .873 .830 .085 8.102 Marital Status DGENDER 2.949 1.141 6.667 1 .010* 19. 086 2.039 178.689 Gender of Respondent ETHHHLD 105 .781 .018 .893 1.111 .240 5.131 Ethnicity of Household DIN COME .178 .824 .047 .829 1.195 .238 6.014 Household income FORMID 2.985 .967 9.531 1 .002* 19.987 2 974 131.649 Spanish or English FASCHOOL -.507 .882 .331 1 .565 602 .107 3.391 Educational level of father MOSCHOOL -.071 .798 .008 1 .929 932 .195 4.453 Educational level of mother 172

PAGE 188

Table 5.18: Logistic Regression: Demographic Variables and Smoking Bans (Cont). Variable Estimated Standard Wald d. f. Sig. Exp (B) 95%C.I. Coefficient Error (OR) (B) (SE) Lower Upper Score of -.073 .190 .147 .701 .930 640 1.350 Knowledge Items (12) Score of -.181 052 12.388 000* .834 .754 .923 attitude items (13) Constant 7.042 4 068 2 996 I .083 1144.209 Factors Associated With Smoking Bans Three variables were significant in the model: gender of respondent, language in which survey was completed (Spanish or English), and attitudes toward smoke exposure as measured by a 13-item scale. Four health measures were then added to the model simultaneously in a second block. The -2 Log likelihood of this model started at 74.620 and ended at 64.239. The model chi-square was significant (x2=51.780, d.f. 16, p<.OOO) and Hosmer and Lemeshow test was not significant (x2=7.616, d.f. 8, p=.472), both indications that the model was a good fit. None of the health variables were significant in the model, nor did their addition improve the model. The demographic variables, significant in the model, remained so when the health variables were 173

PAGE 189

entered; the predictive ability of the model actually decreased with the addition of the health measures (86.9% to 84.8%). The goodness of fit statistics (model chi-square and Hosmer and Lemeshow Test ) remained unchanged from the first iteration. Only the variables, language in which survey was completed, negative attitudes, and gender were predictive of no/partial bans in the household. (See Table 5.18). Additional analyses were run to further examine the relationships between stated ethnicity and language in which the survey was completed. Stated ethnicity was not a significant predictor in the model even when the language in which the survey was completed variable was removed. Interestingly enough, there was no improvement in the model when the ethnicity variable was removed as measured by the ability of the model to predict complete smoking bans versus no/partial smoking bans. 174

PAGE 190

Table 5.19: All Variables and Smoking Bans in Households Variable S.E. Wald d. f. Sig. Exp (B) 95% C.l. EXP (B) Lower Upper DGENDER 3.025 1.219 6.154 1 .013* 20.585 1.887 224.578 Male= I Female=2 ATTITUDE -.I85 .053 I2.213 I .000* .83I .749 .922 FORMIDCA 2.60I 1.007 6.669 1 .010* 13.473 1.872 96.978 English =I Spanish=2 Variables Which Were Not Significant In Regression Model MARCATE -.OI6 1.233 .000 I .990 .984 .088 11.037 Married/Single DOLD -.073 .074 .989 I .320 .929 .805 1.074 Age FAMLIVE .I78 .223 .634 I .426 1.194 .77I 1.850 Residents in household HOUSECAT -1.668 .843 3.914 1 .048 .I89 .036 .985 Single Mulitfamily INCOME .119 .845 .020 I .888 1.127 .215 5.905 Below $30,000 Above $30,000 FASCHOOL -.539 .954 .3I8 I .573 .584 .090 3.790 Father school 175

PAGE 191

Table 5.19: All Variables and Smoking Bans in Households (Cont.) Variable S.E. Wald d.f. Sig Exp (B) 95% C.l. EXP (B) Lower Upper ETHNHlll-D .088 .852 .011 1 .918 1.092 .206 5.798 White=O Hispanic= I TOTALKNO -.129 .204 .401 1 .527 .879 .590 1.310 Knowledge Score CATEHLTH .547 2.065 .070 1 .791 1.728 .030 98.917 Not healthy=O Healthy= I TTL ILL .112 .094 1.414 1 .234 1.118 .930 1.344 Total illness CHRONILL -.070 .662 .011 .916 .933 .255 3.412 No illness =0 Illness=1 AVERASTH -.464 .995 .217 .641 .629 .089 4.423 No asthma=O Asthma=1 Constant 10.656 5.221 4.165 .041 42436.912 In a final model, only the significant variables were entered from the first analysis followed by the health measures. All remained significant, but the predictiveness of the model did not increase, nor were the health variables significant. Health variables were not significant predictors of partial/no smoking bans, while 176

PAGE 192

attitudes and language in which the survey was completed, and gender of respondent were predictive of no/partial smoking bans. Smoking behaviors as predictors of cotinine levels of children living in the households were entered into a second logistic model. Measurement of cotinine was established as the outcome variable with no cotinine exposure equal to 0, and measurable levels of cotinine equal to I. Through the use of odds ratio and confidence intervals, measures of smoking in the home could be investigated in their ability to predicted actual smoke exposure as measured by cotinine. No/partial home smoking bans correlated with positive cotinine measures of children living in the household (r=. 455). The last time someone smoked in the home, number of cigarettes smoked in the home, smoking status of the respondent, and smoking situation scores also correlated with cotinine measures of exposure (See Table 5.8) and were entered into the model as potential predictors. Correlations between smoking measures were less than .9, the value at which mulitcollinearity is considered substantial (Hair et al, 1998). All smoking measures were entered into the. model as predictors simultaneously since there were no specific hypotheses about their order of importance (Tabichnik & Fidell, 2001). 177

PAGE 193

Table 5.20: Smoking Behaviors And Cotinine Levels Variable B S.E. Wald d. f. Sig Exp 95.0% C.I. (B) EXP(B) Lower Upper SMOKHOM .861 .509 2 .867 1 .090 2.366 .873 6.413 Last Time Someone Smoked in Home SMRCAT -.363 .795 .208 1 .649 .696 .146 3.308 No. Smokers in Home CIGCATE -.383 .792 .233 1 .629 .682 .144 3.222 No. cigarettes smoked in past week STATCATE .028 .530 .003 1 .958 1.028 .364 2.907 Smoking status of respondent TOTSITUA .452 .316 2.048 1 .152 1.572 .846 2.919 Situation score SMOKBAN -1.497 .711 4.432 1 .035* .224 .056 .902 Complete ban No/partial ban Constant 1.240 .777 2.547 1 .111 3.456 Although measures of smoking correlated significantly with smoking bans and with cotinine analysis in bivariate analysis, only the smoking ban remained significant in logistic regression. As indicated by odds ratio (OR=.224), households with no smoking bans were only slightly more likely to have positive level of cotinine exposure than those with complete bans. These data support the use of smoking bans as a measure of smoke exposure in households. 178

PAGE 194

Hypothesis Testing Results of data summary and analysis demonstrated that the first hypothesis was partially rejected, the second and third hypotheses were not supported, and the fourth and fifth hypotheses were supported. 1. Households with complete smoking bans will differ in demographic characteristics than households with no smoking bans. When comparing groups regarding smoking bans and no/partial smoking bans using chi-square, significant differences were found in the following variables: gender (p=. 042), ethnicity (p=. 000), language in which the survey was completed (p=. 000) and language spoken in the home (p=. 000). In regression analysis, two variables, gender and language in which survey was completed, were predictive of no/partial home smoking bans. Ethnicity was not a significant variable in the regression model, although the language in which the survey was completed was. Males were twenty times more likely than females (OR= 20.585) to report no/partial smoking bans. This must be interpreted with caution, as the actual n of males completing surveys was low. Those completing the survey in English were ten times more likely to have no/partial home smoking bans than those completing the survey in Spanish (OR=10.4). 2. Children in households with complete smoking bans will have less reported health effects related to smoke exposure than children in households with no smoking bans. This hypothesis was not supported. Four variables measuring 179

PAGE 195

children's health in households showed no significant differences between groups with smoking bans and those without in bivariate analysis (health of children in household, p=.377; illnesses in children in past year, p=.491; children with chronic illness in home, p=.llO; children with asthma in household, p=.l20); health measures in children did not correlate significantly with no/partial smoking bans in households (health of children r=-.061; illnesses in children in past year, r=.078; presence of chronic illnesses, r=.lll; presence of asthma, r=.l06). Health measures were not associated smoking bans in the regression analysis (health of children, sig. =.791; total illnesses in household in past year, sig. =.234; chronic illness, sig.=.916; asthma, sig=.641); in fact, the predictability of the model was reduced (86.9% to 84.8%) when health measures were entered. 3. Households with complete home smoking bans will have greater knowledge of effects of smoke exposure than households with no smoking bans. This hypothesis was not supported in the study. The level of knowledge as measured in the survey between groups with and without household smoking bans was significantly different (p= .009) in bivariate analysis, but knowledge was not a significant predictor of households with no/partial home smoking bans. The level of knowledge as measured by the knowledge items in the survey was high with the mean percent score-88%. 4. Households with complete home smoking bans will have greater negative attitudes toward smoke exposure than households with no smoking bans. 180

PAGE 196

Attitudes scores predicted home smoking bans; the hypothesis was supported. The mean score on the attitude items in the survey was 55.88, higher than the neutral score of 39. The mean score for each question on the survey was 4.31, when the neutral score for each question was 3. The higher attitude score, indicative of more negative attitudes toward smoke exposure correlated negatively with no/partial smoking bans in households (r=-.468). Less negative attitudes as a measure was a significant predictor of no/partial smoking bans in households (sig. =.000). 5. Households reporting complete home smoking bans will have less smoke exposure as measured by cotinine levels than households reporting no home smoking bans. Households with complete smoking bans had significantly lower cotinine measurements of urine samples than those with no/partial smoking bans (r=. 472). Of the four other smoking measures (last time someoiJ,e smoked in home (r=.628), number of cigarettes smoked in home in past week (r=.454), number of smokers living in home (r=.187), smoking status of respondent (r=.488), smoking situation (r=. 701 ), four correlated with complete home smoking bans and only one did not (the number of smokers in the home). This hypothesis was supported. 181

PAGE 197

CHAPTER6 DISCUSSION AND CONCLUSIONS Overview This research explored household characteristics and smoking policies in households, how they were related and how they were enforced, and the relationship of smoking policies and actual smoke exposure through cotinine levels. The bivariate and multivariate analyses assisted in testing the hypotheses and results are discussed as they relate to hypotheses, other fmdings, and relevance for future research. While the original purpose of the study involved the examination of factors in a population without specific regard to ethnicity, its importance emerged through the course of the and results are discussed in that context. Discussion of Findings: Hypotheses Socio-demographic characteristics were not significantly different in households in this study as indicated by the first hypothesis. When households with complete smoking bans were compared to those without bans, only ethnicity characteristics were significantly different. The stated ethnicity of the household (p<.OOl), language spoken in the household 182

PAGE 198

(p<.001), and the language in which the survey was completed (p<.001) differed between those with smoking bans and those without. More Hispanic households (78%) reported complete smoking bans than not Hispanic Whites (53%); yet, they were oflower income than not Hispanic White households. No income differences were detected in groups with and without smoking bans when all subjects (p=.202) were compared. Only language in which the survey was completed was a significant predictor of smoking bans (sig. = .01 0). Stated ethnicity was not a predictor of smoking bans. Sixty-two percent of families identified their ethnicity as Hispanic and 38% completed the surveys in Spanish. Differences in acculturation may exist between these two groups, not measured in this survey. Income levels were not related to the presence or absence of smoking bans among all subjects of this study; Hispanic families had significantly less annual income than did not Hispanic White groups. Households who self identified as Hispanic and those who completed the survey in Spanish reported significantly more complete smoking bans at all income categories than did not Hispanic White households. Within group comparisons of smoking bans revealed no differences in Hispanic households based on income (x2=.360 d.f. 2, p=.835). Not Hispanic White households reported more smoking bans only at income levels above $30,000; those households below $30,000 reported fewer complete smoking bans than no bans 183

PAGE 199

(x2=.14.368, d.f. 2. p<.OOI). The reporting of smoking bans in subjects completing surveys in English and Spanish were consistent with smoking bans based on ethnicity. Ethnicity in this study, when measured by language in which the survey was completed, was predictive of smoking bans. Several measures of ethnicity were used; there may be variations in acculturation, not captured by the measures used in this study, which impact smoking bans and smoking behaviors. The results may also have been impacted by collinearity of ethnicity variables. The second hypothesis addressed the differences in health among children in homes with smoking bans and those without. No differences were found in reported health of children in households with smoking bans and those without. These data were derived from recall of parents regarding children's health and may have been subject to recall bias. Measures of health also did not correlate with cotinine measurements. Reports of chronic illness (p=.ll 0) and asthma (p=.120) approached significance among those with no/partial smoking bans in the households and might be significant in a study with a larger n. However, these two measures were not predictive of no/partial smoking bans in the logistic regression (sig. =.916 and .641) analysis. The fmdings associated with this hypothesis must be interpreted considering the levels of measurement of the health variables. The questions, 184

PAGE 200

designed to be answered by subjects with relatively low literacy levels, were summarized categorically and analyzed using non-parametric tests, resulting in less powerful analysis. The third hypothesis, stating that the levels of knowledge of households with complete smoking bans would be greater than those with no smoking bans, was partially supported in this study. The level of knowledge measured by the 12 items in the survey was high (mean=l0.58, range of3-12, 88% correct); there was a significant difference in the mean score (t-value =2.642, d. 223, p= .009) between households with complete smoking bans (M=10.76, SD=l.636) and those with no bans (M=lO.ll, SD=l.572). There was also a low negative, but significant correlation (r=-.179) between level of knowledge and no/partial smoking ban in the home in bivariate analysis, but it was not a significant predictor (p = .752, OR= .950 {CI-.694-1.302}) of no smoking bans. While significant differences were found in scores between households with bans and those without bans, all scores were high, making interpretation of those differences more difficult. The limitations of the knowledge scale as evidenced by reliability analysis (KR 20=.6003) also may have impacted results. With a consistently high level of knowledge as occurred in these findings, differences between groups may disappear or become less evident. Knowledge did not remain a predictor of smoking bans 185

PAGE 201

in multivariate analysis. Consequently, the hypothesis was only partially supported. The fourth hypothesis, related to attitude scores of households was supported by results of bivariate analysis and attitudes scores were predictors of smoking bans. Attitude scores of households with complete smoking bans were significantly higher (higher score indicates a more negative attitude) than households with no home smoking bans (M=58.12 versus 49.18, p< .001). The attitude scores also correlated negatively with no home smoking bans (r= -.468) and were significant predictors of no smoking bans in logistic regression (p = .000, OR=.833 {CI-.762-.910}). The attitude measure was one of three variables predicting no/partial smoking bans in households. The last hypothesis stated that households with complete home smoking bans would have less smoke exposure as measured by cotinine levels than those with partial/no smoking ban. Ninety-seven samples ( 48%) tested positive for cotinine and 106 (52%) were negative. Positive smoke exposure as measured by cotinine in urine samples correlated positively and significantly with homes with no smoking bans (r=.472). More households reported complete smoking bans than did households who tested negative for smoke exposure through cotinine. No smoking bans correlated with other measures of smoking behaviors in bivariate analysis (last time someone smoked in home, r= .625; number of smokers living in the home, r= .339; 186

PAGE 202

number of cigarettes smoked in past r= .492; smoking status of respondent, r= .440; and score of smoking situations in households, r=.694). Measurement of smoke exposure through cotinine levels of urine validated self-reports of household smoking bans. Complete smoking bans were the only measure of smoking behaviors that significantly predicted cotinine levels (p = .035, OR= .224, Cl = .056.902). Because smoking bans correlated only moderately with other household smoking measures (see above), all were entered into the regression model as independent variables, and only smoking bans was predictive. The use of home smoking ban as a measure of association with smoke exposure in households is an important finding in this study. Of the smoking measures in the survey, it was most predictive of smoke exposure as measured by positive cotinine levels. Smoke exposure in homes is assessed using a variety of measures in the literature, some of which have been confirmed by actual measures of smoke exposure and some of which have not. This is the first study that demonstrates no smoking bans as predictive of actual smoke exposure in households. Findings Compared to Other Studies In the semi-structured interviews, the "respect" of their wishes and rules were identified by families as important in keeping their homes smoke 187

PAGE 203

free. Both Hispanic and not Hispanic White families discussed the importance of friends and relatives respecting their rules about smoke exposure and not smoking in their homes. Gilpin et al (1999), in analyzing the California Tobacco Survey (1996), found that ''family preference for not smoking" was associated with complete home smoking bans in homes where smokers lived. Perez-Stable, Marin, & Posner, (1998) investigated attitudes and beliefs about cigarette smoking in Latino and non Latino White populations; they found that Latinos cite reasons to quit smoking that reflect concerns about families and interpersonal relationships identified in the cultural dimension ofjamialismo (the importance of the family and collective well-being) and simpatia, a Latino cultural script that emphasizes minimizing interpersonal conflict in families. Factors identified by Perez-Stable et al. (1998) that related to famialismo included effect on others, damaging children's health, damaging others' health, and being a good example for children. Reasons for quitting smoking in subcategories of simpatia included smell in hair and clothes, bad breath, get more wrinkles, burning clothes. In their study, factors infamialismo and simpatia categories were identified significantly more by Latinos than Whites suggesting that cultural influences affect attitudes and beliefs related to smoking. Factors of cost, concerns for future health, and damage to own health were identified similarly by Latinos and Whites in Perez-Stable's study. 188.

