Citation
The relationship between maternal expectations for premature infant health outcome and mother-infant interaction

Material Information

Title:
The relationship between maternal expectations for premature infant health outcome and mother-infant interaction
Creator:
Gjerde, Jill
Publication Date:
Language:
English
Physical Description:
ix, 96 leaves : ; 28 cm

Subjects

Subjects / Keywords:
Mother and infant ( lcsh )
Premature infants -- Health and hygiene ( lcsh )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Bibliography:
Includes bibliographical references (leaves 89-96).
General Note:
Department of Psychology
Statement of Responsibility:
by Jill Gjerde.

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:
62768696 ( OCLC )
ocm62768696
Classification:
LD1193.L645 2005m G43 ( lcc )

Full Text
THE RELATIONSHIP BETWEEN MATERNAL EXPECTATIONS FOR
PREMATURE INFANT HEALTH OUTCOME AND MOTHER-INFANT
INTERACTION
by
Jill Gjerde
B.A., Wartburg College, 2003
A thesis submitted to the
University of Colorado at Denver
in partial fulfillment
of the requirements for the degree of
Master of Arts
Psychology
2005


This thesis for the Master of Arts
degree by
Jill Marissa Gjerde
has been approved
by
Li'S\'OS
Date


I
j
i
Gjerde, Jill (M.A., Psychology)
The Relationship Between Maternal Expectations for Premature Infant Health
Outcome and Mother-Infant Interaction
Thesis directed by Professor Peter Kaplan
ABSTRACT
This pilot study explored the relationship between maternal expectations for
premature infant health outcome and interactions with infant. Ten mother-infant dyads
were assessed longitudinally on measures of maternal expectations, quality of
interaction, and maternal mental health. The Parental Expectation Scale (PES),
developed specifically for this study, and the Caregiver Contribution Coding System
were examined for reliability and validity. Factor analysis o f the PES revealed the
PES is a valid and reliable measure for parental expectations in a NICU setting.
Results of the study indicate no relationship between expectations and interaction, but
that more optimistic expectations are significantly correlated with lower levels of
parental stress and depression, and better infant medical risk status. Long-term studies
with larger sample sizes are needed to examine if a relationship exists between
maternal expectations and mother-infant interaction in the NICU or post-discharge.
This abstract accurately represents the content of the candidates thesis. I recommend
its publication.
Signed
Peter Kapl5

iii


ACKNOWLEDGEMENT
This study was conducted with the financial support of grants from Newborn Hope,
Inc. and the Developmental Psychobiology Endowment Fund. Additional
recognition should be given to Aye let Talmi and the Center for Infant and Family
Interaction, The Childrens Hospital, Denver, Colorado, with whom this research
was conducted.


CONTENTS
Figures....................................................viii
Tables.....................................................ix
CHAPTER
1. INTRODUCTION..............................................1
Developmental Outcomes of Premature Infants............4
2. PARENT-INFANT INTERACTIONS................................6
Caregiver Contributions................................9
Adjusting Environment..........................12
Pacing and Timing..............................12
Continuity and Predictability..................13
Supporting Self-Regulation.....................13
Supportive Movement, Handling, and Positioning.14
Supports State Organization....................14
3. MATERNAL EXPECTATIONS....................................17
Do Expectations Affect Interaction?...................21
4. POTENTIAL INFLUENCES ON MATERNAL EXPECTATIONS...24
Maternal Mental Health................................24
Depression.....................................25
Parenting Stress...............................26
v


Socio-demographics.........................................28
Infant Medical Risk........................................28
Hypothesis.................................................29
5. METHODS........................................................31
Participants...............................................31
Procedure..................................................32
Measures...................................................35
Socio-demographic Information.......................35
Life Skills Progression......................35
Parent Infant Interaction...........................36
Caregiver Contribution Coding System.........36
Maternal Expectations...............................37
Parental Expectation Scale...................37
Maternal Mental Health..............................38
The Parental Stress Scale: Neonatal
Intensive Care Unit..........................38
Perinatal Posttraumatic Stress
Disorder Questionnaire.......................38
The Parenting Stress Index-Short Form........39
The Center for Epidemiologic
Studies-Depression...........................39
Infant Medical Risk.................................40
vi


Score for Neonatal Acute Physiology....40
The Neurobiologic Risk Score...............41
Statistical Analyses.....................................42
6. RESULTS.....................................................46
Data Cleaning and Reduction..............................46
Descriptive Statistics...................................47
Reliability and Validity of the PES and CC Coding System.49
Correlational Analyses...................................54
Case Studies.............................................58
Case Study A......................................58
Case Study B......................................60
7. DISCUSSION AND CONCLUSIONS..................................72
Limitations..............................................79
Future Directions........................................81
APPENDIX
A. Parental Expectation Scale...........................85
B. Caregiver Contribution Coding System.................86
C. Caregiver Contributions Observational Details........88
REFERENCES.....................................................89


FIGURES
Figure
6.1 Infant Medical Risk....................................................64
6.2 Parental Stress Index.................................................65
6.3 Center for Epidemiologic Studies Depression.........................66
6.4 Parental Expectation Scale............................................67
6.5 Caregiver Contributions...............................................68
6.6 Parental Stress Scale: NICU...........................................69
6.7 Perinatal Posttraumatic Stress Disorder Questionnaire.................70
6.8 Life Skills Progression...............................................71


TABLES
Table
5.1 Descriptive Statistics of Participant Demographics...................34
6.1 Means and Standard Deviations for Measures at Time 1 and Time 2......48
6.2 Parental Expectation Scale Time 1 Reliability Analysis..............50
6.3 Parental Expectation Scale Time 2 Reliability Analysis..............51
6.4 Correlations Between Measures at Time 1 and Time 2..................57


CHAPTER 1
INTRODUCTION
In the United States, an average of 9,246 babies are bom prematurely each
week (National Center for Health Statistics, 2004). According to the National Vital
Statistics Report (Martin et al., 2003) on prematurity, a preterm infant is one bom
prior to the 37th week of pregnancy. Of all infants bom in the United States in 2002,
12.1% were bom prematurely. The percentage of all babies bom preterm in the US
in 2002 increased by 14% from 1990 and 29% from 1981 (Martin et al., 2003).
Birth of multiples, illicit drug use, infection, bleeding, stress, smoking, maternal age
extremes (less than 15 years and 45 years and older), and a history of preterm
delivery are all major risk factors for preterm birth (Martin et al., 2003). According
to the national vital statistics reports for 2002, although the rate of infant mortality
in the US has declined in the last decade, one of the major contributors to infant
mortality remains prematurity/low birthweight (Martin et al., 2003).
Advancements in medical technology in the Neonatal Intensive Care Unit
(NICU) have improved the prognosis of premature infants over recent years
(Martin, Leverette, & Ferlauto, 1996). Improvements in treatment and technology
have led to longer stays in the NICU, and higher survival rates for preterm infants
(Bruns, McCollum, & Cohen-Addad, 1999). Longer stays occur in part as a result of
increased survival rates of babies bom earlier in gestation. Infants bom under 28
1


weeks from post-menstrual age were once considered unlikely to survive. Now
infants as young as 23 to 24 weeks gestational age and weighing as little as 500
grams, may survive (Goldberg & DiVitto, 1995). However, despite advances in
technology and a decrease in the rate of mortality among premature infants,
morbidity, including cognitive delays, behavioral problems, and
neurodevelopmental problems remain high (Melnyk et al., 2001).
Neonatal intensive care units, designed with modem medical technology, are
able to provide specialized care to premature infants. With an increase in survival
rate for premature infants, and longer hospital stays, parents are faced with coping
with the stress of the often unexpected early arrival of their baby, the hospital
experience, and the developmental and behavioral patterns of a preterm infant
(Goldberg & DiVitto, 1995).
For an infant in the NICU, approximately 2 hours of his or her day is
comprised of medical and nursing procedures. With the use of developmentally
supportive care techniques, interventions can be less disruptive for the infant,
increasing the chances that the infant will demonstrate organized behavior and
stable physiologic organization. Parents want and expect to establish a relationship
with their newborn (Goldberg & DiVitto, 1995). However, this is difficult to
accomplish due to the infants lack of responsiveness and critical medical needs of
the infant, which may limit the infants ability to interact and respond.
2


The current study focuses on how a mothers expectations for the health
outcome of her premature baby relate to the quality of the mother-infant interaction.
The quality of mother-infant interaction is a critical factor that leads to optimal
development. The following section explores how the developmental outcome of a
premature infant differs from that of a full-term-bom infant, and how the premature
infants course of development is dependent on the quality of mother-infant
interaction. While the mother-infant interaction has been studied in populations of
premature infants, there is little research on how a mothers expectations affect the
relationship. It is hypothesized that the poorer the expectations for the health
outcome of the premature infant, the less optimal the quality of mother-infant
interaction.
3


Developmental Outcomes of
Premature Infants
Outcomes of being bom prematurely are greatly influenced by infant,
parental, and demographic characteristics. An infants low birth weight, gestational
age, and medical complications pose an extra threat to the infants development
(Breslau & Chilcoat, 2000). Infants bom at a very low birth weight (< 1500 g) and
those bom prematurely have a higher rate of severe neurologic problems, physical
impairments, and cognitive problems, as well as increased rates of behavioral and
cognitive problems at school age (Hack et al., 1994; Martin et al., 1996).
Additionally, low birth weight infants are at an increased risk for a later diagnosis of
attention-deficit/hyperactivity disorder (Breslau & Chilcoat, 2000). Younger
gestational age at birth is related to a higher rate of medical complications and lower
scores on measures of mental and motor development (Miceli et al., 2000). Medical
events such as ventilation, seizure, and infection may do harm to brain tissue, thus
permanently altering development (Brazy, Goldstein, Oehler, Gustafson, &
Thompson, 1993). Other complications include visual impairments, mental
retardation, cerebral palsy, hydrocephalus, bronchopulmonary dysplasia,
intraventricular hemorrhage, and necrotizing enterocolitis (Lorenz, Wooliever,
Jetton, & Paneth, 1998; Martin et ah, 1996). Increased survival rates have resulted
in increased morbidity.
4


Long-term development can be negatively affected by risk factors to which
premature infants are exposed. Among these risk factors are the highly stressful
medical environment, parental mental health problems, and poor socioeconomic
status (Breslau & Chilcoat, 2000; Lorenz et al., 1998; Martin et al., 1996; Miceli et
al., 2000). Effects of a preterm birth can last into childhood, affecting the childs
physical, academic, and emotional development.
In summary, premature infants are at risk for ongoing and long-term
difficulties. Developmentally, they tend to have lower cognitive and motor skills
throughout childhood and problems associated with attention. In addition to risks of
premature birth itself, difficulties in the mother-infant relationship may be a
precursor for irregularities in developmental outcomes.
5