PAGE 204

In qualitative interviews in this study, families also identified bad smell, cost, health of kids as reasons not to smoke. Reasons to smoke included stress relief, feels good, occupies time, distraction, and to prevent eating. Similarly, in both studies, cost factors were given as reasons for not smoking and prevent eating and/or not to gain weight were identified as reasons to continue smoking. These findings suggest that factors associated with family considerations for not smoking influence smoking bans in both not Hispanic White and Hispanic households; they are expressed differently and perhaps arise from different cultural beliefs and both should be considered when developing culturally sensitive strategies for dealing with smoke exposure in families. In this study, 174 households (73%) reported that they had no smoking in the household, with 67 of 163 households (31%) making exceptions to the ban. Fifty-four households (24%) allowed smoking with restrictions and 7 (3%) households reported unrestricted smoking making a total of61 households (27%) with partial or no restrictions on smoking. The prevalence data are consistent with previous studies done in this country and lower than studies in Europe, Australia, and Hong Kong. Noted exceptions to smoking bans are greater in this study. Gilpin et al, (1999) reported that 38% of people surveyed in California in 1996 had a complete smoking ban, 26% had a partial smoking ban, and 36% had no restrictions. In a different statewide survey in 189

PAGE 205

California, Norman et al, (1999) reported 76% with a home smoking ban, 10% smoking with restrictions, and 14% unrestricted smoking. Wakefield et al, (2000) found 40% homes in an urban community in Australia had a total smoking ban with 16.5% making exceptions to the ban; 16.8% partial restrictions and 26.5% had unrestricted smoking. Okah et al, (2002) reported 38.2% of households with smokers in an inner city neighborhood in Kansas City, Missouri had a complete smoking ban. In Australia, Daly et. al, (2001) found that 38% of households contained current smokers; Lam et al, (2001) reported that 41.2% households allowed smoking in Hong Kong. Crone et al, (200 1) in the Netherlands reported a 44% rate of smoking in households. Only prevalence rates are reported, without investigation of relationships with other variables. No significant differences were found between groups with or without smoking bans in educational levels, income levels, age of parents, or marital status in this study. Significant differences in income (p<.OOl) and educational levels (p<.001) existed between the ethnic groups in this study, but not between households with smoking bans and those without. Jordaan et al, (1999) found that the number of residents in the home predicted smoke exposure, measured by cotinine levels. Whitlock at al (1998) found an inverse relationship between socioeconomic status and reported smoke exposure with a greater association between educational level of parents and cotinine levels 190

PAGE 206

than between median income and cotinine. Erickson & Bruusgard (1995) found that older parents with higher education (long education) were less likely to smoke Both Erikson & Bruusgard (1995) and Lund et al, (1998) found that parents in Norway were more likely to expose their children to ETS if they were young and living without a spouse or cohabite. While no differences in socio-demographic characteristics were found between subjects in this study with smoking bans and those without, there were significant differences in ethnic groups in socio-demographic characteristics and in reported smoking bans. Hispanics had significantly less education (p< .001) and lower income (p=<.001) than not Hispanic White subjects in this study. However, they reported a significantly greater number of complete home smoking bans (p< .001),less cigarettes smoked in the house in the past week (p= .001), and more households with no-smokers living within (p<.001) than not Hispanic White households. There were no significant differences in exceptions made for smoking between the groups when there were complete home smoking bans; however, differences in cotinine levels based on ethnicity approached significance (r=3.851, d.f. 1, p=.05) with Hispanic households demonstrating less smoke exposure as measured by cotinine than not Hispanic Whites. The differences found between not Hispanic White and Hispanic households suggest that further investigation into household practices based on cultural and ethnic practices 191

PAGE 207

should be considered. Hispanic families, even though they were of significantly lower income than not Hispanic White families reported more complete home smoking bans, had more negative attitudes toward smoking, lower knowledge and had significantly less smoking behaviors in all smoking measures except the last time that someone smoked in the home (p=.336). Previous studies (Gilpin et al, 1996) have identified differences in prevalence of smoking practices in Hispanics and not Hispanic White groups, but none to date have confirmed these with documented differences in cotinine measurements. Smoking behaviors including number of cigarettes smoked in past week, smoking status of respondent, last time someone smoked in the home, smoking situations correlated with presence of smoking bans. Gilpin et al, (1999) found that smokers with total smoking bans in homes had higher levels of quitting behaviors than smokers with partial bans or smokers with no bans. Knowledge about smoke exposure was not a predictor of smoking bans in this study. The knowledge level of subjects in this study was high as evidenced by scores (M=10.96, 88% correct). Although measures used for assessment of knowledge in the survey had relatively low reliability (KR 20=.6003), previous studies reveal similar findings. In a qualitative investigation consisting of interviews with pregnant women, Oakley (1993) found that all were aware of the medical message about the health effects of 192

PAGE 208

smoking, but associated smoking with stressful life events such as relationship problems, financial difficulties, and violence. Goldstein (1994) found that although 90% of parents in a study demonstrated an awareness of dangers of smoke exposure, only 24% actually implemented activities to prevent smoke exposure of their children. Brownson et al. (1992) identified that 87.2% of a population studied were aware of the dangers of smoke exposure for young children. Sorum and Bruusgard (1996) and Greenburg et al, (1994) demonstrated limited success in changing behaviors when information based interventions to reduce smoke exposure of children were implemented. Arborelius et al, (2000) investigated interventions useful in preventing smoke exposure of young children and concluded that interventions need to focus on smoke-free environments for children, not on smoking cessation of parents. They found no effects for interventions focusing on providing factual information about dangers of ETS exposure, but did fmd demonstrable effects for interventions geared to behavioral strategies that considered parental beliefs and effects on children. The previous research of Prochaska et al. ( 1997) on stages of change further supports that knowledge alone is not sufficient for behavior change in the smoke exposure. The attitude scale was a predictor of no smoking bans in households in this study. Based on theoretical constructs, attitudes of parents toward health protection of children, attitudes toward smoke exposure of children, beliefs 193

PAGE 209

about legal issues, and negative attitudes toward smoke exposure were measured. Significant differences in negative attitudes were found between households with smoking bans and those without. Negative attitudes were one of three variables found to be a significant predictor of no smoking bans. Fearn ow (1998) found that health dangers of smoking were highly correlated with parental activism related to smoking; health beliefs about smoking and parental values related to smoking were only marginally significant. Parents with negative attitudes toward smoking demonstrated stronger relationships between values and actions than those with less negative attitudes. Gilpin et al. (1999) found that beliefs in the harmfulness of smoke exposure ofbabies and children were associated with a report of smoking restrictions in the home. Gilpin et al. (1999) also found that family preference for not smoking was associated with smoking bans in the home when smokers lived within. Health measures in this study showed no significant relationships with smoking bans; measures of health of children, number of minor acute illnesses, chronic illnesses, or asthma and smoking bans in households did not correlate with nor were they predictors of no/partial household smoking bans. Previous research has repeatedly demonstrated relationships between minor acute illness such as otitis media, respiratory illnesses including asthma, and smoke exposure in households (Cook & Strachan, 1999). The lack of significance in this study may have been related to the level of measurement 194

PAGE 210

and the use of recall for collection of these data or the relatively small number of subjects. To facilitate obtaining the information from parents, the questions related to health were only retrievable in categorical variables providing a less precise method of measurement, resulting in less powerful analysis because of decreased specificity and variability. In conclusion, demographic findings such as income level, educational level, and age of parents associated with smoking behaviors in previous studies were not demonstrated in this study. Prevalence of smoking bans were slightly higher in this study than those previously reported in US and in Europe; relationships between knowledge and smoking bans were similar to previously reported studies. Attitudes were a predictor of smoking bans, consistent with what has been reported. The importance of family preferences in home smoking bans has been reported and consideration of family unity has been identified as important in Hispanic families. Comparisons of findings in this study with previous findings are limited by the variation in reported measures of smoke exposure; smoking status of parents, proximity of children to someone smoking, and cotinine measurement of exposure are all measures that have been used. The findings in this study are consistent with previous studies when reported smoking measures are compared. Only a few studies have validated results of reported smoking behaviors with actual measurement of smoke exposure through a 195

PAGE 211

measure such as cotinine; all of these used smoking measures other than smoking bans. Strengths and Limitations of the Study The use of more than one method in this study design helped ensure the identification and measurement of appropriate variables for measurement and analysis. This mixed method approach has not been used before to examine the behavioral factors associated with smoke exposure. Qualitative methods confirmed the appropriateness of variables for use in the survey that were identified in the literature. From the semi-structured interviews, new perspectives emerged related to "respect" of household rules and family preferences about smoking practices that have been previously identified. Smoking situations in the household (situations in which smoking was allowed related to presence of children, relatives, and friends) were assessed and discovered to be a significant measure of smoking behaviors. The use of reliability analysis and factor analysis in evaluating knowledge and attitude measures helped ensure the validity of the survey instrument designed to test the hypotheses. Finally, validation of reports of smoking bans through measurements of cotinine levels increased the rigor of the study. Conducting only one interview with families on a sensitive topic such as smoking and smoke exposure meant that there was little opportunity for 196

PAGE 212

exhaustive description of the meaning of this health behavior with families. While these one-shot interviews did provide basic information from which more quantitative questions could be developed, they provide a limited amount of information on how households perceive and deal with smoke exposure, especially as it relates to children. Parents, in their desire to demonstrate their capability to care for their children, attempt to present themselves and their families in the best light possible, which can interfere with the accuracy of their reports. Several interviews conducted over time would have provided richer description and greater understanding of family dynamics when addressing smoke exposure. Distinguishing the dual roles of the researcher, that of being a health care provider for some of the families and a researcher, was necessary in this study. In her role as health care provider, the researcher gained access to families, especially those with whom qualitative interviews were conducted. However, the familiarity with the researcher as a previous health care provider created the potential for bias in the interviews. The researcher attempted to differentiate those roles and repeatedly interpreted the differences in her role as a researcher and a health care provider. She assumed that respondents would be honest and forthright in the dialogue, in spite of the previous history with families for whom she was the primary health care provider. Methodological considerations the quantitative phase of the study 197

PAGE 213

limited the generalizability of the findings. The cross sectional nature of the design did not allow for conclusions regarding causality to be drawn, only assessments of associations between variables, and differences which may exist between households with home smoking bans and those without. The sample was small, limiting the power of analysis and generalizability of findings. Data were collected by the researcher at preschool/Head Start sites and school-based health centers. All families who came to the sites were invited to be in the study and not all chose to participate. No data are available on those who did not participate; of interest would be to know if they had smoking in their household or if they were smokers themselves and if those factors influenced their decision not to participate. All families received a flyer informing them of the study, but not all came to the sites on days when the researcher was there. Selection bias was an issue about which little data are available. Based on total enrollments, only a small percentage of eligible families participated, and they were selected as a convenience sample. Consequently, we cannot investigate differences that might have influenced findings nor do we have a well-defined study population to which we can generalize findings. Although 226 surveys were collected, statistical analyses relied on a smaller sample of households with no/partial smoking bans (n=61), limiting 198

PAGE 214

generalizability. While the number of subjects from whom urine was obtained for cotinine was adequate (n=203) for analysis, the actual n upon which primary analysis for cotinine was conducted was much less (n=I06). The validation of smoke exposure through measurement of cotinine levels was possible in this study, largely because of the sample size and increased the validity ofthe fmdings in spite ofthe limited number of subjects. A validated instrument was not used; several measures were instituted to increase validity and reliability. Questions from validated surveys, expert review and pilot testing of the instrument, factor analysis and reliability analysis were used to increase both validity and reliability of the instrument. The reliability of the group of knowledge questions (KR 20 = .60) was lower than is recommended as a measure in surveys. Consequently, the reliability of the measure of knowledge as a factor in smoke exposure was compromised, making interpretation difficult. The outcomes of the knowledge measurement were consistent with measures in previous studies. The cotinine measurement used in this study was a screening tool with 85% sensitivity and 100% specificity. It was selected because it was inexpensive, easy to use in the settings in which data collection was done, and is considered a valid measure that uses body fluids (urine) easily attainable from subjects. The testing for cotinine on site and the ability to provide immediate results to families was not an original purpose of this study, but 199

PAGE 215

was important in the recruitment of subjects. It was noted anecdotally by the researcher that this is a potentially effective method of providing information to families about the effects of passive smoke exposure on children. The use of a screening measure of cotinine levels as validation for self-reports was appropriate due to the exploratory nature of the study. As validation for self reports of parents, it was important that it have high specificity to prevent the possibility of false positives. The lower sensitivity increases the chances for false negatives and no back-up testing was done to confirm the results of false negative readings. This potentially reduces the accuracy of cotinine testing and therefore, the validity of interpretation of the cotinine measures in confirming self-reports of smoke exposure. This study only provided limited control of a possible confounder, sources of smoke exposure outside of the household. The survey obtained data about smoke exposure in daycare, but did not investigate other possible sources of smoke exposure, such as possible exposure in multi-family dwellings. Several subjects anecdotally indicated potential exposure from this source, but it was not measured in the study. Cotinine measurements may have been altered by this factor. While most families with smoking bans and whose children were in daycare chose smoke free environments for their children, seven did not. Six of 7 households who reported smoke exposure in daycare had complete home 200

PAGE 216

smoking bans. Further investigation of this influence on ETS exposure is warranted. A large number of socio-demographic variables, each containing several categories, were investigated in the survey questionnaire. Because of the multiple comparisons of these variables, there exists a greater chance of an inflated Type I error in which variables would be identified as significant when they are not. In this study, the chances for rejecting the hypothesis addressing the differences between households with smoking bans and those without is greater because of the multiple comparisons of variables entered therein. Based on a bonferroni adjustment, alpha was set at .001 for socio demographic variables to reduce the chance for such an error. This adjustment resulted in the identification of fewer significant characteristics related to smoking bans or smoking behaviors than in previous studies. Findings were interpreted accordingly. In future studies, attempts to determine core factors underlying the outcomes of the socio-demographic variables explored in this study would be appropriate. An alternate approach could include selection of specific variables as measures of characteristics to be studied to avoid multiple comparisons. Multiple measures of similar variables increased the possibility of collinearity among variables which influences predictability of variables in multiple regression. Whenever possible, correlations between measures were 201

PAGE 217

calculated prior to multivariate analysis and if r was .8 or above, the measure was not entered into the logistic regression model. In the case of the six measures of smoking in households, correlations between measures were significant but remained below .8 and all were entered as predictors in the model. In bivariate analysis, the language in which survey was completed and the language spoken in the home were consistently different in groups with smoking bans and those without Although, each may have been a proxy for ethnicity, both entered into the regression model to investigate possible differences in predictability. Only the language in which the survey was completed remained a significant predictor. With this finding, the question as to whether these two measures were assessing different concepts must be addressed. The results of this study suggest that Hispanic households who speak and communicate in Spanish may differ from those households who are proficient enough in English to complete a survey. Further study is needed to identify which of these measures is a better measure of ethnicity. The types of measurement used for health indicators may have impacted the results as well. General categorical variables were used to measure this indicator and may not have provided precise data to obtain valid results. The use of survey questions providing metric variables would have facilitated multivariate analysis. 202

PAGE 218

The exploratory nature of this study provided opportunity to identify new characteristics and assess previously identified ones that may impact smoke exposure of children in households. However, the broad investigation of several groups of characteristics increased the chances for Type I error and collinearity among measures, limiting interpretation of findings. With this in mind, precise measurement of these factors and of measures of cotinine should be considered in future research. Future Research Questions Not all children in households with no/partial smoking bans had measurable smoke exposure and some children in homes reporting complete bans tested positive for smoke exposure in this study. Research that investigates the relative merits of different measures of smoke exposure in household environments is needed. Differences in household characteristics impacting smoke exposure based on ethnicity also call for further investigation. Smoking bans are increasingly used as a measure and/or predictor of smoke exposure in households. While the results of this study confirm the use of smoking bans both as a measure and a predictor of smoke exposure, replication with other ethnic groups and groups of different socioeconomic status will increase the generalizability of the findings. 203

PAGE 219

Further research to improve the instruments for measurement of smoke exposure is needed. Continued reliability and validity testing for survey items will insure improved measurement of behavioral factors. Increased sensitivity of cotinine screening tools is needed if these are to be used effectively as a measure of exposure. The development of cheap, valid, and reliable cotinine measurements that provide for control of dilution of urine in testing (creatinine to cotinine ratios) that could be used in data collection settings is needed for confirmation of parental self-reports. Great potential exists for using the results of cotinine testing in providing households with feedback regarding smoke exposure; this is only possible with specific and sensitive measures, or if back-up testing is available for those results which may be inaccurate. Finally, more research is needed to determine how data obtained from qualitative and quantitative methods can be used in the development of strategies resulting in reduced passive smoke exposure of children as a method of enhancing their health at present and in the future. In addition to the measurement issues associated with cotinine testing, the question of whether or not the knowledge of cotinine levels makes a difference in smoking bans and smoking behaviors in households needs to be addressed. Does knowledge of cotinine levels make a difference in household behaviors related to smoke exposure? Additionally, do cotinine level measurements make a difference in an intervention to reduce smoke exposure 204

PAGE 220

for children? Answers to both of these questions are needed to determine the usefulness of cotinine measurement in strategies addressing smoke exposure with young children. The relationship between knowledge and behaviors of adults in households related to smoke exposure presents some interesting paradoxes that need further study. Most studies support a general awareness among adults of the health effects of smoke exposure; most conclude that knowledge has little to do with what actually happens related to smoke exposure. Oakley (1993) suggests that smoking behaviors may be one manifestation of deeper problems embedded in health care and political systems. These complex issues need to be understood and disentangled if long-term solutions to smoke exposure are to be realized. The measurement of negative attitudes was a strong predictor of no smoking bans in this study. The concepts investigated within this construct have not been studied previously, especially those related to beliefs related to legal issues surrounding environmental tobacco smoke exposure. While internal consistency measures supported the use of this scale, testing with other populations will increase its validity and reliability. The qualitative findings in this study also suggested that previous experience with smoke exposure in adults influenced their decisions to protect themselves and their children from smoke exposure. The negative expressions of these experiences 205

PAGE 221

need to be further investigated; of interest, also, are the strong negative emotions expressed by ex -smokers relating to smoking behaviors and smoke exposure. Both of these may need to be considered in strategies for smoke reduction of children. The issues of ethnicity and smoke exposure need further exploration. Differences in prevalence of smoking have been reported (MMWR, 1999) between not Hispanic Whites, African American, Asian, and Hispanic groups. Differences in prevalence are known to exist, but less is known about differences in smoke exposure based on household practices in various ethnic groups. Further investigation of cultural meanings of "respect" and family influences on smoking bans through qualitative methods in a variety of ethnic groups needs to be considered. Replication of this study with other ethnic groups will increase understanding of group differences necessary for development of strategies for reducing smoke exposure in these settings. Finally, consideration needs to be given to the study of smoke exposure in. households within a harm reduction framework. Harm reduction approaches to nicotine products focus on reducing the harms to the user as well as to the inhaler of second hand smoke (Riley et al, 1999). Research fmdings related to smoke exposure are beginning to show that the focus of protection of children is more effective in reducing smoke exposure than information on smoking cessation strategies for parents (Sorum and 206

PAGE 222

Bruusgard, 1996). Within the framework of harm reduction, smoking bans in households can be implemented "for the good of children" and allow parents to alter their behavior resulting in reduction of smoke exposure, without primary regard to whether or not smoking cessation has occurred. Harm reduction has application for reducing involuntary smoke exposure both in public places and in private households. Conclusions This study investigated characteristics of families related to smoke exposure at the household level, both qualitatively and quantitatively. The description of how families address smoking and smoking bans in their homes provides important information upon which strategies for reducing smoke exposure can be developed and tested. Quantitative findings, measured whenever possible at the household level, were similar to previous findings and confirmed relationships between reported and actual smoke exposure. Both qualitative and quantitative findings elucidated the differences and similarities between Hispanic and not Hispanic White households in understanding smoke exposure in households within a cultural context. Efforts for reduction of smoke exposure at the household level hold promise for increasing effectiveness in keeping children smoke free. 207

PAGE 223

While this study did not confirm previous findings associating smoking bans with educational level, income level, and age of parents, it represents one of the few efforts at validating reported smoking bans with actual measurements of smoke exposure through cotinine levels. The use of smoking bans as a measurement of smoke exposure clearly needs further refmement and replication of studies in which that can be investigated are needed. Based on social ecology, in which factors at many levels can be considered as they relate to health indicators, this work contributes to greater understanding of the role of behavioral factors in passive smoke exposure of young children. It provides data for development of strategies to prevent and/or counteract the effects of passive smoke exposure for children and their families. Future research needs to include investigation of factors in the household production of health framework among other ethnically diverse groups so that findings can be compared and culturally competent strategies developed for interventions in households with smoke exposure. Methodological issues evident in this research can be resolved through further testing of reliability and validity of knowledge and attitude measures, revision of survey questions to facilitate measurement of discrete variables for increased power in analysis. Continued investigation of behavioral facets of smoke exposure using qualitative and quantitative methodologies will ensure 208

PAGE 224

that smoke exposure and the resulting health problems will be prevented and/or minimized in young children. The focus of smoking bans as a method for predicting smoke exposure was central to this study. How useful is this as a method for predicting smoke exposure? Does it have applicability for interventions where reduced or no smoke exposure is the goal? The findings from this study indicate that the idea of smoking bans has potential usefulness in predicting smoke exposure in households where children reside. The presence and implementation of these bans can assist families in their ability to ensure a smoke free environment for their children. The findings from this study contribute to the growing body of knowledge of smoking bans and their importance in the continued smoke free environments for young children and their families. 209

PAGE 225

APPENDIX A INFORMED CONSENT AND SURVEY INSTRUMENT FOR SEMISTRUCTURED INTERVIEWS 210

PAGE 226

UCD Human Subjects Research Committee "Household Policies and Environmental Tobacco Smoke Exposure ofPre-Scbool children". Principal Investigator: Yvonne Yousey RN, PbDc SUBJECT CONSENT FORM Ma 2001 Project Purpose You are being asked to take part in a research study of smoking behaviors and rules about smoking in families with pre-school children. You are asked to be in this study to answer questions about passive smoke exposure in your home because you have pre-school age children under age 4. The purpose of this study is to explore how your family deals with sei:ond"hand smoke exposure in your home and with your children. It makes no difference in this study if there are people who smoke in your home or not. This research is being done to help our understanding of how parents with children, age 0-4 years, deal with exposure to second-hand smoke. This is important because some illnesses in young children may be caused or made worse by exposure to tobacco smoke. Knowing how parents deal with tobacco smoke expi>sure in their homes and with their children will help in developing better ways to support them in preventing illnesses in th eir children related to tobacco smoke exposure Participation in this study is entirely voluntary. The purpose of this form is to provide you with information so that you can decide whether or not you would like to be in the study. If you choose not to be in this study, your child will still receive services at the clinic This study will enroll 25 families from faculty practice clinics of University of Colorado Health Sciences Center, School of Nursing, of which your clinic is one Expected Duration of the Study The part of the study in which parents/caregivers are interviewed will take approximately 4 months to complete. Each interview will take approximately 45 minuteS. Project Procedures If you agree to take part in this study, you will be asked to answer questions about how smoking happens or does not happen at your home. Your answers to questions will be audiotaped and transcribed so they can be studied more at a later time. All information you provide in the interview will be confidential. There will be no follow-up visits. Study Duration The first part of the study will be completed over a 4 month period at your home or at a place that is convenient for you to be interviewed. You will have one interview which will last approximately 45 minutes. The information from this interview will be used to develop a survey on factors which affect smoking practices in homes. Discomforts and Risks Completing the interview may cause slight discomfort to parents/caregivers if talking about smoking and smoke exposure is a sensitive topic. Discussing smoking habits and practices may be slightly uncomfortable for families who have friends and relatives who smoke. Efforts to keep information confidential will reduce this discomfort. The researcher has no control over whether you may share your involvement in the study with others. COMIRB Consent Form Template F-017 Revision 001, Effective 1..08-2001 211

PAGE 227

Benefits There is no guarantee that you will receive any health benefit from participating in this research study. Referrals It is not anticipated that there will be any need to refer you or your child for any other care. If such a need should arise, the researcher will refer you to appropriate sources of care. Confidentiality The investigator in this study will treat your identity with professional standards of confidentiality. The information obtained in this study may be published in professional journals, but your identity will not be revealed. Audiotapes obtained during semi-structured interviews will be stored in a secure, locked cabinet available only to the researcher or designated research assistants. Once transcribed, the audiotapes will be erased. Hard copies of the transcription will be kept until the research project is completed and for three years thereafter. Coding. files of the transcription will be the property of the researcher. Completed surveys will be identified by code number and not by name. These will become property of the researcher and will be kept for three years after the study. Study Participation and Withdrawal Your participation in this study is totally voluntary. You may choose not to enter the study or withdraw from the study at any time without any loss of benefits to which you are entitled. Any significant new findings that relate to your participation in this study will be discussed with you. Invitation for Questions You will receive a copy of this consent form to keep. Please ask questions about this research or consent either now or in the future. You may ask questions before, during, or after the study and these questions will be answered unless doing so would compromise the methods used in the study. You may direct your questions to Yvonne Yousey at (303)266-1869. If you have questions regarding your rights as a research subject, please call the Office of Academic Affairs, CU-Denver Building, #700, (303) 5562550. AUTHORIZATION: I have read this paper about the study or it wai read to me. I understand the possible risk and benefits of that being in this study is voluntary. I choose to be and choose to have my child in this study. I know I can stop being in this study and my child will still get the usual medical care. I will get a copy ofthis consent form. (Initial all the previous pages of the consent form). Signature:. ___________ Print Name ________ Date. ____ subject; parent or guardian Consent form explained by:----------Print Name _______________ Date-----,-------Investigator_--'-----------Date _____ 212 2

PAGE 228

Consent Fonn Approval Allan Prochazka. MD/Stcphen Bartldl. R .Ph. Co-Cbairs, COMIRB Christopher Kuni, MDIKn Easterday, R.Ph. Co-Chair.;, COMIRB Adam RoSCIIberg. MD/David lawellin, Ph. O . Co-Chairs, COMIRB Date: Valid Through: ___ COLORADO MULTIPLE INSTITUTIONAL REVIEW BOARD "Household Policies and Environmental Tobacco Smoke Esposure orPre-Sebool Children". Project Description Principal Investigator: Yvonne Yousey RN, PbDe SUBJECT CONSENT FORM September 2001, Venion I You are being asked to take part in a research study of smoking behaviors and rules about smoking in families with pre-school children. You are being asked to be in this study to answer questions about passive snioke exposure in your home because you have pre-school age children under age 4 and bC:cause you speak and write English. The purpose of this study is to explore how your family deals with second hand smoke exposure in your home and with your children. It makes no difference in this study if there are people who smoke in your home or not. The information .from this interview will be used to develop a survey on factors which affect smoking practices in homes This research is being conducted to increase our understanding of how parents and/or caregivers of children age 0-4 years deal with exposure to second hand smoke in their homes. Knowing how parents and caregivers deal with tobacco smoke exposure in their homes and with their children will help in developing better ways to suppon parents and caregivers in deciding how to deal with tobacco smoke exposure. Participation in this study is entirely voluntmy. The purpose of this form is to provide you with information so that you can decide whether or not you would like to be in the study. If you choose not to be in this study, your child will still receive services at the clinic. This study will enroll 20 subjects who receive care at the faculty practice clinics ofUniversity of Colorado Health Sciences Center, School ofNursing. Procedures If you agree to take part in this study, you will be asked to answer questions about how smoking happens or does not happen at your home. You will be asked to allow the researcher to come to your home to discuss with you how you deal with smoke exposure in your home. If you prefer not to meet in your home, you may meet the researcher at another place which is quiet and private to have the interview. Your responses to questions will be audiotaped and transcribed for further study. All information provided in the interviews will be confidential. Information gathered in this studf::will be collected through one interview with you which will last approximately l hour. Discomforts and Risks COMIRB Consent Form Template F-017 Revision OOf, Eflec:tive 1-08-2001 Initials: ___ 213

PAGE 229

behavior to others, where the subject has told a health care provider about a serious threat of imminent physical violence against a Specific person or persons, it is required by law that this be reported to the proper authoritieS. In addition, should any information contained in this study be the subject of a court order or lawful subpoena, the University of Colorado might not be able to avoid compliance with the order or subpoena. AUTHORIZATION: I have read this paper about the study or it was read to me. I understand the possible risk and benefits of that being in this study is voluntary I choose to be in this study. I know I can stop being in this study and my child will still get the usual medical care. I will get a copy of this consent form. (Initial all the previous pages of the consent form). Signature: Print Date. ____ subject; parent or guardian Consent form explained by: __________ Print Name...,...-----Date ______ Investigator----------.,....----COMIRB ConSent Form Template f-{)17 Revision 001, Effective 1.{)8-2001 Date _____ 214

PAGE 230

INSTRUMENT for Use in Interviews Interview Protocol A semi-structured interview will conducted with 20 study subjects who are parents/caregivers of children, age newborn to 4 years of age. The purpose of this interview is to discover factors which influence how decisions are inade regarding environmental tobacco smoke exposure in family controlled settings to include the home and automobile (if applicable). The interview will be audiotaped for later transcription and analysis. I am interested in learning more about how you and your family make decisions that might affect your health and that of your children. I am especially interested in learning about what you do to keep your children healthy in your home. One of the things that families with young children deal with is deciding whether or not smoking is allowed in their home I would like to ask you some questions about this. I will tape your responses so that I can listen to them and record on paper what you tell me today I will not share this information with anyone and it will be kept in my records by number and not by your name. Please remember that it makes no difference to me if anyone here smokes Informant's relationship to children in home .......... Parent___ Caregiver ___ How many adults live in your home? __ How many children? __ Who in your family is most concerned about keeping your children healthy? What are the most important things that you do to keep your children as physically healthy as possible? (Please give me some examples). If someone in your family wants to do something that you know will not be good for your child's physical health and may harm your child's health, what do you do? What would be an example of something that you would consider harmful to your child's pl:iysicai health? If one or some of your friends does something around your child that you know is not good for your child's physical health and may be harmful to your child what do you do? What is an example of that? What have you h eard about the effects of others smoking around your children? 215