CHAPTER 2
PARENT-INFANT INTERACTIONS
Early mother-infant interactions provide the setting for the infants
development (Goldberg, 1978; Kelly & Barnard, 2000). Effective interactions occur
when an individuals cues are readable (clearly defined) and predictable (in order to
anticipate future behavior), and when the individual is able to respond to stimulation
(Goldberg, 1978). Interactions foster the development of emotion regulation, self-
efficacy, and social competence (Field, 1994; Goldberg, 1978). The mothers
contribution to reading her infants signals and providing optimal stimulation allow
the infant to become or remain physiologically and behaviorally organized (Browne,
MacLeod, & Smith-Sharp, 1996; Field, 1994), which is important for development.
If a mother is physically or emotionally unavailable for interaction, as she may be
during times of early separation or if she is depressed, she may not be able to
provide optimal stimulation for her child or be attentive to and respond
appropriately to infant cues. The mothers role thus plays an intricate part in
contributing to infant development (Field, 1994).
Given that an infants most salient learning environment is the time spent
with its primary caregiver, the interaction, or pattern of behavior between the
mother and infant is important for optimal development. How change in the
mothers or the infants behavior affects change in the dyadic interaction has not
6


J
I
f
I
I
I
i
i
been studied exhaustively. Most research has examined change in maternal, infant,
and social variables, but focused little on the change in the interaction. Research on
models of infant development suggests that a mothers interactive style is influenced
by her personality, social-contextual conditions, and the infants temperamental
dispositions (Feldman, Greenbaum, Mayes, & Erlich, 1997).
Mothers vary in how they respond to their childs communicative behaviors
(Tamis-LeMonda & Bomstein, 2002). Additionally, the level of attunement to the
childs behavior varies among mothers. A mothers responsiveness, defined as her
prompt, contingent, and appropriate response to her infant, can affect a childs
development, particularly development of language. According to researchers,
promptness refers to the timing of mothers replies vis-a-vis their childrens
overtures, contingency refers to the dependence of maternal reactions on child
behavior, (Tamis-LeMonda & Bomstein, 2002, p. 93), and appropriateness refers
to maternal replies that are conceptually and positively connected to the childs
behavior (p. 94). The terms responsiveness, reciprocity, mutuality, and sensitivity
have been used interchangeably in literature and represent similar concepts.
Coding systems that measure mothers reactions to their childs behavior
have been used for measuring maternal responsiveness. Tamis-LeMonda and
Bomstein (2002) used a coding system that counted the frequency of infant acts,
mothers responses to the acts, and the different types of maternal responses. The
7


Home Observation Measurement of the Environment (HOME) Inventory (Caldwell
& Bradley, 1984) measures parental responsivity within two domains: verbal
responsivity and emotional responsivity (Bradley, 1989). Bell and Ainsworth (1972)
studied maternal responsiveness to infant crying and found that the more responsive
a mother is to her infants crying, the less likely the infant will cry, and the more
likely the infant is to learn to communicate using a variety of modes.
Data on preterm infants indicates that mothers of preterm infants are less
likely to attempt to engage the infant in social play, and that preterm infants have
more social interaction difficulty with their mothers, when compared to full-term
infants and their mothers (Goldberg, 1979; Goldberg & DiVitto, 1995).
Additionally, preterm infants often show less organization, less alertness, less
responsiveness, and less positive affect in comparison with full-term infants
(Beckwith & Cohen, 1978; DiVitto & Goldberg, 1979; Field, 1977; Stevenson,
Roach, Ver Haeve, & Leavitt, 1990). The home environment provided by mothers
of preterm infants is also typically less responsive and stimulating than that
provided by mothers of full-term infants (Feldman, Eidelman, Sirota, & Weller,
2002). It may be that mothers react differently to preterm infants and stimulate their
infants less due to the lack of reciprocation during interaction. However, it may be
that the mothers expectations of how the baby will react produces a self-fulfilling
prophecy that perpetuates the infants lack of response.
8


The mother-infant relationship is dyadic, with each member actively
participating in the relationship. This being the case, the interaction between the two
is affected by a variety of factors that pertain to the individual, such as physical
health, mental health, sensitivity, and attentiveness/awareness, and to the dyad, such
as attachment, responsiveness, mutuality, and reciprocity. Understanding how these
factors affect the interaction will inform our understanding of how the interaction
then impacts the infants development.
Caregiver Contributions
Caregiver contributions are in essence the caregivers awareness and
responsiveness within the dyadic relationship. Relationships are based on the idea
that those involved respond to and affect one another (Browne et al., 1996).
Mutuality and connectedness are the key components of a relationship (Als, 1995)
in which the individuals work toward co-regulation, fostering optimal development
for both caregiver and infant (Browne et al., 1996).
A caregiver who is aware of her infants behavioral cues and able to focus
attention on her baby amidst distractions will be more likely to detect signs of infant
disorganization and readiness for interaction (Browne et al., 1996). This is key to
providing the appropriate response, so as not to overstimulate or understimulate the
9


infant, but rather provide optimal levels of stimulation and support according to the
infants needs (Browne et al., 1996; Kelly & Barnard, 2000).
The second main component of a caregivers contribution to the relationship
is responsiveness. How a caregiver responds is important so that an infants
development is supported, but not inhibited. Thus, supporting the infants
development as he or she matures can involve adjustments to the environment and
the interaction through pacing and timing, continuity and predictability, supportive
movement, handling, and positioning, support of the infants self-regulation, and
support of the infants state organization (Browne et al., 1996). Adapted from Als
and Gilkersons (1995) categorization of caregiver responses, Browne et al. (1996)
describes these six general caregiver contributions that are specific to the preterm
infant-parent relationship.
An infant is viewed as an active collaborator in his or her care, based on the
idea that an infants behavior is an indication of what works best to continue the
fetal development that began in utero (Als & Gilkerson, 1997). Thus, caregivers
should observe an infants behavior to discover what strategies the infant prefers to
use. According to relationship-based developmentally supportive care, caregivers
respect and promote the use of the infants strategies in order to foster development.
Behaviors to move away from or avoid stimuli can be considered stressful. When
10


these behaviors successfully remove or reduce the stress, they become strategies of
self-regulation (Als & Gilkerson, 1997).
Guidelines of The Newborn Individualized Developmental Care and
Assessment Program (NIDCAP), which can be used in the NICU, suggest pacing
caregiving based on an infants cues and communication, ensuring caregiving
consistency, structuring an infants daily schedule individually, supporting the
infant between caregiving activities, using feeding support, using skin-to-skin care,
collaborating on care, and planning discharge (Als & Gilkerson, 1997). Techniques
used in developmentally supportive caregiving involve altering the environment,
promoting containment and flexion through positioning aids, clustering caregiving
activities together to reduce interruption of sleep, promoting self-regulation,
recognition of readiness for oral feeds, and involving parents in caregiving activities
(Jacobs, Sokol, & Ohlsson, 2002).
The nature of studying caregiver contributions requires that one understand
the infants behaviors and responses to parental contributions in order to determine
the appropriateness of the caregivers behaviors. Infant behavior has been classified
into four systems: physiologic or autonomic, motoric, state, and interactive-
attentional and self regulatory (Als, 1982). Following are descriptions of the 6
caregiver contributions (as described by Browne et al., 1996) to be examined in the
current research.
11


Adjusting Environment
Adjustment of the environment includes alterating sound levels, visual
stimulation, and activity level. Als (1999) suggests that lighting be indirect and dim
because intense lights can be harmful and intrusive for premature infants. Other
adjustments to the environment (Als, 1999; Browne et al., 1996) include dampening
or eliminating extraneous sound in the NICU, covering incubators with thick
blankets, using rooms designed with sound-absorbent materials, moving
unnecessary conversations away from the bedside, and keeping monitor alarms,
overhead pagers, running sinks, telephones, and radios at an appropriate sound
level. Smells can also be noxious to an infant and strong perfumes, as well as the
odor of tobacco on a caregivers clothes, should be avoided (Als, 1999). Adjusting
the environment can aid the preterm infant in maintaining energy, thereby being
able to focus on an interaction.
Pacing and Timing
Throughout the interaction, a responsive caregiver takes breaks to allow the
infant time to recover and rest. Caregiving activities should be decided upon based
on the infants availability. Each infants state of alertness, sleep-wake cycle,
medical needs, and feeding ability should be considered in planning the appropriate
time for interventions. If an infant is alert and can tolerate several interventions
12


contiguously, then clustering caregiving activities is appropriate. Providing the
infant time to rest is one of the main goals of structuring an infants day (Als, 1999).
Continuity and Predictability
A consistent daily schedule and consistent caregivers who know the babys
cues will help the infant prepare for interaction. An infant will be more comfortable
with and trusting of social interactions if her surroundings, schedule, routines, and
caregivers are predictable (Browne et al., 1996).To better prepare the infant for
interaction a caregiver can approach the infant by gently touching the infant,
introducing oneself in a soft voice, and introducing new stimuli calmly and slowly
(Als, 1999).
Supporting Self-Regulation
Self-regulation pertains to the infants ability to organize and regulate his
physiologic, motor, and state systems as a means of comforting and calming himself
(Browne et al., 1996; VandenBerg, Browne, Perez, & Newstetter, 2003). To reduce
stress, infants often engage in sucking or mouthing, cover their face with their hand,
brace feet and legs, tuck their body, and clasp hands or feet together (Als &
Gilkerson, 1997; VandenBerg et al., 2003). Respiration rate and heart rate, as well
13


as time spent crying, are reduced by using pacifiers, indicating that physiological
stress for infants is reduced (Field, 1992).
To support self-regulation the caregiver can offer a finger to hold onto, a
pacifier to suck, or simply position the infant in a way that allows the infant to reach
its hands to its face or brace itself against a blanket or the caregivers body (Als,
1999). Hands-on containment, blanket rolls, swaddling, and nesting are all effective
ways to aid the infant in self-regulation.
Supportive Movement. Handling, and Positioning
Supportive handling is especially important during transitions between
wakefulness and sleep. As the infant moves between states or between interactions
the caregiver needs to be sensitive and supportive to ease the transition (Als, 1999).
Properly supporting a softly flexed and aligned position during care, movement, and
sleep is recommended for optimal developmental care. Meanwhile, the caregiver
should keep the infant contained and embraced close to their body to promote
organization (Als, 1999).
Supports State Organization
Preterm infants have more difficulty maintaining states of alertness than do
full-term infants and maternal responsiveness is often challenging because preterm
14


infants make fewer positive bids and are less socially active. Responsiveness is the
mothers behavior to signals from the infant. Prompt and appropriate behaviors are
components of responsiveness (Beckwith & Cohen, 1989). Due to the dyadic nature
of the mother-infant relationship, responsiveness of the mother often depends on the
behavior of the infant. Unfortunately, the mothers ability to interpret a premature
infants cues and interact appropriately is disturbed by the infants decreased ability
to focus and organize attention (Feldman et al., 2002). An infants limited ability to
mutually attend to his mother may affect the mothers response (Beckwith & Cohen,
1989).
Kangaroo care (skin-to-skin contact) has been noted to contribute to
improving the infants attention system, cognition, and physiologic maturation.
Additionally, mothers using KC may have better perceptions as they report having
more positive feelings toward their infant (Feldman et al., 2002).
Although a primary caregiver may be related or unrelated to the infant, the
majority of research on parent-infant interactions is done on the infant and its
mother. Therefore, the current research will be focused on the mothers role as
caregiver. Furthermore, the current research will use the Caregiver Contributions as
a means of studying the mother-infant interaction because the categorization of
contributions reflects the reciprocal nature of a relationship and is useful in a NICU
setting. In the interest of understanding how the quality of interactions between
15


mother and premature infant are affected, the next section explores maternal
expectations and how they may affect the mother-infant dyad.
16