PAGE 231

What effect do you believe that exposure to tobacco smoke has on your children? What kinds of rules does your family have about smoking? Tell me about the rules that you have. Who do these rules apply to? If you do not have any, what smoking do you allow in your home? Where are people allowed to smoke in your home? Nowhere __ In specific rooms __ Which rooms (please list------------' Anywhere they want Some family members smoke in a variety of places where they live-outside, inside, in the car, etc. If there are smokers in your family, tell me all of the places that they smoke. How do you decide if you allow someone to smoke or not in your home? Demographic information Smokers living in home ......... Yes __ No__ How many? __ Informant a srnoker ...... Yes __ No __ Informant's relationship to children in home . ........ Parent Caregiver ____ Ages of children living in horne ______ Observations made while in the home: Type of home. Number of rooms. of smoking paraphernalia. Presence of people smoking in the home; Odor of smoke. 2 216

PAGE 232

APPENDIXB CONSENT FORMS FOR SURVEY QUESTIONNAIRE SURVEY QUESTIONNAIRE 217

PAGE 233

. Survey Number: ___ COLORADO MULTIPLE INSTITUTIONAL REVIEW BOARD Household and Environmental Tobacco Smoke Exposure of Pre-School children". Project Purpose Principal Investigator:. Yv:onne Yousey RN, PhDc REVISED SUBJECT CONSENT FORM June 2002 Protocol 01-624 You are being a.Sked to take part in a research study of smoking behaviors and rules about smoking in families with pre school children You are asked to be in this study to provide information about passive smoke exposure in your home because you have pre-school age children under age 5 The purpose of this study is to find out what your family does about second-hand smoke exposure in your home. It makes no differencein this study ifthere are people who smoke in your horne or not. This research is being done to help our understanding of how parents with children, age 0-5 years; deal with exposure to second-hand smoke This is important because some illnesses in young children may be affected by exposure to tobacco smoke Knowing how parents deal with tobacro smoke exposure in their homes will help in keeping children healthy related to tobacco smoke exposure PartiCipation in this study is entirely voluntary. The pJJrpose of this form is to provide you with information so that you can decide whether or not you would like to be in the study. If you choose not to be in this study, your child will still receive services at the clinic or may still attend school or Headstart This study will enroll 200 families from various locations around the Denver area. These include faculty practice clinics of University of Colorado Health Sciences Center, School ofNursing, Pre-schools,.Headstart facilities . Expected Duration of the Study This study in which a urine sample is obtained from the child of a parent completing a survey will take approximately six months to complete The time to obtain a urine sample will take from 15-30 minutes depending upon the cooperation of the child Project Procedures If you agree fo take part in this study, you will be asked to help your child get a urine sample which will be tested for cotinine, a way of measuring exposure to secondhand smoke A urine sample will be taken even if no one smokes in your home The results of the urine sample will be kept confidential. There will be no follow-up visits Discomforts and Risks Efforts to obtain a urine sample from your child may be a little uncomfortable for you and your child There is a slight chance that information obtained through your child's urine sample may not be kept confidential. Efforts by the researcher to store this jnformation using codes and keeping the results in a locked cabinet will reduce this as much as possible. This study may involve risks that are currently not known or predicted. If any of these occur, the researcher will notify you immediately and take steps to reduce these risks Initials : __ 218

PAGE 234

BenefitS There is no guarantee that you or your child will receive any health benefit from participating in this research study. This study is designed for the researcher to learn more about second hand smoke exposure. This study is not designed to treat any illness or to improve you or your child's health Also there are risks as mentioned in the Risk Section Source of Funding Funding for this research is provided by the Colorado Tobacco Research Program Support for the research is provided by Program for Health and Behavioral Science, University of Colorado, Denver and University of Colorado Health Sciences Center, School ofNursing, Faculty Practice, Denver, Colorado. Cost to Subject There is no cost to you or your child for participating in this study You will be paid up to $20 for you and your child's participation in this study. Referrals It is not anticipated that there will be need to refer your child for any other care. If such a need should arise, the researcher will refer you to appropriate sources of care Study Participation and Withdrawal Taking part in this study is voluntary. You have the right to choose not to take part in this study If you do not take part in the study, your clinic or doctor will still take care of you. You will not lose any benefits or medical care tc:i which you are entitled . If you choose to take part, you have the right to stop at any time If there are new findings during the study that may affect whether you want to continue to take part, you wili be told about them. The study investigator may decide to stop your participation without your pennission, if she thinks that being in the study may cause you harm, or for any other reason. Also the sponsor may stop the study at any time, Invitation for Questions The researcher carrying out this study is Yvonne Yousey. You may ask any questions you have now. If you have questions later, you may call the Colorado Multiple Institutional Review Board (COMIRB) office at (303) 724-1055. Confidentiality We will make every effort to keep your research records confidential, but it cannot be assured Records that identifY you and the consent form signed by you, may be looked at by a regulatory agency such as The Food and Drug Administration Department ofHealth and Human Services Colorado Multiple Institution Review Board Your records may also be looked at by Colorado Tobacco Research Program and the Faculty of Program for HCalth and Behavioral Sciences, University of Colorado, Denver. Initials: ___ 219 ...

PAGE 235

The rc:Sults of this research may. be presented at meetings or in published articles However, your name will be kept private Any data obtained in the study will b(;C()me property of the researcher and will be kept for three years after completinglhe study. Mandated Reporting of Child Abuse, Neglect or Threatened Violence Some things we cannot keep private If you give us any information about child abuse or neglect we have to report that to Social Services. If you tell us that you are going to physically hurt someone, we have to report that to police. Also, if we get a court order or subpoena to twn over your study records, we. will have to do that. AUTHORIZATION I have read this paper aboui the study or it was read to me. I understand the possible risk and benefits that being in this study is voluntary I choose to be and choose to have my child in this study. I know I con Slop be_ing in this study and my child will still getlhe usual medical care. I will gel a copy of this consent form (Initial all the previous pages of the consent form). Signature: ..,....,.,-.----::----... suardian Print name : ________ Date: ___ Consent form exj>lained by : ________ Print name: _______ Date:_ _____ Investigator: ___________ Date ____________ Survey Number :'-----220

PAGE 236

Approved Sc p \) n :l:;i1? 1_. tJz. Cuestionario Numero:._ ___ r ,"". ;:-\ ,tt, :.; ::.1 "-'''''COLORADO MULTIPLE INSTITUTIONAL REVIEW BOARD Exposici6n a) Humo del Cigarrillo en los Hogares yen el Medio Ambiente en Niiios basta Nivel Pre-Escolar. lnvestigador Principal: Yvonne Yousey RN, PbDc FORMA REVISADA DE APROBACJON Junio 2002 Protocol 01-624 Prop6sito del Proyecto A usted se le ha pedido que participe en este estudio de investigacion de habitos de fumadores y reglas de fumar establecidas en familias con niiios de edad pre-escolar. A usted se Je ha pedido tomar parte en este estudio. para proporcionar informacion acerca de exposicion pasiva aJ humo del cigarrillo en su hogar porque usted tiene nifios(as) rnenores de 5 afios. El prop6sito de este estudio es saber como se cop1porta su familia acerca de Ia exposicion de segunda mano al hurno del cigarrillo. No hay diferencia alguna en nuestro estudio si hay fumadores o no fumadores en su hogar. Esta investigacion nos ayudara a entender como los padres de nifios(as) de edades 0 a 5 afios reaccionan a Ia exposici6n de segunda mano al humo del cigarrillo .. Es importante, porque algunas enfermedades en nifios pequefios pueden verse afectadas porIa exposicion ai humo del cigarrillo. Aprender acerca de como se comportan los padres con Ia exposicion al hurno del cigarrillo en sus hogares no ayudara a mantener sanos a los niiios en relacion a Ia exposicion al humo del Cigarrillo. Su particfpacion es este estudio es completamente voluntaria. El objetivo de esta forma es proporcionarle informacion para que usted pueda decidir si desea participar en eJ estudio. Si usted decide no participar en este estudio, su niiio(a) alln recibir.i los servicios de Ia clinica o podr.i seguir atendiendo Ia escuela Headstart. Este estudio observara a 200 familias de diferentes regiones alrededor de Denver. Esto incluye clinicas de practica del Centro' de Ciencia y Salud y Ia Escuela de Enfermeria de Ia Universidad de Colorado de Denver, escuelas a nivel pre. escolares y ubicaciones de Hi:adstartc Duraci6n Estimada del Estudio Este estudio en el cual se obtiene una muestra de orina del nifio(a) y del padrelmadre.un cuestionario Completo toma aproximadamente seis ineses en total. El tiempo que toma obtener'Ia rnuestra de orina es de 15 a 30 minutos dependiendo del nivel de coope_racion del niiio(a). Procedimientos del Proyecto Si usted participa en este estudio, usted tendra que ayudar a su niiio(a) aobtener uria muestra de orina que ser.i analizada para e) contenido de citinina, una forma de medir Ia exposicion de segunda mano al humo del cigarrillo. Una muestra de orina ser.i tomada aun sin haber fum adores en .el hogar. Los resultados ser.in confidenciales. No habr.i visitas de seguimiento. Iniciales __ 221

PAGE 237

lliesgos e Incomodidades El esfuerzo para obtener una muestrn de orina de su niiio(a) puede ser inc6modo para usted y su pequeiio. Existe una pequeiia posibilidad de que Ia informacion obtenida a travis de su muestra de orina no sea completamente confidencial. Los esfuertos que hace el investigador para mantener Ia informacion de manera confidencial usando codigos y guardando los resultados en un gabinete cerrado reducira el riesgo lomas posible Este estudio puede tener riesgos hasta ahora desconocidos o irnprevisibles. Si alguno de estos riesgos ocurre, el investigador le notificara inmediatamente y tomara las acciones indicadas para disminuir este riesgo. Beneficios No haya garantia de que usted o su niiio(a) recibiran beneficios de salud por participar en este estudio de investigaci6n. Este estudio esta diseiiado para ayudar al investigador a aprender mas acerca de Ia exposicion de segunda mano al humo del cigarrillo. Este estudio no esta diseiiado para dar tratamiento para enfermedades o para mejorar su salud o Ia salud de su niiio(a). Tam bien e xi sten los riesgos mencionados en Ia secci6n .de riesgos. Fuentes de Fioanciamiento Los fondos para esta investigaci6n son proporcionados por el Programa Colorado Tobacco Research Program. El apoyo para esta investigaci6n es proporcionado por el Prograrna Program for Health and Behavioral Science, Ia Universidad de Colorado de Denver y por el Centro de Ciencia y Salud y Ia Escuela de Enfermeria de Ia Universidad de Colorado de Denver y por Pnicticas Privadas de Medicina de Denver, Colorado. Costo del Estudio No existe costo alguno para usted o su niiio(a) por participar en este estudio. Usted podni recibir hasta $20 por su participaci6n y Ia de su niiio en este estudio. Referencias Nose anticipa que se tenga que referir a su niiio(a) a otro tipo de cuidado. En caso de que sea necesario, el investigador le indicani l.as fuentes de cuidado apropiadas. Participacioo y Renuocia al Estudio Participar en este estudio es voluntario. Usted tiene el derecho a decidir no participar en el estudio. Si usted no toma parte en el estudio, su clinica o doctor aun le seguiran atendiendo Usted no perdera 1os beneficios de salud o cui dado medico a que tiene derecho. Si usted desea participar, tiene derecho a parar cuando lo desee. Si Ia informacion encontrada por medio del estudio le afecta para continuar tomando parte de este estudio, usted sera informado. El investigador puede tambien evitar su participacion sin su perrniso;si el investigador pieilsa que puede causarle alglin daiio al ser parte de este estudio o por cualquier otra raz6n Tambien los patroc i nadores pueden detener este estudio en cualquier momento. Invitacion a Preguntas El investigador efectuando este estudio es Yvonne Yousey. Usted puede hacer cualquier pregunta ahora. Si usted tiene preguntas despues, usted puede llamar a Ia oficina de Colorado Multiple InstitUtional Review Board (COMIRB) al (303) 724-1055. lniciales: __ 222

PAGE 238

Confidencial Haremos todo Jo posible por mantener los archivos de manera confidencial, sin embargo no podemos asegurarlo. Los archivos que lo identifican y Ia forma de aprobacion firmada por usted, puede ser revisada por algtin organismo regulador, tales como: The Food and Drug Administration (La Admirustracion de Comida y Medicarnentos) Department of Health and Human Services (Departarnentos de Salud y Servicios Hurnanos) Colorado Multiple Institution Review Board Sus archivos tarnbien pueden ser revisados por Colorado Tobacco Research Program y porIa facultad del Programa Program for Health and Behavioral Sciences de Ia Universidad de Colorado de Denver. Los resultados de este estudio pueden ser presentados en reuruones o publicados en articulos. Sin embargo, su nombre sera manterudo en privado. Cualquier dato obterudo del estudio sera propiedad del investigador y sera mantenido por un periodo de tres afios despues de haber completado el estudio. Reporte Obligato rio de Abuso a Menores, Negligencia o Violencia Existen situaciones que no pueden mantenerse en privado. Si usted nos da informacion de abuso a su nino( a) o negligencia; nosotros tenemos que r.eportarlo a Ia oficina de Servicios Sociales. Si usted nos dice a! go que esta haciendo para herir a alguien fisicamente, nosotros tenemos que reportarlo a Ia policia. Tam bien; si nosotros recibimos una orden de Ia corte o citacion para entregar los archivos de su estudio, los entregaremos. AUTOR1ZACION He leido o me /zan leido este documento acerca del estudio. Entiendo los posibles riesgos y beneficios y que tomar parte en este estudio es de manera voluntaria.. Yo quiero y. quiero que mi nino( a) participe en este estudio. Entiendo que puedo renunciar al estudio y que mi nino aun recibira el cuidado medico usuaL Yo recibire una copia de esta Forma de Aprobacion. (lnicialice todas las ptiginas anteriores de esta Forma de Aprobacion) Firma:. __ Sujeto, padre o guardi:1n Nombre Escrito: __________ ...,.______ Aprobacion explicada por: _______ Nombre Escrito:. _______ Fecha:. _____ Investigador:. ___________ -'--Fecba:. _______ Cuestionario Nfunero: 223

PAGE 239

Smoke Expostn Survey Dear Parent: Approved SEP 25 2fD This is a survey about smoking policies in your home. I am interested in learning about what you do about people smoking in your home. It does not .matter to me if there are people who smoke in your home or not. Please answer the questions in the survey. Your name will not be used and your answers on the survey will be confidential. You may choose not fill out this survey. Or you may choose to skip questions that you don't want to answer. Your child will receive normal school health or health services whether or not you decide to participate. After you finish the survey, I will ask that you help me get a urine sample from your preschool child. The urine will be tested for cotinine, which is a way of checking for smoke exposure. When you return the survey and get a urine sample from your child, you and your child are considered a part of the research study.. I will ask you to sign a special consent form to get a urine sample from your child. Or I may ask you to give me information so that I can contact you to get the urine sample from your child. If you agree to anSwer the survey and get a urine sample from your child, you will receive a grocery certificate for $20. The information from the survey and the results of the urine test will provide us with information about how to help families keep children as healthy as possible related to passive smoke exposure. If you have any questionS or concerns about participating, please call Yvonne Yousey at 303-266-1869. If you have any questions about your rights as a research participant, please call the Colorado Multiple Institutional Review Boord (COMIRB) office at 303-724-1055. Please return the survey to your school or clinic. And complete the iftformation form. so that I can contact you to arTange for obtaining the urine specimen. THANK-YOU SO MUeHl. Funded by Colorado Tolx=o l'rogrUII, Boulclu. Colorado 224

PAGE 240

s To be by researcher! Survey ___ IL ---L--J 61575 Please answer the following questions about your household by filling in the box. rn m 1. How .. many adults live in your home, at .leaat part of the t i me? 2. How many children live in your home, at least part of the time? 3 In years, how old is each child living in your home? (If a child is less 1 year old, please X the hoX.) Child Ill rn Child 112 rn Child 113 rn Child 114 rn Child 115 rn .Child #6 []] SEr 2}1002 4 ilhat is the zip code of your heme? I s In general, how healthy is each child living in your home? (ile mean children under age 18 who live with you at least part of the time.) Child 1 Child 2 Is this child usually healthy? Yes No Y -------------4---_ ...J._ es No In the past year, i l bow many illnesses 1 ITJ 1 (such ae colcil, sore j eb.roatB ear : infections has this I child bad? Provide : ; ___ ..;.,___..., Doea thia chiid : : Yes No _; _____ I health problem? Does -this child have asthma? No 1 rn . Ye s No Ye S No _Child 3 Child .. i I I Yes No Yes No I IITJ rn I I I Yes No \ Yes No I I Yes NO I Yes No Sa. If you have any children over age 10, do th'U' spoke cigarette_s? Ch:l.ld 5 Child 6 I Yes No I Yes No rn I rn I Yes No Yes No I I Yes No Yes No I 0No 6. When was the last time anyone smoked YO'\%' home? one choice only. Owithin the past 24 hours 0 Between 24 hours and one week ago 0 Between _one week and one month ago 0 Betwe_en one month and one year ago 0 More than one year ago 225

PAGE 241

.E; survey rol 3 61575 7. Is there a child in your home who is currently breast fed? 0Yes 0No B .. Including.yourself, how manypeople living in your home currently smoke ? sa. For those people who smoke and live in your household, please identify their relationship to child(ren) living in the home. Cbeck al l that apply. 0 Mother (Stepmother) 0 Father (or Stepfather) Qother 0 (s) (or relative) 0 Friend (a) (or similar person) sb. If people smoke in your home, bow many cigarettes did they smoke in the past week? Only one choice. 0 Less tha n 20 (1 pk) 40 (1-2 pks) 041-60 (2-3 pks) Omore than 60 (more than 3 pks) 9. What is your current smoking status? only one choice. 0 Neve r smoked (less than 100 cigarettes in your life) 0 Ex-smoker, (more than 100 cigarettes in your life, but none in the past 30 days) 0 current smoker (smoked within the past 30 days) 10. Which of the following best describes the rules for smoking in your home? Check only one. 0 No one is ever allowed .to smoke in my home. 0 People can smoke sometime s or some places in my home 0People can smoke whenever and wherever they want in the bouse. WHERE do people .smoke in your home? Please go through each line, and mark one' choice that is truefor your household. 11. Smoking is allowed in a room with an open window. 0Yes 0No 12. Smokiilg is okay in whi .ch rooms? A Kitchen QYes 0No B. Living or family room OYes 0No c. Bedroom QYes QNo o. Bathroom 0Yes 0No E. Basement 0Yes 0No F Garage 0Yes 0No G. Porch/deck 0Yes QNo : .. 226

PAGE 242

61575 This is about SITUATIONS when people 13. Smoking is allowed indoors when children are at home 14. Smoking is allowed indoors when children are in the room. 15. Smoking is allowed indoors when children are asleep. 16. Most families I know allow cigarette smoking in their homes. 17. Some relatives are allowed to smoke in my home. 18. Some friends are allowed to smoke in my home. 19. Does your family have a vehicle {car, van, tr.uck)? smoke ;'' survey IDI in your home. Make one choice per 0Yes 0No 0Yes 0No O'Yes 0No 0Yes 0No 0Yes 0No 0Yes 0No 0Yes 0No Answer the following only if you answered yes to the question 19. 4 question. Where do people smoke in your family's vehicles {car, van, or truck)? Please select YES or NO for each of the following. 20. 21. Smoking is ai1owed in our carts) when no children are in the vehicle. Smoking is allowed in our car{s) when children are in the vehicle, if the window is rolled down. 0Yes 0Yes 22. Smoking is allowed in our car(s} when.. children are in the vehicle regardless if the window{s) is up or 0Yes down. 23. Smoking is never allowed in our vehicles at any time. 0Yes O _No 0No 0No 24. How often do you make exceptions for people to smoke in your home? One cboice only. O'Never Q Once a year to a {ew times a year 0 Several times a month 0 More than once a week 2 '4a. If you make exceptions, please select YES or NO for each of the following: It is too cold/rainy/windy to 0Yes QNo go outside. A grandparent or other relative has a right' to smoke if they want. QYes 0No I don't feel comfortable telling someone not to smoke in my house 0Yes 0No 227

PAGE 243

survey roiTTJ 5 61575 25. Who rules for smoking or not smoking in your home? Check all that apply. 0 Mother (or Stepmother) 0 Father (or Stepfather) 0 Children 0Grandparent or other.older relative 0 Other 25a. If there are rules about smoking in your home, who makes sure that people follow the rules? Check all that apply. 0 Mother (or Stepmother) 0 Father (or Stepfather) 0 Children 0 Grandparent or other older-relative The next set of questions is about what you know about smoke exposure and children. 26. If a child is healthy, tobacco smoke will not harm them. 27. The scientific evidence doesn't really prove tobacco smoke is harmful. 28. Indoor tobacco smoke makes children's asthma worse. 29. children who are often around indoor tobacco smoke have more colds and coughs. 30. People who are exposed to tobacco smoke when they are children are more likely to get cancer as adults. 31. Children who are often around indoor tobacco smoke h ave more earaches 32. Smoke exposure hurts children only if they already are sick. 33. Smoke exposure hurts children only if they a chronic illness like asthma. 34. A little smoke exposure not harm a baby. 35. A little smoke exposure will not harm a child. 36. Smoke exposure harms children whether they are sick or healthy. 37. The dangers of tobacco smoke have been exaggerated 228 0 True 0 False 0 True 0 False 0 True 0 False 0 True 0 False 0 True 0 False 0 True 0 False 0 TrUe 0 False 0 True 0 False 0 True 0 False 0 True 0 False 0 True 0 False 0 True 0 False.

PAGE 244

survey ml 6 61575 For each of the following statements about tobacco smoke exposUre, mark the response that fits best for you. One choice per question only. Strongly Agree Neither agree Disagree Strongly agree somewhat nor disagree somewhat disagree 38. I believe that smoke exposure D D D D D is harmful to children. 39. 1 hate it when 1 see adults D D D D 0 smoking around children. 40. It makes me mad when people D D D D D smoke indoors around children. 41. It's not a big deal if adults D D D D D smoke around children. 42. It's okay for people to smoke D D D D D around children as long as they don't smoke around my kids. 43. I don't mind when people smoke D around me. D D D D 44. Children should never be exposed 0 0 0 D D to environmental tobacco smoke. 45. I get mad or upset when 1 see D D D D D someone smoking close to a baby. Please go to next page ..: 229

PAGE 245

survey roO_=rJ 7 61575 Here are some statements about what some people do about other people smoking . For each statement, please mark the response that is best for you. Strongly Agree Neither agree Disagree Strongly agree somewhat nor disagree somewhat disagree 46. A parent should protect D D D D D their children from smoke exposure. 47. Parents have the right to D D 0 D 0 decide whether or not they will smoke.around their children. 48. We should have laws against D D D D D smoke exposure in homes just like in public places and work places. 49. smoking indoors should be 0 D D D D illegal in homes where children live. so. It is my job to keep my D D D D D child(ren) away from tobacco smoke. 51. It is my job to teach my D D D D D child(ren) about dangers and health effects of tobacco smoke. Please go to next page 230

PAGE 246

Survey roLI __ -L __ 8 61575 Please answer the following questions about you and your family: 01. How old are you? Ooivorced 02. Are you now? One choice only 0 Single OMarried 03. What is your sex? One choice only FemaleO Male 0 04. How much school have you completed? Please select only one. 0 Less than 6th grade 0 Between 6th and 8th grade 0 Some high school 0 High school graduate or GEO 0 Technical or trade school 0 Some college 0 College graduate 0 Graduate study 0 Living with partner OS. If you have a spouse/partner, how much school has he/she "completed? .Please select only one. 0 Less than 6th grade 0 Between 6th and 8th grade 0 Some high school 0 High school graduate or GED 0 Technical or trade school 0 Some college 0 College graduate 0 Graduate s .tudy Plea.se go to the next page 231

PAGE 247

liEd 61575 survey miT lJ 9 06. Please mark whether each of the following p eople is Hispanic/Latino or not. 07. Hispanic or Latino You Your spouse/partner (if you have one) Child 1 Child 2 Child 3 0 0 0 0 0 Not Hispanic or Latino 0 0 0 0 0 ----------------------Child 4 0 0 Child 5 0 0 -----------------------------Child 6 0 0 Please mark the race of each of the following people. Native Black or Hawaiian or White African Asian Pacific American Islander You 0 0 0 0 ----------Your. spouse/ 0 partner 0 0 0 (if you have one) -----------Child 1 0 0 0 0 ---------. -----------------Child 2 0 0 0 0 ----------------.. ------Chiid 3 0 0 0 0 -----------------Child 4 0 0 0 0 Child 5 0 0 0 0 Child 6 0 0 0 0 232 Am. Indian or Alaka Native 0 0 0 0 0 0 0 0 Other 0 0 0 0 0 0 0 0

PAGE 248

61575 Su:rvey IDj .____.___..___. DB. What is your total yearly household income, including money that you, your spouse/partner, or other adults in your household make? Make one choice. () c:$10, 000 per year ()$10,000 $20,000 ()$20,001 $30,000 0$30,001 $40,000 09. Do you rent or own your home? ()Rent ()own D10. Do you live in: Choose one only. ()House ()Apartment ()Trailer Home ()Duplex or Tri-plex or Townhome D11 Do the windows in your home open? ()Yes ()Yes, they open, but not easily () Yes, but we don't open them "()No ()$40,001-$50,000 0$50,001 $70,000 ()Above $70, 000 D12. _How many rooms are in your residence? Mark all that apply. () Living Room ()Family Room () Dining Room OKitchen ()Bathroom. ()Bedrooms ()Garage ()other Please list ------------------Number ofbathrooms, if more than .Number of bedrooms, if more than 233 10

PAGE 249

. .-.A Survey 61575 D13. Does your child spend time in daycare on a regular basis? Dl7a. If so, how much? 0 Daily, for part of every day. 0Daily, all day. 0Less than every day but more than 2 days per week. 0 Less than 2 days per week. Dl4. Does your child have contact with someone who smokes in daycare? OYes DlS What languages do. you speak?-0English 0 Spanish Oother (list) 11 0No Dl6 Please indicate what languages are spoken by the people living at your house? Both Engli.!'h' only More English and More Only English than Spanish/Other Spanish/Other Spanish/Other Spanish/Other Lanaguage Language than Language Language Equally English You 0 0 0 0 0 -----Your spouse/ 0 0 0 0 0 partner (if you have one) --------Child 1 0 0 0 0 0 --...... ---------Child 2 0 0 0 0 0 ------------------. Child 3 0 0 0 0 0 -------Child 4 0 0 0 0 0 -------childS 0 0 0 0 0 --------Child 6 0 0 0 0 0 END OF SURVEY THANK .YOU for this survey. 234

PAGE 250

For Researcher Only Survey IDI I I I 12 61575 Todays Site C Baker SBHC 0 Sheridan PreSchool 0 Carin Ciinc 0 Adams County Heads tart 0 Gregory Hill SBHC 0 Other 0 Community Health Services D District so Preschool D District 14 PreSchool Cotinine level of urine L__l__j Age of child L__l__j 235

PAGE 251

Cuestionorio de Exposicion ol Humo del Cigorrillo Estimado(a) Padre/Madre: Approved OCT COMIRB ESte es un cuestionario relacionado a las costumbres y reglas de fumar en su hagar. Tengo interes en saber Ia que usted hace cuando las personas fumon en su hagar. No es importante para ml si hay o no hay personas fumadoras en su hogar. Por favor responda a las preguntas del cuestionario Su nombre no sera usado y sus respuestas sercin confidenciales. Usted tiene Ia opci6n de no responder a este cuestionario. Usted tambien puede dejar de contestar algunos de estas preguntas. Su nino( a) recibirci el servicio normal de salud en Ia escuela o cllnica sin importar si usted decide o no participar Despues de terminar este cuestionario, le pedire que me ayude a obtener uno muestra de orina de su nifio(a) en edad pre-escolar. La orina sera examinada para determiner el contenido de Cotinina, que es una manera de exposici6n al humo del cigarrillo. Cuando usted nos devuelva el cuestionario y Ia muestra de orina de su nifio(a), usted y su nino( a) seron considerados parte de este estudio de investigacion. Le pediri que firme una forma de autorizacion para obtener Ia muestra de orina de su nifio(a. 0 Je pedire que me de informacion para poder comunicarme con usted y obtener Ia muestra de orina .de su nifio(a). Si uSted ocepta completer este cuestionario y proporcionar Ia muestra de orioo de su nino(a), usted recibirci Lin certificado de compra de comestibles de $20 dolores La informacion del cuestionario y los resultados del examen de orina nos ayudarcin a obtener informacion acerca de como ayudar a las familias a mantener a sus pequenos sanos. Si usted tiene cualquier pregunta o duda acerca de partidpar en este estudlo, por favor llame a Yvonne Yousey al303-266-1869. Si usted tiene preguntas acerca de sus derechos como participant"e en este estudio de investigaci6n por favor llame a Ia oficina de Colorado Multiple Institutional Review Board (COMIRB) al303-724-1055. Por favor devuelva este cuestionario a su escuela o clinica. Complete Ia forma de informacion para que nosotros podamos comunicornos con usted y obtener el especimen de orina. IMuchisimas Gracias! Este proyecto esta patrocinado por el Programa Tobacco Research Program. Boulder, Colorado 236

PAGE 252

2 Para ser completado por el encuestador. Cuestionario No. !.___...___,___, Por favor responda a las siguientes pregunta acerca del su hogar completando los siguientes espacios. 1. i. Cuantos adultos viven en esta casa, al menos tiempo parcial? m 2. (.Cuantos niiios viven en esta casa, al menos tiempo parcial? OJ 3, (.Que edad (en aiios) tienen los niiios que viven en su casa? (Si el niiio(a) tiene menos de un afio por favor marque el espacio con una "x") Nino( a)# 1 ITJ Nino( a)# 4 ITJ Nino( a) # 2 [I] Niiio(a) # 5 [I] 4. (.Cual es su c6digo postal? Ll -L.......J...--1.--l.---' Niiio(a) # 3 [I] Niiio(a) # 6 [I] En general, (.QUe tan saludables son los niiios que viven en su hogar? (Ninos menores de 18 anos que vivan con usted por lo menos parte del tiempo) 5 NinOfaT1 Ninola 2 Nino(a)3 Nino(a)4 Nino(a)5 Nino(a)6 (.Este nino normalmente Si No Si No Sl No Sl No Sl No Sl No tiene buena salud? Durante el ano pasado, (.cuantas enfermedades, tales como resfriados, dolor. de garganta, e infecciones de oldo ha tenido ITJ [I] ITJ UJ UJ OJ este nino(a)? Proporcione un numero. (.Tieneeste nino(a) un problema cr6nico desalud o enfermedad Sl No Si No Sl No Sl No Sl No Sl No seria? l. Tiene asma este nino(a) Sl No Si No Sl No Si No Sl No Si No 237

PAGE 253

53 i_Aiguno de los niiios mayores de 10 aiios que viven en su hogarfuman cigarrillos? OSi ONo 6. i_Cuando fue Ia ultima vez que alguien fum6 dentro de su casa? Una respuesta o Dentro de las ultimas 24 horas. o Entre las ultimas 24 horas y hace una semana. o Entre hace una semana y hace un mes. o Entre hace un mes y hace un aiio. o Hace mas de un aiio. 7. i_Aigunos de sus niiios estan siendo alimentados con leche materna? OSI ONo B. lncluyendolo a usted, i_Cuantas de personas que viven en su hogar fuman? ITJ :3 B.a. Para aquellas personas fumadoras que viven en su hogar, por favor identifique Ia relaci6n con los niiios que tambiem en su hogar. lndique todas las opciones que apliquen. o Madre (Madrastra) 0 Abuelos(as) (o pariente) o Padre (Padrastro) o Amigos(as) (o personas similares) ootro B.b. Si alguien fuma en su casa, i_Cuantos cigarrillos fumaron durante Ia ser'nana pasada? Una respuesta unicamente. o Menos de 20 (1 paquete) o 20-40 (1-2 paquetes) o 41-60 (2-3 paquetes) o Mas de 60 (mas de 3 paquetes) 9. l,Cual es su situaci6n actual? Una respuesta unicamente. o Nunea he fumado.(menos de 100 cigarrillos en su vida) .. o Ex fum!!dor (mas de 1 00 cigarrillos en su vida, pero ninguno en los ultimos 30 dlas). o Fumador (fum6 en los u.ltimos 30 dlas) 238