CHAPTER 3
MATERNAL EXPECTATIONS
Studies have demonstrated that a mothers representation of her infant
affects the interactive behavior between mother and child (Keren, Feldman,
Eidelman, Sirota, & Lester, 2003; Stem & Karraker, 1990). Thus, a mothers
expectations may affect a childs development indirectly, by means of the mother-
infant interaction. In the current study the mothers expectation for the health
outcome of her premature infant will be explored in relation to the mother-infant
interaction.
Preconceived ideas about childbirth and typical child development are
expectations that many mothers hold. However, when the experience of childbirth
or development is disrupted, maternal expectations are impacted as well. Traditional
rituals of childbirth, such as the transition to parenting, giving the baby its first bath,
baby showers, and other normative changes associated with pregnancy are dismpted
(Talmi & Harmon, 2003). Thus, expectations about childbirth, parenting, and
development are violated (Macey, Harmon, & Easterbrooks, 1987). How these
expectations then contribute to or relate to the quality of interaction is of interest.
Although little research has been done on expectations relating to a preterm
population, several researchers have conducted studies on prematurity stereotypes
(Stem & Karraker, 1990), prematurity prejudice (Miller & Ottinger, 1986), and
17


maternal perceptions (Epps, 1993). Although these may all be related, it is important
to distinguish expectations from prematurity stereotypes and prematurity prejudice.
The prematurity stereotype is a preset conventional idea about premature infants and
a prematurity prejudice is an attitude about prematurity that is held in spite of
evidence to the contrary (Epps, 1993). Caregiving may be negatively affected by a
prematurity stereotype because perceptions, expectations, and behavior toward the
infant are affected (Stem & Karraker, 1990). In contrast, expectations are the
behaviors and outcomes a parent anticipates from their child. Perceptions are closely
related to expectations because they examine concerns a parent has related to the
infant or the infants circumstance (Simons, Ritchie, & Mullett, 1998). However,
expectations seems to be a better term because it implies that the parent is looking
toward the future and forming ideas of what will occur as the child develops. This
study will use the term expectations to define what a parent views will be the future
outcome for the health of their premature infant.
Experiments have been done using full-term infants to study the prematurity
stereotype. In the experiments, the same full-term infants are labeled as either
premature or full-term. Differences in adult reactions to the infants can be attributed
to preconceptions and expectations about the abilities of premature infants (Stem &
Karraker, 1990). In comparison with full-term-labeled (FTL) infants, premature-
labeled (PL) infants are rated as less sociable, less behaviorally active, less
18


physically developed, less cognitively competent, and less liked. (Stem &
Hildebrandt, 1984; Stem & Karraker, 1988). The negative set of beliefs that adults
may hold regarding premature infants can produce a self-fulfilling prophecy, in that
adults treat premature infants differently based on expectations (Stem & Karraker,
1990).
The birth of a premature infant has been compared with a crisis situation.
When the ideal baby is not bom, mothers have to revise their expectations
(Chatwin & MacArthur, 1993). The impact of giving birth to an infant prematurely
can create a sense of loss of control over caring for a child who requires specialized
interventions. Furthermore, expectations for the childs normal development are
altered and there is the uncertainty of morbidity and mortality (Epps, 1993).
A mothers perception of her infants severity of illness changes several days
after admission. While at the time of admission, a mothers perception is more
related to the infants birth weight, 2 to 5 days after admission, a mothers
perception is more related to actual medical severity (Simons et al., 1998). Parental
perceptions, defined by Simon et al. (1998), are (a) the number of medical diagnoses
reported by a parent, (b) whether a parent felt that neonatal diagnoses were affecting
their infant currently, and (c) whether a parent felt that neonatal diagnoses were
going to affect the infant in the future.
19


Some aspects of parental perceptions, as measured by Simons et al. (1998)
seem to tap into what may also be considered as expectations, especially as they
pertain to the future status of the infants health. Parental perceptions examined in
the study reveal a relationship between parental reports of current or future effects
of neonatal diagnosis and other worrisome perceptions concerning their infant. For
instance, in comparison to those parents who did not report current or future effects
of neonatal diagnosis, those parents who did report effects also tended to have fears
for their infant while in the hospital, reported that they imposed more restrictions on
their infant due to neonatal diagnosis, and rated the current health status of their
infant as less optimal (Simons et al., 1998).
A parents current perceptions about his or her child may affect expectations
and thus current perceptions may be an essential component to appreciating,
interpreting, and understanding expectations. Epps (1993) studied parental
perceptions of infant general health, attractiveness, behavior, and caregiving in
regards to preterm infants. Data analysis showed that providing parents with
information on the health status of the infant was the only variable that significantly
affected parental perception of infant behavior.
The Neonatal Perception Inventory (NPI) has been used to study mothers
perceptions of preterm infants. Using this instrument, studies have found that over
time, perceptions of preterm infants become more positive and at 1 month exceed
20


perceptions of full-term babies (Alfasi et al., 1995). Additionally, Stem, Karraker,
Sopko, and Norman (2000) determined that mothers of premature infants had fewer
stereotypes of premature-labeled infants, in comparison with mothers of full-term
infants. Although prematurity stereotypes may decrease over time and with
increased exposure to the infant, prolonged hospitalization limits the amount of time
parents have with their infant, thus increasing the likelihood that stereotypes
develop (Goldberg & Divitto, 1995). It could be hypothesized that just as
prematurity stereotypes may develop due to limited interaction, poorer expectations
of an infants health condition may also develop with minimal or limited interaction,
a common problem in a NICU setting.
Do Expectations Affect Interaction?
Based on the literature about expectations, it can be hypothesized that
expectations for the health outcome of a child impact a parents behaviors toward
and interaction with the child. Ones actions toward another are greatly affected by
expectations one holds about the others behavior or personality. A self-fulfilling
prophecy results when ones expectations influence perceptions in such a way that
confirmation of the expectations occurs. In relation to parents of premature infants,
parents form expectations and beliefs based on knowledge of the infants
prematurity, which then guide their behavior toward the infant. A confirmation of
21


the adults expectancies occurs over time as the adults behaviors begin to elicit the
expected behavior from the infant (Stem & Karraker, 1990).
Interestingly, Stem et al (2000) did not detect a difference between mothers
behaviors toward infants labeled as full-term and infants labeled as premature.
However, the researches did report that mothers of full-term infants tended to touch
the limbs of female babies labeled full-term more than they touched the limbs of
female babies labeled premature. Additionally, the infants themselves actually
behaved differently depending on their birth status label. Those infants labeled as
premature showed less positive emotion than full-term labeled babies.
If interruptions in the interactive process between mother and child occur
during hospitalization, it is more likely that maternal engagement with the child 3
months post-discharge will also be less-than-optimal (Keren et al., 2003). Moreover,
mothers with positive representations of their infants have more optimal interactions
with their infants than mothers with negative representations. Readiness for
motherhood, which taps readiness for discharge, belief in childs survival, and sense
of parenting, positively predict maternal positive touch. Withdrawal behavior, or
negative interactive behavior on the part of the infant, is predicted by negative
maternal representations. Previous research supports the notion that a mothers
expectations affect her behavior toward her preterm infant. However, there is a lack
of research on how specific types of expectations affect the interaction between a
22


mother and her preterm infant. The goal of the current study is to determine how the
mother-infant interaction is related to maternal expectations for the health outcome
of her premature infant. Based on the literature, it is hypothesized that poorer
expectations for the outcome of the infant will negatively affect the dyadic
relationship.
23


CHAPTER 4
POTENTIAL INFLUENCES ON MATERNAL
EXPECTATIONS
Other potential influences, such as maternal mental health,
sociodemographics, and infant medical risk may also impact how a mother forms
her expectations. Much research has been done on the effect of maternal mental
health, socioeconomic status, and infant medical risk on infant development and
quality of mother-infant interaction. However, these variables have not been studied
in conjunction with maternal expectations. It is unknown how or if these factors
contribute specifically to maternal expectations for the health outcome of her infant.
It is possible that these variables impact how a mother forms her expectations.
Further research should be done to determine where maternal mental health,
sociodemographics, and infant medical risk factor into the expectation-interaction
model. The current research will utilize case studies to explore how these variables
may relate to interaction and maternal expectations. Following is a brief description
of how these influences impact interaction.
Maternal Mental Health
Infants of mothers with a psychiatric illness tend to be more dysregulated
and have disturbed emotional functioning (Weinberg & Tronick, 1998). A mother
24


who is emotionally unavailable is more likely to be noncontingent and disruptive in
her attempts to stimulate her baby, thus interrupting the flow of interaction and
limiting the mutuality of sharing behavior (Field, 1994). Furthermore, maternal
distress (stress and depression) related to having a preterm birth appears to be
related to child behavior problems at 3 years of age, when the maternal distress is
present at 4 months postpartum (Miceli et al., 2000). Compared to mothers of
healthy full-term infants, mothers of premature infants are more likely to have
higher levels of depression, anxiety, hostility, and stress during the first year after
giving birth (Mew, Holditch-Davis, Belyea, Miles, & Fishel, 2003).
Depression
An infant learns about his or her environment from interaction with the
mother. If a mother is depressed and withdrawn, the infants learning about his
environment may be impaired (Kaplan, Bachorowski, Smoski, & Hudenko, 2002).
Mothers of preterm infants have a higher rate of postpartum depression than
mothers of full-term infants. Maternity blues, postpartum depression (PPD), and
postpartum psychosis are the most common postpartum mood disorders among
mothers (Mew et al., 2003). Depressed mothers often fall into two categories in
terms of behavior: intrusive (overstimulating) and withdrawn (understimulating)
(Jones et al., 1997). It has been proposed by Kaplan, Bachorowski, Smoski, and
25