PAGE 254

1 O.{.Cual de las siguientes oraciones describe mejor las reglas de fumar en su hogar? [Marque una respuesta unicamente } o Nadie esta autorizado a fumar dentro de mi casa. o Las personas pueden fumar a veces y en algunos lugares en mi casa. o Las personas pueden fumar donde sea y cuando sea en mi casa. l,D6nde furrian las personas dentro de sucasa? 4 [Por favor revise cada una de Jas.preguntas, y marque las secciones que sean verdad en su casa ] 11. Fumar esta permitido en una habitaci6n mientras Ia ventana esta abierta 0 Sl o No 12. (.En cual de las siguientes habitaciones de Ia casa esta permitido fumar? a. Cocina OSI ONo b. Sala o cuarto de recreo OS! ONo c Recamara OSi ONo d.Bano OSI ONo e. S6tano OSi ONo f Garaje (cochera) osr ONo g. Patio (porche) OSI ONo Lo siguiente se trata de situaciones en que las personas fuman en su hogar. [Por favor marque una respuesta por oraci6n ] 13. Fumar esta permitido adentro cuando los nirios estim en casa. OSI ONo 14. Fumar esta pel'mitido adentro cuando los nirios estari en el cuarto. OSi ONo 15. Fumar esta permitido adentro cuando ios nifios estan dormidos OSi ONo La mayoria parte de las familias que conozco permiten fumar dentro de su casa. OSi ONo 17. Algunos de mis parientes tienen permitido fumar dentro de mi casa. OSi ONo 239

PAGE 255

-----------.nr----------------------------------------------------------------18. Algunos amigos tienen permitido fumar dentro de mi casa. 19. i, Tiene su familia un vehiculo (carro, van o camioneta)? OSI ONo OSI ONo [Conteste las siguientes preguntas solamente si su respuesta a Ia pregunta 19 fue afirmativa.] i,D6nde fuman las personas dentro de su vehlculo familiar (carro, van o camioneta)? Por favor seleccione sl o no a las siguientes preguntas. 20. Fumar esta permitido en el vehlculo cuando los nifios estan adentro. 0 Si 21. Fumar estfl permitido en el vehlculo cuando los ninos estan adentro si Ia ventana esta abierta 0 Si 22. Fumar esta permitido en el vehlciJio cuando los ninos estan adentro si Ia ventana esta cerrada. 0 Si 23. Fumar no esta permitido en el vehiculo en ningun memento. 0 Sr 24. (,Que tan seguido hace usted una excepci6n para que una persona pueda fumar en su casa? o Nunca o Una o varias veces al afio. o Varias veces al mes. o Mas de una vez a Ia semana ONo ONo ONo ONo 24.a. Si usted hace excepciones, por favor conteste sf o no a las siguientes preguntas: muy frio o lh:Nioso o ventoso para salir afuera 0 Si 0 No Un abuelo o pariente tiene derecho a fumar adentro 0 Si 0 No No me siento c6modo(a) de decirle a alguien que no fume en mi casa OSi ONo 5 25. {.Quien pone las reglas de fumar o no fumar en su casa? lndique todas las opciones que apliquen OMadre 0 Nifios OPadre 0 Abuelos 0 Otro 25.a. Si ex i sten reglas acerca de fumar en su hogar, j,quien se asegura de que las reglas sean cumplidas? lndique todas las opciones que apliquen. o Madre o Padre ONifios .. OAbuelos 240

PAGE 256

Las siguientes preguntas se refieren a su conocimiento acerca de exposici6n al huino del cigarrillo en los ninos(as). 26. Si un nino( a) tiene buena salud, el humo del cigarrillo no le va a hacer dafio. 0 Cierto 0 Falso 27. Las pruebas cientificas realmente no prueban que el humo del cigarrillo es dafiino. 0 Cierto O Falso 28. El humo del cigarrillo dentro de Ia casa hace que el asma de los niflos empeore 0 Cierto OFalso 29. Los ninos que estim expuestos frecuentemente al humo del cigarrillo dentro de Ia casa tienen mas los resfriados y tos. 0 Cierto 0 Falso 30. Las personas que estim expuestas ai humo del cigarrillo cuando son pequenas tienen mas probabilidades de tener cancer cuando son adultos. 0 Cierto OFaiso 31. Los nifios que estan expuestos frecuentemente al humo del cigarrillo dentro de Ia casa tienen mas dolores de oidos. 0 Cierto 0 Falso 32. El humo del cigarrillo les hace daflo a los ninos(as) solamente si ya estan enfermos. 0 Cierto 0 Falso 33. El humo del cigarrillo les hace dafio a los niilos(as) solamente si ya tienen una enfermedad cr6nica, como el asma. 0 Cierto 0 Falso 34. Un poco de humo de cigarrillo no dafia a un bebe. 0 Cierto 0 Falso 35. Un poco de humo de cigarrillo no dana a un niflo(a). 0 Cierto 0 Falso 36. El humo del cigarrillo les hace dafio a los ninos(as) si estan sanos o enfermos. 0 Cierto OFalso 37. Los peligros del humo del cigarrillo han sido exagerados. 0 Cierto 0 False 241 6

PAGE 257

.. Marque Ia respuesta que sea mas apropiada para usted en las siguientes ali nnaciones porcada acerca de exposici6n a/ humo del cigarril/o Marque una respuesta unicamente linea. T Completamente I De I Neutral I En I C de acuerdo acuerdo desacuerdo e ompletamente n desacuerdo 38. Yo creo que el humo del 0 D 0 0 cigarrillo dana a los nifios(as). 0 39. Detesto ver a adultos fumar alrededor de nifios(as). 0 0 D 0 0 40. Me enoja cuando los adultos fuman dentro de un D 0 0 0 Iugar donde hay nifios(as) D 41. No es algo grave cuando los adultos fuman alrededor 0 0 0 0 de los nifios(as). 0 42. Las personas pueden fumar alrededor de algunos nil\os(as), 0 D 0 D mientras no lo hagan alrededor 0 de mis nil\os(as). 43. No me molesta que las personas fumen a ml alrededor. 0 0 D 0 0 44. Los nifios(as) no deben estar expuestos nunca a un 0 0 D 0 ambiente de humo de cigarrillo. D 45. Me enoja cuando veo a alguien fumar alrededor de 0 [j D 0 un bebl!. 0 Por favor pase a Ia siguiente pagina 242 7

PAGE 258

8 Marque Ia respuesta que sea mas apropiada para usted en /as siguientes afirmaciones acerca de Jo que a/gunas personas hacen cuando otros fuman. Completamente en desacue.rdo 46. Un padre/madre debe proteger a sus hijos contra 0 0 0 0 0 Ia exposici6n al humo del cigarrillo. 47. Los padres deben decidir si quieren o no fumar alrededor 0 0 0 0 0 de sus hijos. 48. Deberiamos de tener leyes contra Ia exposici6n al humo del cigarrillo dentro 0 0 0 0 0 de los hogares, al igual que Ia que tenemos en lugares publicos y en el trabajo. 49. Fumar deberia ser ilegal en hogares donde viven nilios(as). 0 0 0 0 0 50. Es mi deber mantener a mis hijos alejados del D 0 0 0 0 humo del cigarrillo. 51. Es mi deber ensefiar a mis hijos acerca de los peligros y efectos secundarios del humo del 0 0 0 0 0 cigarrillo. Por favor pase a Ia siguiente pagina 243

PAGE 259

Por favor responda las siguientes preguntas acerca de usted y su familia: D1. l,Cuantos afios tiene? [IJ D2. Es usted (una respuesta unicamente). 0 Casado(a) 0 Soltero(a) 0 Divorciado(a) D3. Sexo (una respuesta unicamente) 0 Masculine 0 Femenino D4 . Nivel de educaci6n (una respuesta unicamente) o Menos del 6" grado. o Entre el 6" y el a grado. OVrve con su Pareja o Algo de escuela preparatoria (high school) o Diploma de escuela preparatoria (high school) 6 GED o Escuela tecnica o de oficio o Algo de universidad o Graduado de Ia universidad o Educaci6n post-universitaria D5. Si usted tiene un esposo(a) 6 pareja l,Cual es su nivel de educaci6n? (una respuesta unicamente). o Menos del 6" grado. o Entre el 6" y el a grado. o Algo de escuela preparatoria (high school) o Diploma de escuela preparatoria (high school) 6 GED o Escuela tecnica o de oficio o Algo de universidad o Graduado de Ia universidad o Educaci6n post-universitaria D6. Por favor indique cual de las siguientes person.as es Iatino o no. Usted Hispano o Latino 0 Su esposo(a)/ Pareja (si tiene) 0 Nifio(a) 1 0 Nino( a) 2 0 Nifio(a) 3 0 NifioCal 4 0 Nino! a) 5 0 NifioCal 6 0 No Hispano o Latino 0 0 0 0 0 0 0 0 Por favor pase a Ia siguiente pagina 244 9

PAGE 260

07. Por favor indique Ia raza de las siguientes personas. Anglo Negro o Asiatico Nativo Indio Americano Africa no Hawaiano o Nativo de Americano o Islas del Alaska Pacifico Usted 0 0 0 0 0 Su esposo(a)/ Pareja (si tiene) 0 0 0 0 0 Nifio{a) 1 0 0 0 0 0 Nifio{a) 2 0 0 0 0 0 Nifio(a) 3 0 0 0 0 0 Nifio(a) 4 0 0 0 0 0 Nifio(a) 5 0 0 0 0 0 Nifio(a) 6 0 0 0 0 0 DB. l,Cual es su salario total anual del hogar, incluyendo dinero que usted, su esposo(a)/pareja, u otros adultos en su casa obtienen? Una respuesta (.micamente. 0 < $10,000 por aiio 0 $10,000 "'$20,000 0 $20,001 $30,000 0 $30,001 $40,000 0 $40,001 $50,000 0 $50,001 $70,000 0 Mas de $70,000 09. l,Usted renta o es dueiio de su casa? 0 Rento 0 Dueno(a) 010. Usted vive en: escoja una respuesta OCasa 0 Apartamento 0 Casa Trailer 0 Casa duplex o triples 0 Condominia o fraccionamiento 011. l,laS ventanas de su hogar se abren? OSI. 0 sr abren, perb diflcilmente. 0 Si, pero no las abrimos. ONo. 245 10 Otro 0 0 0 0 0 0 0 0

PAGE 261

012. l,Cuantos cuartos tiene su residencia? Marque todos los que apliquen. 0 Sala 0 Garaje/cochera 0 Cuarto de Estar 0 Comedor 0 Cocina 0 Otro Por favor indique: 0 Bano 0 Recamaras Numero de banos Numero de recamaras ITJ ITJ 013. l,SU hijo(a) pasa tiempo en Ia guarderia regularmente? 0 Si 0 No D13.a. Si contest6 "sl" t.cufmto tiempo? 0 Diario, parte del dia de cada dia. 0 Oiario, todo el dia. 0 Menos de cada dia, pero al menos 2 dias por semana. 0 Merios de dos dias por semana. 014. l,En Ia guarderia, tiene contacto su hijo(a) con alguien que fuma? OSI 0 No 015. {.Que idiomas habla usted? Marque todos los que apliquen. o Ingles o Espanol 0 Otro ______________ ,.(mencione} 016. Por favor indique cuales idiomas se hablan las personas que viven en su hogar. Solo Mas Ingles Ambos Ingles Mas EspafioV Ingles que EspaiioV y EspanoVOtro Otro ldioma Otro ldioma ldioma lgualmente que Ingles Usted 0 0 0 0 Su esposo(a}/ Pareja (si tiene} 0 0 0 0 Nifio(a}1 0 0 0 0 Nifio(a}2 0 0 0 0 Nifio(a}3 0 0 0 0 Nifio(a}4 0 0 0 0 Nifio(a}5 0 0 o 0 Nifio(a}6 0 0 0 0 Final del Cuestionario Gracias por Completar este Cuestionario 246 11 Solo EspafioV Otro ldioma 0 0 0 0 0 0 0 0

PAGE 262

Cuestionario No.D D D Para ellnvestigador/Encuestador Solamente Fecha del Dia de Hoy ITJDJITJ Ubicaci6n OBaker SBHC OPre-escolar Sheridan OCinica Carin OHeadstart del Condado de Adams 0Gregory Hill SBHC OOtro -----Paid $20 D OServicios Sociales Comunitarios 0 Pre-escolar del Distrito 50 DPre-escolar del Distrito 14 Nlvel de Cotinina en Ia Orina Urine obtained on site 0 Urine brought in D Edad del nino(a) [LJ 247 12

PAGE 263

APPENDIXC RECRUITMENT FLYERS FOR QUANTITATIVE PHASE 248

PAGE 264

Ao::;roveci Do you have a child under age 5? Smokers and Non-Smokers Tell us what you knowHHU Fill out a survey Help us get a urine sample from your $20 when you If interested, ask for a survey at (specify locatior\). : or Call Yvonne Yousey 303-.266-1869 249 ......

PAGE 265

WHAT IS IT? ADo roved COMIRB ! This is a study to learn more about how families handle tobacco smoke exposure in their households. It is important because knowing more about how families handle smoke exposure around children will help in keeping children healthy. ! This study includes answering a questionnaire and getting a urine sample from a young child in the household. The urine will be tested for cotinine which comes from nicotine. ! All information collected in the study will be kept confidential. WHO CAN PARTICIPATE? ! Any family that has a child between ages of newborn and 5 years. ! Families with smokers and families with no smokers are welcome. WHY PARTICIPATE? : Completing the survey will help us know more about how families deal with smoke exposure in their homes. : Getting a urine sample will help us know whether or not your child might be exposed to smoke. : You will be paid $20 after the survey is completed and we are able to get a urine sample from your child. If you are interested, fill out the survey and leave at (specified location) Someone will contact you about getting the urine from your child. FOR MORE INFORMATION, you may also call Yvonne Yousey at 303-266-1869 Funded by Colorado Tobacco Research Program, Boulder, Colorado. 250

PAGE 266

. siCION AL HUMO ov DE CIGARR:t.
PAGE 267

iDe que se Trata? ! Este es un estudio para aprender mas acerca de como las familias sobrellevan fa exposicion al humo del cigarrillo en sus hogares. Es importante para Ia salud de nuestros hijos saber como se comportan las familias ante fa exposicion af humo del cigarrillo cuando los ninos estan presentes. ! Este estudio incluye responder un cuestionario y obtener una muestra de orino de los ninos pequenos en el hogar. La muestra de orina sera examinada para determinar ef contenido de cotinina que proviene de Ia nicotina. ! Toda Ia informacion encontrada por medio de este estudio es confidenciaf. iQUIEN PUEDEPARTICIPAR? ! Cualquier familia que tenga ninos de 0 a 5 anos. ! Las familias con miembros fumadores o no-fumadores son bienvenidas . iPORQUE PARTICIPAR? ! AI compfetar este cuestionario usted nos ayuda a entender como las familias sobrellevan Ia exposicion al humo del cigarrillo en sus hogares. ! AI tener Ia muestra de orina nos ayudara a saber su nino( a) ha estado expuesto(a) al humo de cigarrillo. ! A usted se le pagaran $20 despues de que ef cuestionario sea completado y de que obtengamos una muestra de orina de su nino(a). Si usted esta interesado(a), complete y entregue este cuestionario y alguien le llamara pcira obtener Ia muestra de orina de su nino(a). PARA MAS INFORMACION, Usted pLiede !lamar o Yvonne Yousey al 303-266-1869 Patrocinado por el Programa Colorado Tobacco Research Progi'Grn. Boulder Colorado. 252

PAGE 268

BIBLIOGRAPHY Aday, L. (1996). Designing and Conducting Health Surveys. 2nd Ed. San Francisco: Jossey-Bass Publishers. Al-Delaimy, W., Crane, J. & Woodward, A. (2001). Passive smoking in children: effect of avoidance strategies at home as measured by hair nicotine levels. Archives ofEnvironmental Health 56 (2), 117-122. Al-Delai,ny, W., Luo, D., Woodward, A., & Howden-Chapman, P. (1999). Smoking hygiene: a study of attitudes to passive smoking. New Zealand Medical Journal, 112, 33-36. Aligne, C.A. & Stoddard, J. (1997). Tobacco and children: an economic evaluation of the medical effects of parental smoking. Archives of Pediatric Adolescent Medicine, 151, 648-653. Anderson, P. & Hughes, J. (2000). Policy interventions to reduce the harm from smoking. Addiction, 95, (Supplement 1 ), S9-S 11. Arborelius, E., Hallberg, A. C., & Hakansson, A. (2000). How to prevent exposure to tobacco smoke among small children: a literature review. Acta Paediatrica, 89 ( 434), 65-70. Ashley, M.J., Cohen, J., Ferrence, R., Bull, S., Bondy, S., Poland, B., & Pederson, L. (1998). Smoking in the home: changing attitudes and current practices. AJPH 88 (5), 797-799. Ashley, M.J. & Ferrence, R. (1998). Reducing children's exposure to environmental tobacco smoke in homes: issues and strategies. Tobacco Control, 7 (1), 61-65. Bahceciler, N. Barlan, LB. Nuhoglu, Y. & Basaran, M. (1999). Parental smoking behavior and the urinary cotinine levels of asthmatic children. Journal of Asthma, 36 (2), 171-175. Barbour, R.S. (1998). Mixing qualitative methods: quality assurance or qualitative quagmire? Qualitative Health Research, 8 (3), 352-361. 253

PAGE 269

Barbour, R.S. (2001). Checklists for improving rigour in qualitative research: a case ofthe tail wagging the dog? BMJ, 322, 1115-1117. Benowitz, N .L. (1996). Cotinine as a biomarker of environmental tobacco smoke exposure. Epidemiological Reviews, 18 (2), 188-204. Benowitz, N.L. (1999). Biomarkers of environmental tobacco smoke exposure. Environmental Health Perspectives, 107 (Suppl2), 349-355. Benowitz, N.L. & Jacob, P. III. (1994). Metabolism of nicotine to cotinine studied by a dual stable isotope method. Clinical Pharmacology Therapeutics, 34 (5), 483-493. Berman, P., Kendall, C., & Bhattacharyya, K. (1994). The household production of health: integrating social science perspectives on micro-level health determinants. Social Sciences and Medicine, 38 (2), 205-215. Biener, L., Cullen, D., Zhu, X.D., & Hammond, S.K. (1997). Household smoking restrictions and adolescents' exposure to environmental tobacco smoke. Preventive Medicine, 26, 358-363 Borland, R. (1999). Theories of behavior change in relation to environmental tobacco smoke control to protect children. International Consultation on Environmental Tobacco Smoke and Child Health, WHO. Background paper. Borland, R., Mullins, R., Trotter, L., & White, V. (1999). Trends in environmental tobacco smoke restrictions in the home in Victoria, Australia. Tobacco Control, 8, 266-271. Bronfenbrenner, U. (1986). Ecology of the family as a context for human development: research perspectives. Developmental Psychology, 22 (6), 723-742. Brownson, R.C., Eriksen, M.P., & Warner, K.E. (1997). Environmental tobacco smoke: health effects and policies to reduce environmental tobacco smoke exposure. Annual Review ofPublic Health, 18, 163-185. Brownson, R.C., Jackson-Thompson, J., Wilkerson, J.C., Davis, J.R., Owens, N.W., & Fisher, E.B. (1992). Demographic and socioeconomic differences in beliefs about the health effects of smoking. American Journal of Public Health, 82 (1), 99103. 254

PAGE 270

CDC. (Centers for Disease Control and Prevention). (1989). Reducing the High Consequences of Smoking: 25 Years of Progress-A Report of the Surgeon General. US Department of Health and Human Services, Public Health Service, CDC, DHHS publication no. (CDC) 89-8411. Washington, DC. CDC. (1994). Current trends attitudes toward smoking policies in eight states-United States, 1993. Morbidity Mortality Weekly Report, 43 (43), 786-789. CDC. (1999). Cigarette smoking among adults-United States, 1997. Morbidity Mortality Weekly Reports, 48, 993-996. CDC. (2000). Strategies for reducing exposure to environmental tobacco smoke, increasing tobacco-use cessation, and reducing initiation in communities and health-care systems. Morbidity Mortality Weekly Report, 49, 1111112000 (RR12), 1-11. CDC. (2002). Cigarette smoking among adults-United States, 2000. Morbidity Mortality Weekly Report, 51 (29), 642-645. Colley, J.R.T. (1971). Respiratory Disease in Childhood. British Medical Bulletin, 27, 9-14. Committee on Assessment of Asthma and Indoor Air, Division of Health Promotion and Disease Prevention, Institute of Medicine (2000). Exposure to environmental tobacco smoke. Clearing the Air: Asthma and Indoor Air Exposure. National Academy Press, 263-297. Cook, D.G. & Strachan, D.P. (1999). Health effects of passive smoke. 10. Summary of effects of parental smoking on the respiratory health of children and implications for research. Thorax, 54 (4), 357-366. Cook, D.G., Strachan, D.P., & Carey, I.M.,. (1998). Health Effects of Passive Smoking: 9 Parental smoking and spirometric indices in children Thorax, 53 (10): 884-893. Cook, D.G., Whincup, P.H., Jarvis, M.J., Strachan, D.P., Papacosta, 0. & Bryant, A. (1994). Passive exposure to tobacco smoke in children aged 5-7 years: individual, family, and community factors. BMJ, 308, 384-389. Corbett, K. (2001). Susceptibility ofyouth to tobacco: a social ecological framework for prevention. Respiration Physiology, 128, 103-118. 255

PAGE 271

---------------------Crabtree, B. & Miller, W. (ed). (1992). Doing Qualitative Research. Newbury Park, Ca: Sage Publications. Crone, M., Reijnevald, S., Burgeijer, R., & Hirasing, R. (2001). Factors that influence passive smoking in infancy: a study among mothers of newborn babies in the Netherlands. Preventive Medicine, 32 (3), 209-217. Cummings, K.M., Markello, S.J., Mahoney, M., Bhargava, A.K., McElroy, P.D., & Marshall, J.R. (1990). Measurement of current exposure to environmental tobacco smoke. Archives of Environmental Health, 45, 74-79. Czaja, R. & Blair, J. (1996). Designing Surveys: A Guide to Decisions and Procedures. Thousand Oaks, Ca.: Pine Forge Press, 1996. DaVanzo, J. & Gertler, P. (1990). Household Production of Health: A Microeconomic Perspective on Health Transitions. Santa Monica, Ca: The RAND Corporation. Denzin, N. & Lincoln, Y. (1998). The Landscape of Qualitative Research. Thousand Oaks, Ca: Sage Publications. DiClemente, C., Prochaska, J.O., Fairhurst, S.K., Velicer, W.F., Velasquez, M.M., & Rossi, J.S. (1991 ). The process of smoking cessation: an analysis of precontemplation, contemplation, and preparation stages of change. Journal of Consulting and Clinical Psychology, 59, 295-304. DiFranzia, J. & Lew, R. (1996). Morbidity and mortality in children associated with the use of tobacco products by other people. Pediatrics, 97 (4), 560568. Doll, R. (1998). Uncovering the effects of smoking: historical perspective. Statistical Methods in Medical Research, 7, 87-117. Drope, J. & Chapman, S. (2001). Tobacco industry efforts at discrediting scientific knowledge of environmental tobacco smoke: a review of internal industry documents. Journal of Epidemiology and Community Health, 55 (8), 588-594. Elder, J., McGraw, S.A., Stone, E.J., Reed, D.B., Harsha, D.W., Greene, T. & Wamgsgans, K.C. (1994). CATCH: Process evaluation of environmental factors and programs. Health Education Quarterly 2, S107-S127. 256

PAGE 272

Emmons, K.M., Abrams, D.B., Marshall, R.J., Etzel R.A., Novotny, T.E., Marcus, B.H., & Kane, M.E. (1992). Exposure to environmental tobacco smoke in naturalistic settings. American Journal of Public Health, 82, 24-28. Emmons, K.M., Hammond, K., & Abrams, D. (1994). Smoking at home: The impact of smoking cessation on nonsmokers' exposure to environmental tobacco smoke. Health Psychology, 13 (6), 516-520. Emmons, K.M., Wong, M., Hammond, S.K., Velicer, W.F., Fava, J.L., Monroe, A.D., & Evans, J.L. (2001). Intervention and policy issues related to children's exposure to environmental tobacco smoke. Preventive Medicine, 32 (4), 321-331. Erikson, P.l., & Kaplan, C.P. (2000). Maximizing qualitative responses about smoking in structured interviews. Qualitative Heath Research, 10 (6), 829-840. Eriksen, W. & Bruusgaard, D. (1995). Smoking behavior in young families. Do parents take practical measure to prevent passive smoking by the children? Scandinavian Journal of Primary Health Care.13 (4), 275-280. Eriksen, W., Sandvik, L., & Bruusgard, D. (1997). Social support and the smoking behavior of parents with pre-school children. Scandinavian Journal of Social Medicine, 25 (2), 93-99. Etzel, R.A. (1997). Active and passive smoking: hazards for children. Central European Journal of Public Health. 2, 54-56. Farkas, A., Gilpin, E., Distefan, J., & Pierce, J.P. (1999). The effects of household and workplace smoking restrictions on quitting behaviors. Tobacco Control, 8, 261-265. Farkas, A., Gilpin, E., White, M., & Pierce, J. (2000). Association between household and workplace smoking restrictions and adolescent smoking. JAMA, 284 (6), 717-722. Fearnow, M., Chassin, L., & Presson, C. (1998). Determinants ofparental attempts to deter their children's cigarette smoking. Journal of Applied Developmental Psychology, 19 (3), 453-468. 257

PAGE 273

Fiore, M.C., Jorenby, D.E., & Baker, T.B. (1997). Smoking cessation: principles and practice based upon the AHCPR Guideline, 1996. Agency for Health Care Policy and Research. Annals ofBehavioral Medicine,19 (3): 213-219. Fontana, A., & Frey, J. (1998). Interviewing: the art of science. InN. Denzin & Y.S. Lincoln (Eds.), Collecting and Interpreting Qualitative Materials (pp. 47-78). Thousand Oaks, Ca: Sage Publications. Fried, P., Perkins, S., Watkinson, B., & McCartney, J. (1995). Association between creatinine-adjusted and unadjusted urine cotinine values in children and the mother's report of exposure to environmental tobacco smoke. Clinical Biochemistry, 28 (4), 415-420. Fukushima, T., Fujiwara, T., Matsumoto, J., Isobe, A., Shiwak:u, K., Ishizak:i, E., Tsuchiya, S., & Yamane, Y. (1997). Children's smoking habits in an agricultural community in Japan and the importance of a no smoking program integrating family, school, and community health. Public Health,111, 289-292. Gaffuey, K. & Lynch, J. (2000). Tobacco smoke exposure and pediatric otitis media: an empirical basis for practice. Lippincott's Primary Care, 4 (5), 508-514, Gergen, P., Fowler, J., Maurer, K., Davis, W., & Overpeck, M. (1998). The burden of environmental tobacco smoke exposure on the respiratory health of children 2 months through 5 years of age in the United States Third National Health and Nutrition Examination Survey. Pediatrics,101 (2), 1-6. Gidding, S.S. (1999). Effects of passive smoking on the cardiovascular system in children and adolescents. WHO. International Consultation on Environmental Tobacco Smoke and Child Health Background paper. Gilpin, E., White, M., Farkas, A., & Pierce, J. (1999). Home smoking restrictions: which smokers have them and how they are associated with smoking behavior. Nicotine and Tobacco Research, 1 (2), 153-162. Giovino, G.A. (2002). Epidemiology of tobacco use in the United States. Oncogene, 21. 7326-7340. Goldstein, J. (1994). Informal smoking controls in Winnipeg households. Canadian Journal of Public Health, 85 (2), 106-109. 258

PAGE 274

Gottlieb, N., Eriksen, M., Lovato, C., Weinstein, R., & Green, L. (1990). Impact of a restrictive work site smoking policy on smoking behavior, attitudes, and norms. Journal of Occupational Medicine, 32 (1), 16-23. Green, J. & Britten, N. (1998). Qualitative research and evidence based medicine. BMJ, 316, 1230-1232. Green, L., Richard, L., & Potvin, L. (1996). Ecological foundations of health promotion. American Journal of Health Promotion,lO (4), 270-281. Greenberg, R., Strecher, V., & Bauman, K. (1994). Evaluation of a home based intervention to reduce infant passive smoking and lower respiratory illness. Journal of Behavioral Medicine,l7 (3), 273-290. Hair, J., Anderson, R., Tatham, R., & Black, W. (1998). Mulitivariate Analysis. 5th Ed. New Jersey: Prentice Hall. Hammond, S.K., Sorensen, G., Youngstrom, R., & Ockene, J. (1995). Occupational exposure to environmental tobacco smoke. JAMA, 274 (12), 956-960. Hauford,V. & Lison, D. (1998). Urinary cotinine as a tobacco-smoke exposure index: a minireview. International Archives of Occupational & Environmental Health, 71 (3), 162-168. Henderson, F.W., Redi, H.F., Morris, R., & Hy, P.C. Helms, R.W., Forehand, L., Mumford, J., Lewtas, J., Haley, N.J., et al. (1989). Home air nicotine levels and urinary cotinine excretion in preschool children. American Review of Respiratory Disease,l40 (1), 197-201. Henriksen, L. & Jackson, C. (1998). Anti-smoking socialization: relationship to parent and child smoking status. Health Communication,lO (1), 87-101. Henschen, M., Frischer, T., Pracht, T., Spiekerkotter, E., Karmaus, W., Meinert, R., Lehnert, W., Wehrle, E., & Kuehr, J. (1997). The internal dose of passive smoking at home depends on the size of the dwelling. Environmental Research, 72 (1), 65-71. Hosmer, D. & Lemeshow, S. (1989). Applied Logistic Regression. New York: John Wiley & Sons, Inc. 259

PAGE 275

Hosmer, D., Taber, S., & Lemeshow, S. (1991). The importance of assessing the fit of logistic regression models: a case study. AJPH. 81 (12), 16301635. Hopkins, D., Briss, P., Ricard, C., Husten, C., Crande-Kulis, V., Fielding, J., Alao, M., McKenna, J., Sharp, D., Harris, J. Woollery, T., & Harris, K. (2001). Reviews of evidence regarding interventions to reduce tobacco use and exposure to environmental tobacco smoke. American Journal of Preventive Medicine, 20 (2S), 16-66. Hopkins, D., Husten, C., Fielding, J., Rosenquist, J.N. & Westphal, L. (200 1 ). Evidence reviews and recommendations on interventions to reduce tobacco use and exposure to environmental tobacco smoke. A summary of selected guidelines. American Journal of Preventive Medicine. 20 (2S), 67-88. Hopper, J. & Craig, K. (2000). Environmental tobacco smoke exposure among urban children. Pediatrics. 106 (4), 1-5. Hovell, M., Zakarian, J., Matt, G., Hofstetter, C.R., Bemert, J.T., & Pirkle, J. (2000). Effect of counseling mothers on their children's exposure to environmental tobacco smoke: randomized controlled trial. British Medical Journal, 321 (7257), 337-342. Hovell, M., Zakarian, J., Wahlgren, D., & Matt, G. (2000). Reducing children's exposure to environmental tobacco smoke: the empirical evidence and directions for future research. Tobacco Control, 9 (Supplement II), ii40-ii47. Huberman, A.M.& Miles, M. (1998). Data management and analysis methods. InN. Denzin & Y.S. Lincoln (Eds.), Collecting and Interpreting Qualitative Materials (pp. 179-210). Thousand Oaks, Ca: Sage Publications. Rudzinski, L. & Sirois, P. (1994). Changes in smoking behavior and body weight after implementation of a no-smoking policy in the workplace. Southern Medical Journal, 87 (3), 322-327. Hughes, J. (1999). Four beliefs that may impede progress in the treatment of smoking. Tobacco Control, 8, 323-326. Institute of Medicine, (2000). Clearing the Air: Asthma and Indoor Exposures. Washington, DC: National Academy Press. 260