Zinser (2001) that infant-directed speech is a form of stimulation that helps regulate
state and attention of the infant. Mothers with diagnosed depression have a smaller
change in fundamental frequency in their speech in comparison with nondepressed
mothers (Kaplan et al., 2001), suggesting that depressed mothers speech is
understimulating, characteristic of withdrawn mothers.
Premature infants have an increased need for stimulation and a decreased
level of responsiveness, which places them at greater risk than healthy full-term
infants for negative effects of maternal depression (Davis, Edwards, Mohay, &
Wollin, 2003). Unfortunately, mothers of preterm infants have a higher rate of
emotional distress than do mothers of healthy full-term infants within the first year
following birth (Mew et al., 2003). Little research has been done on maternal
depression and premature birth. Moreover, there is scarce literature on how maternal
depression affects the mothers interaction with her preterm infant. However,
evidence shows that maternal depression is associated with infant developmental,
cognitive, and emotional delay (Davis et al., 2003, Mew et al., 2003), and it can be
assumed that this would be true for both full-term and preterm populations.
Parenting Stress
Coping with a premature birth has been identified in studies as a source of
significant stress for mothers. Although the physical nature of the NICU is difficult
26


to adjust to, the appearance of a sick, fragile infant; disruption to the parental role;
course of hospitalization; loss of parental role; and separation from the infant during
hospitalization appear to be the greatest sources of stress for mothers of premature
infants (Davis et al., 2003; Hughes et al., 1994). Additionally, unmet expectations
for the typical development of their baby can contribute to parental stress (Goldberg
& DeVitto, 1995).
Disruptions in daily family routines and parent-infant interaction due to a
premature birth can cause psychological distress for parents. Across cultures,
families have some of the same goals for their childs development, including health
and protection. Fear and anxiety due to the infants medically intensive birth and
small size is a common feeling among parents of preterm infants. In a study of
Mexican immigrant families goals for their infants in the NICU (Denney et al.,
2001), all of the parents made positive and reassuring comments about their infants
current condition. Gaining weight, being taken off of life-support machines, and
other milestones were among the hopeful and positive comments made by parents
about their infants progress. Despite feelings of fear, parents were able to recognize
progress and capabilities of their babies.
Being unprepared for a premature birth can be stressful while concern for the
infants medical complications and separation from the infant can increase the stress
(Mew et al., 2003). During the neonatal period, compared to mothers of full-term
27


infants, mothers of preterm infants suffer more severe levels of psychological
distress (Singer, Salvator, Guo, Collin, Lilien, & Baley, 1999).
Socio-demographics
Socioeconomic status (SES) has shown to be one environmental factor that
is linked to neurodevelopmental outcome among infants. Associated with lower
SES are a number of factors such as neglect and abuse, which may mediate poor
cognitive outcomes among children (Martin et al., 1996). Poor economic and social
conditions can create a strain on family members, fostering a feeling of insecurity.
Often a parents interaction with their child is impaired by a parents perception of
lack of control over their life and their childs life. Both mother and well-child and
mother and ill-child are adversely affected by an impoverished environment, and
their interaction is even more impaired when the child is in a state of illness
(McCarthy et al., 2000).
Infant Medical Risk
For seriously ill infants (i.e., infants suffering a life-threatening illness for at
least seven consecutive days), the number of days of prolonged medical
complications is related to the amount of interaction with the mother. For example,
mothers interact significantly less with infants who suffer prolonged medical
28


complications for over 35 days versus those infants who have suffered 17 or fewer
days (Minde, Whitelaw, Brown, & Fitzhardinge, 1983). The degree of illness of a
premature infant appears to have an impact on how the infants and their mothers
interact at 4 and 8 months postpartum. Of the premature infants studied by Jarvis,
Myers & Creasey (1989), the sickest infants scored lowest on measures of mother-
infant reciprocity during interaction. Mothers of sicker babies seem to have more
difficulty adjusting their activities according to the infants cues (Jarvis et al., 1989).
While maternal mental health, socio-demographics, and infant medical risk
all appear to affect the mother-infant interaction, it is hypothesized that they may
also impact expectations. Further research should be done on these variables and
their influence in the expectation-interaction model. As mentioned earlier, we will
use case studies to describe relationships between expectations and the
aforementioned covariates.
Hypothesis
As the number of preterm births in the US increases, it is critical to promote
healthy development among those infants who are bom prematurely. The mother-
infant interactional pattern has been noted as one contributor to optimal
development of a baby. Therefore, the goal of the current study is to determine how
the mother-infant interaction is affected by maternal variables, specifically maternal
29


t
i
i


expectation for the health outcome of her premature infant. We hypothesize that a
positive relationship exists between maternal expectation for health outcome of her
premature baby and the quality of mother-infant interaction. In other words, mothers
with more positive expectations for the health outcome of their baby will also have
more optimal interaction with their infant.
30


CHAPTER 5
METHODS
Participants
Ten mothers and their premature infants participated in this pilot study.
Inclusion criteria for the mothers required that they be mothers of first-born
singletons and 18 years of age or older. Mothers and infants were excluded from the
study if the mother reported a history of psychiatric illness or substance abuse or
dependence, had other children, or had given birth to multiples. The dyad was also
excluded if the infant had grade III or IV intraventricular hemorrhages (bleeding in
or around the brain), congenital anomalies (malformation of a part of the body), or
past or anticipated major surgery. The reason to exclude infants for medical status
reasons is that we want to study a relatively healthy preterm population. By doing
so, we are more confident that interactions, maternal expectations, maternal mental
health, and socio-demographic variables are not largely affected by the degree of the
infants illness severity. Inclusion criteria for infants required that they be bom
between 28 and 33 weeks gestational age and have a birth weight appropriate for
gestational age.
31


Procedure
Participants were recruited from two large level II and level III NICUs in the
Metro Denver Area. A level IINICU is a special care nursery for babies who may
need monitoring, intravenous tubes or needles, an incubator, help with maintaining
body temperature, and feeding assistance. A level III NICU is one in which care
may include the use of a ventilator or other respiratory assistance, monitors for
blood pressure, and close observation. An infants condition in a level III NICU is
more medically unstable than an infant in a level II NICU. All mothers and infants
who met the inclusion criteria were approached by a General Clinical Research
Center (GCRC) nurse within 7 days after admission to the hospital. After consenting
to participate, the mothers were asked to complete a battery of self-report measures.
A researcher read the measures to the mothers while the mothers recorded their own
responses on a separate set of questionnaires. Finally, mothers and infants were
videotaped together for 20-30 minutes during an interaction of feeding, bathing,
socialization, medical care, or other routine caregiving. The infants medical record
was also reviewed and assessed for infant medical risk. All participants met with a
researcher at two separate time points. The first was done as soon as possible after
recruitment, and the second was done when the baby was 35 weeks gestational age.
At both time points the mothers completed the self-report measures, were
videotaped during an interaction with their infant, and received $25 compensation.
32


In addition, the family received a toy and a copy of the videotaped interaction for
participating.
Mothers who participated ranged in age from 19 to 32 years old (M= 25.30).
Of the mothers, 9% were Caucasian and 1 % was classified as Multiracial (one
mother was Caucasian and Filipino). Of the infants, 9% were classified as
Caucasian and 1 % was classified as Multiracial (one infant was Hispanic and
Caucasian). Table 5.1 displays the descriptive statistics of participant demographics.
33


Table 5.1
Descriptive Statistics of Participant Demographics
Variable / Percent
Infant gender
Male 5 50
Female 5 50
Infant ethnicity
Caucasian/Non Hispanic 9 90
Multiracial 1 10
Maternal ethnicity
Caucasian/Non Hispanic 9 90
Multiracial 1 10
Annual family income
$0-$10,000 2 20
$10,000-$25,000 2 20
$25,000-$50,000 3 30
$50,000-$ 100,000 2 20
>$100,000 1 10
Maternal Age
34


Measures
The measures selected for this study, which cover the areas of socio-
demographic information, parental expectations, mother-infant interaction, maternal
mental health, and infant medical risk, provide multi-method, multi-source data.
While the maternal mental health, socio-demographic, and parental expectation
measures are self-report, the mother-infant interaction piece is gathered via
observational ratings, and the infant medical risk scores are gathered from medical
records and physiological scores. Thus, the data gathered on all of the participants
comes from a variety of sources, enhancing its credibility.
Socio-Demographic Information
Life Skills Progression. Demographic information was collected using a
brief demographic questionnaire and The Life Skill Progression (LSP; Wolleson,
1999) measure. Items included in the demographic questionnaire were completed
for both mother and father, as reported by the mother. Demographics collected
were: infant gender; age of parents; household composition; total number of people
living in household; infant, mother, and father e thnicity; and annual income. The
Life Skills Progression (Wolleson, 1999) is a measure that assesses relationships
with friends/family, relationships with child(ren), education, health and medical
care, mental health and substance use, & adequacy of basic essentials. Reliability of
35


the LSP has been established with a coefficient alpha of .81 at baseline, and .84 at
follow-up. Both predictive and construct validity of the LSP have been established.
Parent-Infant Interaction
Caregiver Contribution Coding System. The Caregiver Contributions Coding
System (Talmi & Browne, 2003) was designed to assess preterm infant-parent
interactions in the NICU. The Coding System has 6 domains pertaining to parent-
infant interactions. A mothers awareness of and responsiveness to her infant is
measured within the 6 domains: (1) adjusts environment; (2) pacing and timing; (3)
continuity and predictability; (4) supports self-regulation; (5) supportive movement,
positioning, and handling; and (6) supports state organization. The Coding System
was used to analyze videotaped observations of mother-infant interaction in the
NICU. The coding system has been used in other studies and validated on a sample
of 60 mother-infant feeding interactions in the NICU (Talmi, Browne, Gjerde, &
Lampe, 2004). The system is sensitive to variability in the mother-infant interaction,
within the domains, which had coefficient alphas ranging from .71 to .94. Domain
scores are also significantly correlated with another observational coding system
(NCAFS) that is well-validated. Inter-rater reliability of trained coders was above
90% in a preliminary study used to establish validity and reliability of the system
(Talmi et al., 2004).
36


Maternal Expectations
Parental Expectation Scale. The Parental Expectation Scale was developed
specifically for this study. No other rating scale exists for assessing maternal
expectations for the health outcome of her child. Therefore, the scale was designed
to have high face validity. A self-report measure, the scale consists of 10 items
based on a 4 point likert scale, ranging from 1 {not at all true) to 4 {very true).
Questions include: I am worried about whether my infant will survive, I expect
that my babys development will be good, and I know what the future holds for
my babys development. The Clinical Interview for Parents of High-Risk Infants
(CLIP; Keren et al., 2003) and measures of parental perception of medical condition
and diagnosis (Simons et al., 1998) were consulted for current perception of health
status (i.e., I am happy with how my child is doing medically, I believe my baby
is healthy.) and measures of parental perception of medical condition and
expectations of neonatal diagnosis (Simons et al., 1998) were consulted for items of
expectation of outcome (i.e., I expect that my babys development will be good,
I am worried about whether my infant will survive, my infant is ready for
discharge from the hospital). Preliminary reliability and validity was evaluated
with this study. Theoretically, the scale should have been positively correlated with
mother-infant interaction, such that more optimistic or positive expectations would
be correlated with more optimal interactions.
37


!
!
!
!
S
i


Maternal Mental Health
Maternal mental health was measured using 4 measures: the Parenting Stress
Index, short form (PSI-SF; Abidin, 1995), the Parental Stress Scale:NICU (PSS-
NICU; Miles, Funk, & Carlson, 1993), the Perinatal Posttraumatic Stress Disorder
Questionnaire (PPQ; DeMeir et al., 1996; Hynan, 1998), and The Center for
Epidemiologic Studies Depression (CES-D; Radloff, 1977).
The Parental Stress Scale: Neonatal Intensive Care Unit. The Parental Stress
Scale: Neonatal Intensive Care Unit (PSS: NICU; Miles et al., 1993) assesses
parental perceptions of stress during the hospitalization of their infant in the NICU
and takes approximately 10 minutes to complete. Items, which tap into stresses
arising from the physical and psychosocial environment of the NICU and are rated
on a 5-point likert scale ranging from 1 {not at all stressful) to 5 {extremely
stressful) include: the presence of monitors and equipment, and when my baby
seemed to be in pain. Cronbachs alpha coefficients are acceptable (> .70) for all
scales. Scores on the PSS: NICU scales are significantly correlated {p < .001 and p
< .05) with scores of State Anxiety (Miles, et al., 1993).
Perinatal Posttraumatic Stress Disorder Questionnaire. The Perinatal PTSD
Questionnaire (PPQ; DeMeir et al., 1996; Hynan, 1998) is a 14-item dichotomously
scored instrument that measures symptoms of PTSD relating to childbirth and takes
approximately 5 minutes to complete. Responses are in the form of yes or no
38


and items include: Have several bad dreams of giving birth or of your babys
hospital stay, and Unable to remember parts of your babys hospital stay. Test-
retest reliability (r =.92) and internal consistency (a = .83) are high. Validity is also
high, as the PPQ is significantly correlated with other well-validated measures of
PTSD (DeMier, Hynan, Hatfield, Varner, Harris, & Manniello, 2000; Quinnell &
Hynan, 1999).
The Parenting Stress Index-Short Form. The Parenting Stress Index-Short
Form (PSI-SF; Abidin, 1995) is a 36-item questionnaire that measures stress in a
parent-child system and takes approximately 10 minutes to administer. Items are
rated on a 5-point scale ranging from Strongly Agree to Strongly Disagree, and
include: I dont enjoy things as I used to, and I expected to have closer and
warmer feelings for my child than I do and this bothers me. The test-retest and
internal reliability coefficients for the PSI-SF are .84 and .91 respectively for total
stress scores. The validity of the PSI-SF has not been established independently,
however, it has demonstrated high correlation with the full scale PSI (.94 for total
stress) (Abidin, 1995).
The Center for Epidemiologic Studies Depression. The Center for
Epidemiologic Studies Depression (CES-D; Radloff, 1977) is a 20-item self-report
scale, taking approximately 5 minutes to complete, that assesses the frequency of
depressive symptoms, both somatic and affective over the past week (Posner,
39


Stewart, Marin, & Perez-Stable, 2001; Carpenter et al., 1998). Responses are
categorized into one of four options for each item: Rarely or none (0), Some or a
little (1), Occasionally (2), or Most or all of the time (3). Score totals range from 0
to 60, with a higher score indicating a higher degree of depression (Radloff, 1977).
Items include: During the past week, I was bothered by things that usually dont
bother me, and During the past week, I felt that everything I did was an effort.
Adequate reliability of the scale (a = .83 .92) has been demonstrated (Carpenter et
al., 1998).
Infant Medical Risk
Infant medical risk while in the NICU was assessed using The Score of
Neonatal Acute Physiology (SNAP II; Richardson, Gray, McCormick, Workman, &
Goldmann, 1993), and The Neurobiologic Risk Score (NBRS; Brazy, Eckerman,
Oehler, Goldstein, & ORand, 1991).
Score for Neonatal Acute Physiology. The Score for Neonatal Acute
Physiology (SNAP; Richardson et al., 1993) is a measure of severity of illness and
risk at time of admission. It takes between 5 and 15 minutes to complete and is to be
used with newborns in a neonatal intensive care setting, typically within the first 24
hours of birth. The SNAP is used to assess each organ system for the worst possible
physiologic derangement. Therefore, infants with higher SNAP scores are sicker
40


and have higher rates of mortality. Items measured by the SNAP include mean
blood pressure, white blood count, oxygenation index, and potassium level, among
others (Richardson et al., 1993). There are five risk categories into which infants can
be classified: low risk except for low probability catastrophic event, 5.3 (4.4);
mildly ill, still at small risk, 8.4 (4.5); moderately ill, but excellent chance for
survival, 12.2 (5.0); extremely ill, but with good chance for survival, 16.2
(5.8); and virtually certain death, now or delayed, 23.5 (5.4). The SNAP is
highly correlated with other measures of illness severity (r ranging from .59-.78)
and is capable of distinguishing infants with 2- to 20-fold differences in mortality
(Richardson et al., 1993).
The Neurobiologic Risk Score. The Neurobiologic Risk Score (NBRS;
Brazy et al., 1991) is an assessment scoring system used for very low birth weight
infants that quantifies the impact of neonatal medical events on developmental
outcome (Brazy, Goldstein, Oehler, Gustafson, & Thompson, 1993). Based on 13
items measuring brain cell injury, items are scored as 0 (no risk), 1, 2, or 4
(depending on the degree of risk). The more severe and the longer the duration of
the symptoms are, the higher the score. A total score of <4 is considered low risk, 5-
7 is intermediate risk, and a score >8 is considered high risk (Brazy et al., 1993).
Items measured by the NBRS include Apgar score (score given immediately after
birth to indicate a babys physical condition), ventilation, blood pH, and
41


j
i


intraventricular hemorrhage. The sum of all 13 items is the total NBRS score (Brazy
et al., 1991). Studies have demonstrated high correlations of the NBRS with other
outcome measures of cognitive performance, motor performance, and neurologic
examination (p < .0001) (Brazy et al., 1993).
Statistical Analyses
Statistical analyses used for this research included evaluating the reliability
and validity of the Parental Expectation Scale and the reliability of the Caregiver
Contributions coding system, examining the relationship between maternal
expectations and mother-infant interaction and the relationship of covariates:
maternal mental health, socio-demographics, and infant medical risk to the
expectation-interaction model, and case studies.
As part of establishing the internal consistency of the Parental Expectation
Scale (PES), a reliability analysis of the items of the PES was used to measure the
internal consistency of items at each time point and correlating scores from time 1
and time 2. Coefficient alphas were calculated separately for items that measure
current perceptions and for items that measure future expectations.
To establish validity, the theoretically derived items of the Parental
Expectation Scale were examined in relation to items of other measures.
Theoretically the PES should be related to other measures used in the study. Using
42


the sample mean for the PES, a confidence interval for our population mean (p) was
calculated using a = .05. Therefore the probability of making a Type I error is .05
and we can be 95% confident that the interval contains the p. Reliability was
measured by means of coefficient alphas at time point 1 and time point 2.
Additionally, due to the nature of the study, the Parental Expectation Scale was
administered at two points in time, providing opportunity to run test-retest
reliability. However, it was expected that scores on the scale would change
depending on the amount of lapsed time between administrations, change in medical
status of the baby, and additional factors, such as maternal mental health,
environmental stress (financial stress, prolonged time away from home, etc), and a
decreased support system, that may affect a change in expectations. For instance,
improvement in the medical condition of the infant may predict a mothers change
to more optimistic and positive expectations. Maternal mental health,
sociodemographic status, and infant risk were selected as covariates when
examining time one and time two total scores.
To examine convergent and divergent validity of the PES, correlational
analyses were conducted. Convergent validity was expected to be high if the scale
correlated with other valid measures with which we expected it to be correlated.
Because expectations should theoretically be related to depression and stress, scores
on the Perinatal Stress Scale: NICU measure and the Center for Epidemiologic
43


Studies Depression were expected to correlate with the expectation scale.
Additionally, we predicted that scores on the expectation scale would correlate
positively or negatively with scores on particular items or indices of other measures,
depending on the degree of optimism measured by the other indices.
As an example, low scores on the expectation scale (revealing negative or
low expectations) were expected to be positively correlated with items on the
Parenting Stress Index, such as, It takes a long time and it is very hard for my child
to get used to new things, My child is not able to do as much as I expected, My
child doesnt seem to learn as quickly as most children, and My child doesnt
seem to smile as much as most children.
Discriminant validity is also important because the expectation scale should
not tap constructs other than expectations. So, while we predicted that the
expectation scale would be correlated with the Parental Distress and the Parent-
Child Dysfunctional Interaction subscales, we expected expectations to be less
correlated with the Difficult Child subscale of the Parenting Stress Index.
Furthermore, we expected that the expectation scale would be more highly
correlated with the Infant Appearance/Behavior and Parental Role Alteration
subscales than the Sights and Sounds subscale of the PSS: NICU.
The result of the reliability and validity analysis ensured the use of a
consistent and empirically constructed set of items. The reliability of the Caregiver
44


Contributions (CC) coding system was also evaluated by examining correlations
between items at each time point and across time points.
The Pearson product-moment correlation coefficient was used to determine
the direction and magnitude of a linear relationship between expectations and
interaction. Mean scores, standard deviations, and ranges in scores were calculated
for mothers expectations and for mother-infant interaction.
Case studies were used to illustrate examples of how maternal mental health
and socio-demographic status may affect maternal expectations, and thus mother-
infant interaction. Due to the limited sample size, we could only speculate about the
potential influence of maternal mental health on expectations. Based on the case
examples we were able to formulate hypotheses for future research. Case studies
may not be useful for statistical analysis (Hersen, 1990) and may lack objectivity
(Kazdin, 2003), but the advantage is that case studies contribute to and complement
research, theory, and practice by providing valuable and unique information
(Kazdin, 2003). Contributing to the formulation of ideas and hypotheses, case
studies are very useful for developing therapy techniques and encouraging different
outlooks on problems or on universal beliefs (Kazdin, 2003; Yates, 1970). Case
studies do not necessarily provide causal knowledge, but their ability to persuade
and motivate is a powerful tool for demonstrating a point or inspiring action
(Kazdin, 2003).
45


CHAPTER 6
RESULTS
Data cleaning and reduction were conducted and descriptive statistics on all
measures were run. Information gathered from the 10 mother-infant dyads was used
to assess the reliability and validity of the PES and Caregiver Contributions coding
system. To determine the relationship between maternal expectations on infant
health outcome and mother-infant interaction Pearson product-moment correlations
were calculated. A correlational analysis with maternal mental health, socio-
demographics, and infant medical risk as covariates was also conducted. A
regression analysis was conducted to examine the proportion of variance that
measures of maternal mental health and infant health risk account for in scores of
maternal expectations. Case studies were also developed.
Data Cleaning and Data Reduction
Prior to analysis, specific items on the Parental Expectation Scale and Center
for Epidemiologic Studies Depression were recoded so that higher scores on the
PES indicate more optimism about infant health, while higher scores on the CES-D
indicate higher levels of depression. Four items on the PES were reverse scored: I
am worried about whether my infant will survive, I am disappointed with my
infants current medical status, I expect that my babys development will be
46


poor, and My infants medical status changes day-to-day. Additionally, the
Parenting Stress Index and Perinatal Posttraumatic Stress Disorder Questionnaire
were constructed so that higher scores indicate higher levels of parental stress and
post-traumatic stress, respectively. Furthermore, higher scores on the Parental Stress
Scale: NICU indicate higher levels of stress related to the NICU setting.
Descriptive Statistics
The Parental Expectation Scale average score was 2.89 (SD = .45) at time 1.
At time 2 the average score was 3.49 (SD = .49). The Caregiver Contributions
coding system mean score at time 1 was 3.71 (SD = .56). At time 2, the coding
system had a mean score of 3.85 (SD = .59). Due to the small sample size, the
coding system was not divided into domains, but rather, a total mean score for the
measure was created. Means and standard deviations of all instruments are
displayed in Table 6.1.
!
i
47