PAGE 276

Jackson, C. & Henriksen, L. (1997). Do as I say: parent smoking, antismoking socialization, and smoking onset among children. Addictive Behavior, 22, 107-114. Jarvis, M. (1999). Children's exposure to passive smoking: survey methodology and monitoring trends. WHO. International Consultation on Environmental Tobacco Smoke and Child Health. Joad, J.P. (2000). Smoking and pediatric respiratory health. Clinics in Chest Medicine, 21 (1), 37-46. Jones, R. (1996). Research Methods in the Social and Behavioral Sciences. Sunderland, Mass.: Sinauer Associates, Inc. Jordaan, E., Ehrlich, R., & Potter, P. (1999). Environmental tobacco smoke exposure in children: household and community determinants. Archives of Environmental Health, 54 (5), 319-327. Kaplan, C.P., Napoles-Springer, A., Steward, S., & Perez-Stable, E. (2001). Smoking acquisition among adolescents and young Latinas: the role of socioenvironmental and personal factors. Addictive Behaviors, 26, 531-550. Kegler, M. & Malcoe, L. (2002). Smoking restrictions in the home and car among rural Native American and White families with young children. Preventive Medicine, 35, 334-342. Kline, R. (2000). Smoke knows no boundaries: legal strategies for environmental tobacco smoke incursions into the home within multi-unit residential dwellings. Tobacco Control, 9, 201-205. Kuzel, A. (1992). Sampling in qualitative inquiry. In B.F. Crabtree & W.L. Miller (Eds), Doing Qualitative Research. Newbury Park, CA: Sage Publications, pp.31-44. Lam, T.H., Hedley, A.J., Chung, S.F., & Macfarlance, D.J. (1999). Passive smoking and respiratory symptoms in primary school children in Hong Kong. Human and Experimental Toxicology, 18,218-223. Lam, T.H., Leung, G.M., & Ho, L.M. (2001). The effects of environmental tobacco smoke on health services utilization in the first eighteen months of life. Pediatrics, 107 (6), e91, 1-6. 261

PAGE 277

Leech, J.A., Wilby, K., & McMullen, E. (1999). Environmental tobacco smoke exposure patterns: a subanalysis of the Canadian human time-activity pattern survey. Canadian Journal of Public Health, 90 (4), 244-249. Levesque, L., Richard, L., & Potvin, L. (2000). The ecological approach in tobacco-control practice: health promotion practitioner characteristics related to using the ecological approach. American Journal of Health Promotion, 14 (4), 244252. Li, J.S., Peat, J.K., Xuan, W., & Berry, G. (1999). Meta-analysis on the association between environmental tobacco smoke exposure and the prevalence of lower respiratory tract infection in early childhood. Pediatric Pulmonology, 27 (1 ), 513. Lincoln, Y.S. & Guba, E.G. (1985). Naturalistic lnguiry. Beverly Hills, Ca: Sage Publications. Lund, K.E., Skrondal, A., Vertio, H., & Helgason, A.R. (1998). To what extent do parents strive to protect their children from environmental tobacco smoke in the Nordic countries? A population based study. Tobacco Control, 7, 56-60. Maggs-Rapport, F. (2001). Best practice research: in pursuit of methodological rigor. Journal of Advanced Nursing, 35 (3), 373-383. Mannino, D.M., Caraballo, R., Benowitz, N., & Repace, J. (2001). Predictors of cotinine levels in US children: data from the third National Health and Nutrition Examination Survey. Chest, 120 (3), 1-13. Mannino, D.M., Moorman, J., Kingsley, B., Rose, D., & Repace, J. (2001). Health effects related to environmental tobacco smoke exposure in children in the United States. Archives of Pediatric and Adolescent Medicine 115, 36-41. Mannino, D.M., Siegel, M., & Husten, C. (1996). Environmental tobacco smoke exposure in children: results from the 1991 National Health Interview Survey. Tobacco Control. 5, 13-18. Manuel, J. (1999). Double exposure. Environmental tobacco smoke. Environmental Health Perspectives, 107 (4), 196-204. 262

PAGE 278

Marks, D.F. (1998). Addiction, smoking and health: developing policy based interventions. Psychology, Health and Medicine, 3 (1 ), 97 -Ill. Mascola, M., VanVunakis, H., Tager, 1., Speizer, F., & Hanrahan, J. (1998). Exposure of young infants to environmental tobacco smoke: breastfeeding among smoking mothers. American Journal of Public Health, 88 (6), 893-896. Matt, G., Hovell, M., Zakarian, J., Bemert, J., & Pirkle, J. (2000). Measuring secondhand smoke exposure in babies: the reliability and validity of mother reports in a sample oflow-income families. Health Psychology, 19 (3), 232-241. Matt, G., Wahlgren, D., Melbourne, H., Zakarian, J., Bemert, J., Meltzer, S., Pirkle, J., & Caudill, S. (1999). Measuring environmental tobacco smoke exposure in infants and young children through urine cotinine and memory-based parental reports: empirical findings and discussion. Tobacco Control, 8 (3), 282-289. Mattila-Wiro, P. (1999). Economic Theories of the Household: A Critical Review. Helsinki, Finland: UNU World Institute for Development Economics Research. 1-41. May, C. (1998). The preparation and analysis of qualitative interview data. In B. Roes and C. Webb (Eds. ). Research and Development in Clinical Practice. London: Whurr Publishers. Mays, N. & Pope, C. (1995). Qualitative research: rigour and qualitative research. BMJ, 311 (July 8), 109-112. Mays, N. & Pope, C. (1995). Qualitative research: observational methods in health care settings. BMJ, 311 (July 15), 182-184. Miles, M. & Huberman, A.M. (1994). Qualitative Data Analysis. Thousand Oaks, Ca: Sage Publications. Netting, R., Wilk, R.R. & Amould, E.J. Ed. (1984). Households: Comparative and Historical Studies of the Domestic Group. Los Angeles: University of California Press. Norman, G., Ribisl, J., Howard-Pitney, B., & Howard, K.A. (1999). Smoking bans in the home and car: do those who really need them have them? Preventive Medicine, 29 (6), 581-589. 263

PAGE 279

Norman, P., Bell, R., & Conner, M (1999). The theory of planned behavior and smoking cessation. Health Psychology, 18 (1 ), 89-94. Oakley, A. (1993). Essays on Women. Medicine. and Health. Edinburgh: Edinburgh University Press Ltd. Okah, F., Choi, W., Okuyemi, K., & Ahluwalia, J. (2002). Effect of children on home smoking ban restriction by inner-city smokers. Pediatrics, 109 (2), 244-249. Ownby, D.R., Johnson, C.C., & Peterson, E.L. (2000). Passive cigarette smoke exposure of infants: importance of nonparental sources. Archives of Pediatrics & Adolescent Medicine. 154 (12), 1237-1241. Pederson L., Wanklin, J., Bull, S & Ashley, M. (1991). A conceptual framework for the roles of legislation and education in reducing exposure to environmental tobacco smoke. American Journal of Health Promotion, 6 (2), 105110. Perez-Stable, E., Marin, G., & Posner, S. (1998). Ethnic composition of attitudes and beliefs about cigarette smoking. Journal of General Internal Medicine. ll, 167-174. Pershagen, G. (1999). Accumulating evidence on health hazards of passive smoking. ActaPaediatrica. 88 (3), 491-492. Peterson, E. Johnson, C. & Ownby, D. (1997). Use of urinary cotinine and questionnaires in the evaluation of infant exposure to tobacco smoke in epidemiologic studies. Journal of Clinical Epidemiology, 50 (8), 917-923. Phillips, K., Bentley, M., Mohammed, A., Howard, D., & Cook, J. (1999). Low level saliva cotinine determination and its application as a biomarker for environmental tobacco smoke exposure. Human and Experimental Toxicology, 18, 291-296. Pirkle, J.L., Flegal, K.M., Bemert, J.T., Brody, D.J., Etzel, R.A., & Maurer, K.R. (1996). Exposure ofthe US population to environmental tobacco smoke: the third National Health and Nutrition Examination Survey, 1988-1991. JAMA, 275 (16), 1233-1240. 264

PAGE 280

Pizacani, B., Martin, D., Stark, M., Koepsell, T., Thompson, B., & Diehr, P. (2003). Household smoking bans: which households have them and do they work? Preventive Medicine. 36, 99-107. Pope, C. & Mays, N. (1995). Qualitative research: reaching the parts other methods cannot reach: an introduction to qualitative methods in health and health services research. BMJ, 311 (July 1), 42-45. Pope, C. & Mays, N. (2000). Analyzing qualitative data. BMJ. 320 (January 8), 114-116. Prochaska, J.O., Redding, C., & Evers, K. (1997). The transtheoretical model and stages of change. InK. Glanz, F. Lewis, & B. Rimer. (Eds.), Health Behavior and Health Education. (pp. 60-84). San Francisco: Jossey-Bass Publishers. Prochaska, J. et al. (1994). Stages of change and decisional balance for twelve problem behaviors. Health Psychology, 13, 39-46. Rickert, W.S. (1999). Background paper. Environmental tobacco smoke: properties, measurement techniques, and applications. WHO. International Consultation on Environmental Tobacco Smoke and Child Health. Riley, D., Sawka, E., Conley, P., Hewitt, D., Mitic, W., Poulin, C., Room, R., Single, E., & Topp, J. (1999). Harm reduction: concepts and practice. A policy discussion paper. Substance Use and Abuse. 34 (1), 9-24. Sallis, J. & Owen, N. (1997). Ecological models. InK. Glanz, F. Lewis & B. Rimer (Eds. ), Health Behavior and Health Education: Theory, Research, and Practice (pp. 403-424). San Francisco: Jossey Bass Publishers. Samet, J.M. (1999). Synthesis: the health effects oftobacco smoke exposure on children. WHO. International Consultation on Environmental Tobacco Smoke and Child Health. Background paper. Samet, J., Lewit, E., & Warner, K. (1995). Involuntary smoking and children's health. Current Practices in Pediatrics (July), 189-203. Sasco, A. & Vainio, H. (1999). From in utero and childhood exposure to parental smoking to childhood cancer: a possible link and the need for action. Human and Experimental Toxicology, 18, 192-201. 265

PAGE 281

Scherer, G., Meger-Kossien, 1., Riedel, K., Renner, T., & Meger, M. (1999). Assessment of the exposure of children to environmental tobacco smoke by different methods. Human Experimental Toxicology, 18, 297-301. Schuster, M., Franke, T., & Pham, C. (2002). Smoking patterns of household members and visitors in home with children in the United States. Archives of Pediatric and Adolescent Medicine 156, 1094-1100. Seifert, J., Ross, C., & Norris, J. (2002). Validation of a five-question survey to assess a child's exposure to environmental tobacco smoke. Annals of Epidemiology, 12 (4), 273-277. Sackrider, M. (1996). The respiratory effects of passive tobacco smoking. Current Opinion in Pulmonary Medicine, 2, 129-133. Sorrell, J. & Redmond, G. (1995). Interviews in qualitative nursing research: differing approaches for ethnographic and phenomenological studies. Journal of Advanced Nursing, 21 (6), 1117-1122. Stein, R., Haddock, C., O'Byrne, K., Hymowitz, N., & Schwab, J. (2000). The pediatrician's role in reducing tobacco exposure in children. Pediatrics, 106 (5), 1-17. Stepans, M.B. & Fuller, S. (1999). Measuring infant exposure to environmental tobacco smoke. Clinical Nursing Research, 8 (3), 198-221. Stokols, D. (1996). Translating Social Ecological Theory into guidelines for community health promotion. American Journal of Health Promotion, 10 (4), 282297. Subramaniam, R.P., Turim, J., Golden, S.L., Kral, P., & Anderson, E.L. (200 1 ). An exploratory study of variations in exposure to environmental tobacco smoke in the United States. Risk Analysis, 21 (3, 561-574.) Sweda, E. (1998). Protecting children from exposure to environmental tobacco smoke. Tobacco Control, 7, 1-2. Tabachnick, B. & Fidell, L. (2001). Using Multivariate Analysis. Boston: Allyn & Bacon. 266

PAGE 282

Thompson, B. & Kinne, S. (1990). Social change theory: Applications to community health. InN. Bracht. (Ed.), Health Promotion at the Community Level. New York: Sage Co. US Department of Health and Human Services (1986). The Health Consequences of Involuntary Smoking: A Report of the Surgeon General. Department of Health and Human Services, Public Health Service, Office of Smoking and Health. DHHS Publication No. (CDC) 87-8398, Government Printing Office, Washington, DC. US Department of Health and Human Services (1989). Making Health Communications Work: a Planner's Guide. Bethesda, Md: National Cancer Institute, Office of Cancer Communications. US Department of Health and Human Services (1989). Reducing the Health Consequences of Smoking: 25 Years of Progress. A Report of the Surgeon General. Department of Health and Human Services, Public Health Service, Centers for Disease Control, National Center for Chronic Disease Prevention and Health Promotion, Office of Smoking and Health DHHS Publication No. (CDC) 89-8411, Atlanta, Georgia US Environmental Protection Agency. (1992). Respiratory Health Effects of Passive Smoking: Lung Cancer and Other Disorders. EP A/600/6-90/006F. Washington DC. Wakefield, M., Banham, D., Martin, J., Riffin, R., McCaul, K., & Badcock, N. (2000). Restrictions on smoking at home and urinary cotinine levels among children with asthma American Journal of Preventive Medicine, 19 (3), 188-192. Wakefield, M., Chaloupka, F., Kaufman, N., Orleans, C., Barker, D., & Ruel, E. (2000). Effect of restrictions on smoking at home, at school, and in public places on teenage smoking: cross sectional study BMJ, 321, 333-336. Wang, M., Eddy, J., & Fitzhugh, E. (1995). Application of odds ratio and logistic models in epidemiology and health research. Health Values, 19 (1), 59-62. Warner, K. (2000). The need for, and value of, a multi-level approach to disease prevention; the case of tobacco control. Promoting Health: Intervention Strategies From Social and Behavioral Research. Washington, DC.: Institute of Medicine. 267

PAGE 283

Weinstein, N.D.& Klein, W.M. (1996). Unrealistic optimism: present and future. Special issue: unrealistic optimism about personal risks. Journal of Social and Clinical Psychology, 11, 97-105. Wewers, M.E. & Uno, M. (2002). Clinical interventions and smoking ban methods to reduce infants' and children's exposure to environmental tobacco smoke. JOGNN, 31 (5), 592-598. WHO. (1999). International Consultation on environmental Tobacco Smoke and Child Health. Consultation Report. Geneva, Switzerland. Willers, S., Axm.on, A., Feyerabend, C., Nielson, J., Skarping, G., & Skerfving, S. (2000). Assessment of environmental tobacco smoke exposure in children with asthmatic symptoms by questionnaire and cotinine concentrations in plasma, saliva, and urine. Journal of Clinical Epidemiology, 53, 715-721. Woodward, A. & Al-Delaimy, W. (1999). Measures of exposure to environmental tobacco smoke. Annals of New York Academy of Sciences, 895, 156-172. Consultation Report. (1999) WHO. International Consultation on Environmental Tobacco Smoke and Child Health. Geneva, Switzerland. Zapka, J., Fletcher, K., Pbert, L., Druker, S., Ockene, J., & Chen, L. (1999). The perceptions and practices of pediatricians: tobacco intervention. Pediatrics, I 03 (5), 1-10. Zimmerman, R. & Vernberg, D. (1994). Models of preventive health behavior: comparison, critique, and meta-analysis. Advances in Medical Sociology, 1. 45-67. 268