Table 6.1
Means and Standard Deviations for Measures at Time 1 and Time 2
Measure M SD
PSI Total Score T1 63.10 17.20
PSI Total Score T2 62.88 11.22
CES Total Score T1 38.10 8.93
CES Total Score T2 31.25 5.82
PSS:NICU Total Score T1 107.50 20.38
PSS:NICU Total Score T2 117.38 20.67
PPQ Total Score T1 18.20 3.79
PPQ Total Score T2 16.75 2.31
PES Mean Score T1 2.89 .45
PES Mean Score T2 3.49 .49
CC Mean Score T1 3.71 .56
CC Mean Score T2 3.85 .59
LSP Total Score 121.20 10.15
SNAP Total Score 10.67 7.50
NBRS Total Score 1.63 1.50
Note. T1 = Time 1; T2 = Time 2.
48


Reliability and Validity of the PES and CC Coding System
The Parental Expectation Scale was developed for the purpose of this study.
Internal consistency analyses of the PES at time 1 (Chronbachs Alpha = .75)
indicated that item number 10 (my infants medical status changes day-to-day)
could be eliminated for better inter-item reliability (Chronbachs Alpha = .77).
When item number 10 was eliminated the mean item score became 2.97 (SD = .49)
at time 1. At time 2, however, item number 10 correlated more highly with the other
items than it did at time 1. Internal consistency analysis of the PES at time 2 yielded
a stronger alpha of .84. If item 10 were still deleted at time 2 the alpha would
become .80. As such, a decision was made to retain all 10 items of the PES and
construct a mean scale score. Item means, standard deviations, and alphas if the item
were deleted are shown in Table 6.2 and 6.3 for time points 1 and 2, respectively.
Test-retest reliability was also evaluated for the PES. Correlation between the PES
at time 1 and time 2 displayed a positive trend (r = .633, p = .092).
49


Table 6.2
Parental Expectation Scale Time 1 Reliability Analysis
Item M SD Alpha if item deleted
Happy Medically 3.70 .48 .73
Survival 2.90 1.45 .71
Dissap Current Med 3.30 .82 .73
Healthy 3.70 .48 .72
Develop Good 3.80 .42 .74
Develop Poor 3.90 .32 .76
Infant Ready for Disch 1.30 .67 .72
Parent Ready for Disch 1.70 1.06 .66
Future Develop 2.40 .97 .74
Med Status Changes 2.20 .79 .77
50


Table 6.3
Parental Expectation Scale Time 2 Reliability Analysis
Item M SD Alpha if item deleted
Happy Medically 3.88 .35 .82
Survival 3.75 .46 .81
Dissap Current Med 3.75 .71 .89
Healthy 3.75 .46 .83
Develop Good 3.88 .35 .82
Develop Poor 4.00 .00
Infant Ready for Disch 2.63 1.19 .78
Parent Ready for Dish 2.88 1.25 .78
Future Develop 2.75 .89 .82
Med Status Changes 3.63 1.06 .80
Note. The following component variable has zero variance and
is removed from the scale: develop poor.
It was hypothesized that specific items on the PES would better tap into
current perceptions, while others tap more into future expectations. Reliability
analysis on items in each of these domains revealed that the items categorized as
current perceptions held together better than the items categorized as future
expectations. The current perception items had a Chronbachs alpha of .69 at time
51


1 and an alpha of .75 at time 2. At both time points, the alpha would be greater if the
item my babys medical status changes day-to-day was eliminated. The future
expectation items were less internally consistent related and at time 1 had an alpha
of .39 and at time 2 an alpha of .65. This may be affected by the difference in scores
from time 1 to time 2 on the discharge items.
Further item analysis of the Parental Expectation Scale involved calculating
the Pearson product moment correlations for items of the PES in relation to specific
items on other measures, in order to establish validity of the PES. As proposed,
items from the Parental Stress Index (for which higher scores reflect higher levels of
stress) were compared to items of the Parental Expectation Scale (for which higher
scores indicate higher levels of optimism). Results indicate a significant correlation
at time 1 between the parental stress item, my child doesnt seem to learn as
quickly as most children, and the parental expectation item, I expect that my
babys development will be good (r = .662,p = .037) at the .05 level. A trend, atp
< .10, also exists between the same parental stress item, my child doesnt seem to
learn as quickly as most children, and the parental expectation item, I believe my
baby is healthy (r = .604,/? = .064). These results did not support our hypotheses.
We expected that the items would be negatively correlated, but instead the
correlations suggest that more optimistic expectations are related to higher levels of
stress related to infant learning. Other items of the Parental Stress Index that were
52


compared were, it takes a long time and it is very hard for my child to get used to
new things, my child is not able to do as much as I expected, and my child
doesnt seem to smile as much as most children. Although there were no other
significant correlations, it appears that the majority of the findings are in the
expected direction. For instance, when the items of the Parenting Stress Index
indicate higher levels of stress, the items on the Parental Expectation Scale are
typically showing lower expectations.
Compared with the subscales of the Parenting Stress Index, it was expected
that maternal expectations should be more highly correlated with the Parental-
Distress and Parent-Child Dysfunctional Interaction subscales than with the
Difficult Child subscale. Results show that scores on the Parental Expectation Scale
at time 2 were significantly negatively correlated with the Parent-Child
Dysfunctional Interaction subscale (r = -.732, p = .039) and the Difficult Child
subscale (r = -.809,p = .015) at the .05 level. In other words, more optimistic
expectations were significantly related to lower amounts of stress pertaining to
parent-child interaction and difficult children. Comparing the PES to subscales of
the Parental Stress Scale: NICU, as proposed, yielded no significant results.
However, the trend was toward higher correlation between the subscales of Infant
Appearance/Behavior and Parental Role and maternal expectations than between the
subscale of Sights and Sounds and expectations.
53


Reliability analysis of the Caregiver Contributions coding system yielded a
Chronbachs alpha of .95 at time 1. At time 2, Chronbachs alpha was .97. The
author was the primary coder of the interaction video tapes. The author and another
researcher, both previously trained to reliability on the coding system, coded 15% of
the tapes to establish inter-rater reliability. Inter-rater reliability of coders was
greater than 94% on the tapes that were double coded. Test-retest reliability on the
coding system revealed a trend toward statistical significance (r = .639, p = .088).
Correlational Analyses
Scale scores for all measures were correlated at time 1, time 2, and across
the time periods with the mean scores for the Parental Expectation Scale and the
Caregiver Contributions coding system (Table 6.4). Negative correlations were
expected between maternal expectations and scores on measures of maternal mental
health and infant medical risk because higher scores on the Parental Expectation
Scale indicate more optimism, while higher scores on measures of maternal mental
health and infant medical risk indicate higher levels of symptomatology and infant
risk. Time 2 expectation scores were significantly negatively correlated with
parenting stress scores at time 1 (r = -.756; p = .030) and depression scores at time 2
(r = -.777; p = .023) at the .05 level. Additionally time 2 scores on the PES
significantly negatively correlated with the Neurobiologic Risk Score (r = -.963,/? =
54


.00)) at the .01 level. Similarly, scores on the Caregiver Contributions coding
system were expected to be negatively correlated with measures of maternal mental
health and positively correlated with scores on the Life Skills Progression. Scores
on the Caregiver Contributions show significant negative correlations at time 2 with
time 1 scores of parenting stress (r = -.110, p = .049) and significant positive
correlations with Life Skills Progression scores (r = .119, p = .045) at the .05 level.
Time 1 expectation scores displayed a trend with scores on the Life Skills
Progression (r = -.562, p = .091). Time 2 expectation scores also show a trend with
Time 2 parenting stress scores {r = -.665, p = .072), Parental Stress Scale: NICU
scores at time 2(r- -.670, p = .069), and PTSD scores at time 2(r = -.661, p =
.075). Time 1 Caregiver Contribution scores show a trend with Time 2 depression
scores (r = -.104, p = .051). Table 6.4 displays the correlation coefficients for time 1
and time 2 and across time points.
When the covariates maternal mental health, infant medical risk, and socio-
demographics were partialled out, the correlations between expectations and parent-
infant interaction did not change. A hierarchical regression was performed to test
the effects of depression, stress, PTSD, and infant medical risk on maternal
expectations. Depression accounted for 58% of the variability in expectations (ft =
-.759, p = .048). After depression was controlled for, the only other factor that
55


accounted for a significant portion of variance in maternal expectations was infant
medical risk (/? = -.905,/? = .05).
56


Table 6.4
Correlations Between Measures at Time 1 and Time 2
Measure PEST1 CCT1 PES T2 CC T2
PSIT1 -.370 .127 -.756* -.710*
CES-D T1 -.475 .231 -.462 -.037
PSS:NICU T1 -.469 .504 -.263 .416
PPQT1 -.451 .212 -.677 .137
PEST1 - -.079 .633 -.038
CCT1 -.079 - .465 .639
PSIT2 -.467 -.238 -.665 -.159
CES-D T2 -.492 -.704 -.777* -.202
PSS:NICU T2 -.621 .034 -.670 .274
PPQ T2 -.451 -.441 -.661 .041
PES T2 .633 .465 - .181
CC T2 -.038 .639 .181 -
LSP -.562 .475 -.006 .719*
SNAP -.292 -.196 -.638 .115
NBRS -.668 -.154 -.963** .190
Note p < .05. ** p< .01.
57


Case Studies
Case Study A
Infant A (Case 105) was a white male infant bom at 28 weeks gestational
age with a birth weight of 1310 grams. A researcher met with the infant and his
mother when the baby was 28 weeks and 6 days gestational age and again at 35
weeks and 3 days. Birth was vaginal and infant A was initially doing well with
spontaneous respiratory effort. In the delivery room infant A was on a mask of
continuous positive airway pressure (CPAP). He required CPAP for the first 7 days
of his life. During the first 11 days of life, Infant A had increased abdominal
distension and poor stooling for which he received barium enema. Thereafter, his
feeds were given in full, which he tolerated well. Additionally, he advanced to
breast milk and grew appropriately. Over the course of his hospital stay he received
a 7-day course of epoetin injections for anemia and caffeine for apnea of
prematurity. Following birth he was treated with ampicillin and gentamicin for 7
days for pneumonia and presumed sepsis. Further blood cultures remained negative.
An echocardiogram on day three showed a large patent ductus arteriosus (PDA) and
dilated left ventricle and atrium. Following treatment with Indocin he was
asymptomatic. During the first 24 hours of life, a SNAP score of 8 was given, which
is considered mildly ill, still at small risk. At discharge, the NBRS score was 1,
58


indicating low risk. Both of these health risk scores were below the average for
babies in this study, indicating that this baby was a healthy baby relative to the
sample as a whole.
The mother reported strong social support and that the pregnancy was
planned and welcomed. At the time one visit with the researcher, the father and
extended family members were present. Infant As mother did endorse intermittent
substance use (excluding nicotine) during pregnancy. The mother considered her
employment seasonal and the familys annual income fell in the range of $10,000-
$25,000, which is just below the average income of the participants. This mother
was also the oldest in the sample of mothers. In relation to the other mothers, Infant
As mother fell below the mean on measures of stress at both time 1 and time 2. On
the depression measure Mother A was at least one standard deviation below the
mean at both time points. Additionally, she had the lowest score on the depression
measure at time 2 and the second lowest score at time 1. Her scores on the post-
traumatic stress measure were also below average at both time points. Her PES
scores were above the mean at both time points, indicating that she had more
optimism about her babys health outcome, compared to the average mother.
Additionally, her expectations increased from time 1 to time 2. Her scores on the
Caregiver Contributions were also above the mean at both time points, indicating
59


that she responded appropriately to her babys cues more consistently than the
average mother in the study.
Infant A and his mother exemplify what optimal mother-infant interaction
and optimistic expectations look like. This dyad had support of a father/spouse,
extended family, and hospital staff, who trained the mother in caregiving
techniques. From verbalizations during the videotaped interaction we know that the
mother learned from nursing staff about Kangaroo Care and the mother practiced,
enjoyed, and believed in the effectiveness of the developmental technique that is
beneficial to both mother and child. Other verbalizations of the mother during
interactions demonstrate her understanding and awareness of Infant As needs: He
just kinda does this until he gets tired, about 15 minutes, He didnt want to do the
other side. No, hes out, even though his eyes are open. Mother A allowed her
infant time to respond to interaction, and timed interactions according to the babys
needs. She also was consistent in noticing signs of her baby becoming overwhelmed
and attempted to return him to an organized state. The mother was very optimistic
about her childs health outcome and expected that his development would be good.
Case Study B
Infant B (Case 109) was a Caucasian female bom at 30 weeks gestational
age and weighing 1500 grams. The researcher met with the mother-infant dyad
60


when the baby was 31 weeks and 3 days and again at 34 weeks and 6 days age.
Infant B was delivered by cesarean section. Following delivery the infant had
spontaneous respirations and was placed on CPAP. The baby was transferred to
room air on day 3 of life. Medical course was complicated by an intermittent heart
murmur and presumed sepsis. The baby received a dose of saline for hypertension
and tachycardia (i.e. increased heart rate) and was placed on ampicillin and
gentamicin for 48 hours for presumed sepsis. Although anemic following birth, the
hematocrit count was reassuring by day 8. Infant B also had phototherapy on day 5
and day 12. The baby was still on caffeine and had not been discharged at the time
of this publication. The SNAP score for this baby was a 7 following birth, meaning
mildly ill, still at small risk. This babys health risk score indicated that this baby
was a healthy baby relative to the sample as a whole.
Infant Bs mother was 21 years old, one of the youngest mothers
participating. She reported that she had some casual friends and that her family and
husband are supportive, however she also was glad that by being in the hospital she
did not have to share her child with extended family whom she did not feel close
and warm feelings toward. The pregnancy was unplanned, but accepted. The mother
felt that she had limited knowledge about a childs development and felt that her
connection with Infant B was inconsistent. Mother B reported that her employment
meets expenses most of the time, but that her job provides a low income. This
61


mother has been and is currently treated for depression and has a mild learning
disability.
In contrast to other mothers, Mother B was above the mean on a measure of
parental stress at time 1 and time 2. However, her stress levels on a measure of
stress related to being in the NICU were lower than average at time 1 and 2. Her
scores on the depression measure were above the mean, which can be expected with
her history of reported depression. On the measure of post-traumatic stress disorder,
Mother B was above the average at both time points. Mother Bs expectations for
the health outcome of her daughter were also lower than the average at both time
points, but did increase over time. Additionally, her mean scores on the measure of
mother-infant interaction showed a lower than average quality of interaction.
Furthermore, the mixed feelings about social support, young age, and the
presence of depression and a learning disorder could have impacted the mothers
expectations or interaction. This mothers understanding of her babys cues was
inconsistent or nonexistent. She did not attend to the babys needs or support state
organization. The way she moved the baby was more jerky than smooth or gradual.
Although the mother noticed signs of her babys disorganization, she did not seem
to accurately assess the babys efforts at self-regulation: why are you yawning so
much? Mother B was also concerned about whether her infant would survive and
reported that she does not know at all what the future holds for her babys
62


development. This mother-infant dyad is a good example of how someone with low
expectations and a poor quality of interaction looks.
These case studies were selected for their contrasting characteristics. Not
only were the mothers expectations and quality of interactions opposing, but age,
level of social support from extended family, and maternal mental health were
dissimilar. Mother A was older, sought information and practiced developmental
techniques taught by nurses, and had no history of mental illness. Mother B, on the
other hand, was younger, had little background on child development, did not seek
assistance from nurses, and felt less of a bond with her infant. One similarity,
however, is that both infants had the same level of infant medical risk. Therefore,
these cases provide a nice contrast of variation in maternal behavior and
symptomatology when infant medical risk is controlled for. These case studies
provide good qualitative examples of how differences in socio-demographics and
maternal mental health may impact or moderate expectations and interactions.
Graphs 6.1-6.8 display scores for each measure across cases.
63


Figure 6.1 Infant Medical Risk
Infant Medical Risk
NBRS Total
Score
SNAP Total
Score
64


Score
!
i
Figure 6.2 Parental Stress Index
Parental Stress Index
PSI Total Score
Time 1
PSI Total Score
Time 2
i
65


Figure 6.3 Center for Epidemiologic Studies Depression
j
Center for Epidemiologic Studies Depression
!
i
i
i
60.00
50.00
£
O 40.00
W
30.00
20.00
/
\
I I I I I I I I I I
101.00 I 103.00 I 105.001 107.001 109.00 I
102.00 104.00 106.00 108.00 110.00
Score ID
CES Total
Score Time 1
CES Total
Score Time 2
i
66


Figure 6.4 Parental Expectation Scale
Parental Expectation Scale
PES Mean
Score Time 1
PES Mean
Score Time 2
67


Figure 6.5 Caregiver Contributions
Caregiver Contributions
CC Mean
Score Time 1
CC Mean
Score Time 2
68


Figure 6.6 Parental Stress Scale: NICU
Parental Stress Scale:NICU
i
*
i
PSS Total
Score Time 1
PSS Total
Score Time 2
69
1


Figure 6.7 Perinatal Posttraumatic Stress Disorder Questionnaire
Perinatal Posttraumatic Stress Disorder Questionnaire
!
i
i
i
!
26.00-
24.00-
22.00-
22
g 20.00-
V)
18.00-
16.00-
14.00-

I I I I I
101.00 I 103.001 105.00 I 107.001 109.00 I
102.00 104.00 106.00 108.00 110.00
Case ID
PPQ Total
Score Time 1
PPQ Total
Score Time 2
i
i
70


LSP Total Score
Figure 6.8 Life Skills Progression
Life Skills Progression
71


CHAPTER 7
DISCUSSION AND CONCLUSIONS
This study on maternal expectation for premature infant health outcome and
mother-infant interaction examined the reliability and validity of a new measure for
assessing parental expectations as well as the relationship between expectations and
interaction. Factors that contribute to the expectation-interaction model were also
studied: maternal mental health, socio-demographics, and infant medical risk. There
is little research on how a mothers expectations for the health outcome of her
premature infant relate to her interaction with her baby. An infants development is
impacted by its interaction with its mother or primary caregiver (Feldman et al.,
1997; Field, 1994). Thus, studying the quality of interaction and the variables that
affect the quality of interaction is important for promoting and providing optimal
developmental care for a fragile population.
Results of the preliminary analyses indicate that the Parental Expectation
Scale is a valid and reliable measure. To establish internal consistency we looked at
correlations between expectation items at a single time point and across time points.
Results revealed that one item was problematic at time 1: my infants medical
status changes day-to-day. This item was not eliminated from the scale because it
was correlated more highly with other items at time 2. A preterm infants medical
condition is more volatile following birth which could be why a parent would
72


respond to item 10 differently at time 1 than at time 2 when the infant is more stable
and closer to discharge.
Test-retest results showed that the expectation items about discharge: my
infant is ready for discharge from the hospital and I feel ready for my infant to be
discharged from the hospital are noticeably lower at time 1 than at time 2.
However, the scores on the 2 discharge items are expected to change over time, as
the infant nears discharge and the parent gains a better understanding of what to
expect for the future. The low correlation may be explained by this change in
expectations over time. Furthermore, the change in scores on the PES over time is
not unusual considering that as the baby improves, expectations should also become
more optimistic. This is reflected in the increase of mean scores on the PES from
time 1 to time 2.
Test-retest results of the caregiver contribution coding system showed that
interactions between infants and their mothers also change over time. Mean scores
on the quality of interaction increased from time 1 to time 2. The improvement in
quality of interaction may be accounted for by the timing at which the interactions
were captured. Time 1 interactions were captured within the first week of the
infants birth. Typically the capability for interaction with an infant is limited during
the first few days because infants are medically fragile, parents are just becoming
acquainted with their babies and the environment, and medical staff tend to take
73


primary responsibility in caring for the babies. Therefore, the majority of
interactions captured at time 1 consist of holding or touching the infant while the
infant is sleeping or being fed intravenously. At time 2 the parents have taken over
more of the caregiving routines and the baby is more developed and able to interact
with the mother. Improvements in the quality of interaction may be due to the fact
that mothers are able to get to know their baby better over time or the mother feels
more comfortable interacting with the child.
Levels of depressive symptoms changed from time 1 to time 2, decreasing
over time. It may be that mothers display more symptoms of depression
immediately following the birth of their infant, when their expectations are
disrupted. As the mother adjusts to and accepts the situation and becomes more
confident in her childs health status, her depressive symptoms may decrease. Risk
factors for developing postpartum depression include stressful life events,
obstetrical complications, not breast-feeding, difficult infant temperament, and an
unwanted/unplanned pregnancy (Beck, 2003; Yonkers et al., 2001). As each of
these factors changes and improves for the mother over the course of the babys
hospital stay, her depressive symptoms may also decrease.
To establish construct validity of the Parental Expectations Scale an item
analysis of the scale was conducted. Items from the expectation scale were expected
to relate to items from other measures, and thus these items were compared. Only
74


one significant correlation was found between items on the PES and items on other
measures. The items that were significantly correlated were the parental stress item,
my child doesnt seem to learn as quickly as most children and the parental
expectation item, I expect my babys development will be good. The same
parental stress item showed a trend toward significance with the parental
expectation item I believe my baby is healthy. This correlation may indicate a
parents optimism in spite of adverse circumstances. Thus, parents are still proud
and hopeful for their infant even if they are concerned and worried that their child
will not be able to compete or develop at the same level as other children. Possibly,
those parents who are more stressed are more concerned for their infants
development and also more hopeful that their childs development will improve
with time. Furthermore, the expectation item, I believe my baby is healthy may be
perceived by parents as a question about the babys physical/medical status, while
the parental stress item about learning may be perceived as a question about the
babys mental status. Physical condition and mental status are two separate
constructs which may explain why parents would respond differently to questions
pertaining to health and learning.
Two of the subscales of the Parenting Stress Index were significantly
correlated with the Parental Expectation Scale, showing that parents who feel a
bond with their child and believe that their child is behaving normally also tend to
75


have more optimistic expectations. Similarly, parents who feel that their child is
more challenging than other children also tend to have lower expectations for their
infants health outcome. The Parental Stress subscale of the Parenting Stress Index
was not significantly correlated with expectations, which is understandable because
the Parental Distress subscale is related to a mothers conflict with her partner,
confidence related to child-rearing, and how other life roles are restricted due to the
stress of parenting. The Difficult Child subscale taps into parental feelings about
child noncompliance, temperament, and demandingness. The Parent-Child
Dysfunctional Interaction subscale assesses a parents perception that the child does
not meet expectations and that interactions with the child are not reinforcing
(Reitman, Currier, & Stickle, 2002, p. 387).
As anticipated, more optimistic parental expectations appear to be associated
with lower levels of parental stress, depression, and infant medical risk. Contrary to
our expectations, at the beginning of the infants hospital stay, a parents more
optimistic expectations show a trend toward significant correlation with lower
scores on the Life Skills Progression measure. One possibility is that those parents
who have less knowledge of development, less social support, and less education do
not know what to expect for their childs development, and thus their expectations
are optimistic because they have little or nothing with which to compare. Another
possibility is that the newness of the experience leads parents to be more positive or
76


optimistic and as they spend time in the NICU and learn more about the condition of
their baby, their expectations also shift. More optimistic expectations also tend to
occur with lower levels of stress in the NICU setting, and fewer symptoms of post-
traumatic stress disorder.
It was hypothesized that maternal expectations would be related to mother-
infant interaction. This hypothesis was not confirmed. It may be the case that
expectations and mother-infant interaction are not related. Perhaps a mother can put
aside her expectations for her infants medical status when she is interacting with
her infant. Another possible explanation is that those areas that the expectation scale
does not tap, such as information on what medical diagnoses were communicated to
the mother, what experience the mother has with child rearing, or what type of
coping skills the mother employs, may account for the lack of correlation between
expectations and interaction Furthermore, the Parental Expectation Scale does not
gather information concerning what the mother was prepared for prior to giving
birth. It is possible that she was aware she would be giving birth prematurely and
was informed of what to expect for the babys medical status.
Using hierarchical regression, results show that the depression accounted for
a significant amount of variance in parental expectations. After depression scores
were controlled for, only infant medical risk added a significant proportion of
variance accounted for in expectations. The other variables added into the regression
77


model included scores on the PSS: NICU, PSI, and PPQ. These results suggest that
expectations at time point 2 are related to depression and infant medical risk at time
point 2. With a larger sample size we may be able to run multiple regressions that
would show what effect expectations have on levels of depression and other
measures of maternal mental health.
Using case studies, we hoped to demonstrate how depression or anxiety
affect a mothers ability to interact with her infant and to form accurate expectations
of her infants health outcome. Linking maternal mental health and
sociodemographic status to expectations or mother-infant interaction may provide
ideas for future research and revision of conceptualization of expectations. In
addition, case studies may be useful in demonstrating how expectations which are
presumably accurate (as measured by correlation between expectations and infant
medical risk) correlate with interactions.
Due to the limited sample size it is helpful to provide case studies of what
our data might look like if we had more subjects. The case studies provided
examples of those mothers whose expectations may affect interaction. They
represent mothers with optimistic expectations/high quality interactions and mothers
with low expectations/low quality interactions. Rather than throwing out small
sample size data because of lack of significant findings, it is useful to examine
78


i
individual cases and generate new hypotheses (Shih, Ohman-Stricklan, & Lin,
2004).
Limitations
A major limitation of the present study was its small sample size. With a
small sample size, it is difficult to run statistical analyses on data. However, when
small effect sizes are anticipated, as in a pilot study, it is necessary to rely on
statistical analyses. Typically, data from research with small sample sizes are
presented graphically and visual analysis, the most common form of analysis for
small sample studies, is used (Bengali & Ottenbacher, 1998). The graphs provided
were intended to aid in visual analysis of trends in data. An interesting finding in
this study, detected through visual analysis or examination of individual scores, is
that although the total mean score on the Parenting Stress Index from time 1 to time
2 indicates a decrease in scores, only 2 cases actually decreased in their levels of
stress. The other participants showed small increases in their levels of stress from
time 1 to time 2, suggesting that stress in a mother-infant relationship actually
increases with time spent in the NICU. Additionally, stress may increase as the
mother prepares to bring her baby home and become responsible for caring for a
fragile baby.
\
79


Due to the high number of correlations conducted, a limitation of this study
is the potential for error in the correlations. To account for the high number of
correlations computed, planned comparisons or Bonferroni post-hoc comparisons
could be used in future analyses to correct for the potential for correlations to be by
chance alone.
Using line graphs to plot data from small sample studies allows one to easily
monitor, recognize, and display data, and examine the extent and consistency of
change over time (Bengali & Ottenbacher, 1998; Kazdin, 2003). Conducting pilot
studies may not yield results that can be generalized as well as results from a group-
study design, but pilot studies are beneficial for identifying new processes and
theories (Kazdin, 2003). As mentioned previously, a strength of this research is that
several sources (i.e., hospital records, self-report measures, observation) of
information were used to gather data on each of our subjects. A multi-method study
with multiple time-points is very useful for determining effects/relationships within
a small sample size.
Not knowing what medical diagnoses were communicated to the mothers is
yet another limitation. Although the relationship between expectations and infant
medical risk was negative (higher expectations related to lower levels of risk) we do
not know what information the mothers actually had or how well they understood
the information concerning their infants status. Although degree of optimism in
80


expectations may be a better predictor of interaction, accuracy of expectations may
play a mediating role in quality of interactions.
Future Directions
In order to gain a better understanding of whether expectations are related to
mother-infant interaction a long-term follow-up study may be advantageous.
Although this study was longitudinal, a long-term follow-up study that compares the
mothers expectations in the NICU and post-discharge with restrictive behavior
could be valuable. After the transition home has occurred it may be helpful to see
how expectations and interactions carry out. For instance, at 12 months post-
discharge we would hope to find that higher levels of optimistic expectations would
lead to higher levels of quality interactions, and thus a better developmental
outcome for the baby.
This pilot study provides a better understanding of how expectations of a
premature infants health outcome may be related to the mothers mental health and
the babys medical risk. As maternal expectations of infant health outcome have
rarely been studied, this pilot study is valuable for guiding future studies concerning
factors that affect, are affected by, or are related to maternal expectations. Studies
with small sample sizes, such as a pilot study, are not necessarily conducted to
81


confirm hypotheses or determine the average effect, but are designed to explore any
effect or relationship (Shih et al., 2004).
As the results show, depression and infant medical risk are significant
correlates of maternal expectations. Thus, the information that medical professionals
convey to mothers is an important component in affecting the expectations a mother
forms. The PES may be valuable as a tool to identify mothers with expectations that
are incongruent with infant medical risk. Determining what differentiates these
mothers from those who have accurate expectations may be helpful for
understanding coping mechanisms. Additionally, it may be useful to study whether
expectations in the NICU persist post-discharge. Practitioners may want to follow-
up on mothers who have expectations that are incongruent with their infants
medical risk because low expectations that persist beyond discharge could be
harmful for the childs development if the parents behaviors toward the child are
affected by the expectations.
Future studies of expectations should include finding out what medical
information is communicated to parents and how well they understand the
information. Additional study should be done to determine if discriminant
expectations affect interaction in the NICU or over time. Change in expectations
over time should also be evaluated as we know from this study that maternal
expectations for infant health outcome increase over time spent in the NICU.
82


Whether these expectations continue to change post-discharge is yet to be studied.
McCarthy et al. (2000) should be consulted for further reading on comparison of
mother and physician ratings of infant medical status, and the impact of less reliable
judgment on mother-child interaction. Furthermore, studying what knowledge
mothers have of development would aid in understanding how mothers form their
expectations and with what they compare their childs status.
Changes to the expectation scale may include reformatting it as an interview,
similar to the CLIP (Keren et al., 2003). Conducting an interview may provide more
qualitative data and explanations for why the parent holds particular expectations.
This would also be a great way to gather information on what mothers
know/understand about their childs diagnosis. A benefit to using the PES over the
CLIP is that because it is a self-report measure it is less time-consuming and does
not require training for administration. However, incorporating an interview format
for at least part of the PES may be beneficial in future studies. If a longitudinal
study is conducted it would also be useful to compare the PES with the Parental
Questionnaire about Medical Diagnoses and Perceptions (Simons et al., 1998),
which is a measure used to assess parental perceptions about medical diagnoses
post-discharge.
As the percentage of premature births in the United States increases (Martin
et al., 2003), it is essential that research continue to explore and determine how to
83


promote optimal development among premature infants and how to best meet the
needs of infants and their caregivers. The goal of this research was to examine the
relationship between maternal expectations for premature infant health outcome and
mother-infant interaction, in the hope that by examining these variables, we can
better understand how to foster optimal development among premature infants. This
research provides a means for assessing maternal expectations of premature infant
health outcome and demonstrates that infant medical risk and maternal depression
relate to a mothers expectations. Moreover, this research indicates that there is a
need to further study how expectations are affected and the role expectations play in
affecting premature infant development.
i
84


APPENDIX A
PARENTAL EXPECTATION SCALE
ID_
Dale
Parental Expectation Scale
Please read each statement and circle the number that best describes how true each statement is for you.
Not Not really Somewhat Very
at all true true true
true
1. I am happy with how mv child is doing medically. 1
2. J am worried about whether my infant will survive. 1
3. I am disappointed with mv infant's current 1
medical status.
4. I believe my baby is healthy. 1
5. 1 expect that my babys development will be good. 1
6. I expect that my babys development will be poor. 1
7. My infant is ready for discharge from the hospital. 1
8. I feel ready for my infant to be discharged 1
. from the hospital. ,
9. I lenow what the future holds for my 1
babys development.
10. My infants medical status changes day-to-day. 1
3
4
4
4
4
4
4
4
4
4
85


APPENDIX B
CAREGIVER CONTRIBUTION CODING SYSTEM
Study ID_____
Today's Date Coder Initials Interaction Length (minutes)
Filename and Path ________________
Location: G Hospital u Home 0 Other_________________________
People present: Mother Father Sibling(s) Grandparent(s) Other____________________________
Type of Interaction: D Feeding Social Play Routine caregiving (e.g., diaper change) Bathing
Other
Domain 1 i i 2 1 3 I 4 5 | NA
A'cAJ ^ \ ,.V T'_V'Adjusts Environment. Sn "JmS
1. environment quiet and soothing
2. adjusts.environmentto support infant: v-;.rt r.- :y~-y; .-'.'V ; V. . y S : i!..
3. verbalizes awareness-of environmental influence and infant, 'restxmsev:::SSiS->r^ Vn-Y'G
4. limits new environment when possible i
5. adjusts several;domains;simuttaneousiy orcfT3nges;'G'i- environmeritSmorethanoncejf needed;m'-iy. ' rb&m
6. Environment prepared for interaction 1
5tei|3 liSriSlS?
7.:provides:necessitY interventions considering:infant needs = -IV::;:;'.:
8. allows infant time to respond to interactions and attends to infant responses to interactions
; 9.pausesdnteractions:to allowinfant to fecbverwhenJnfant'?::; -showsbegfnning disorganization orsigrisof being'U j -':;;V;:.=;- overwhelmed iftl ssst SiSS
10. times and paces interactions accordinq to infants needs
11. removes support gradually and leaves when infant demonstrates phvsioloqic, motor, and state stability
: 12l:maintains contact after interventions ar^compfeted untrtT^:- '.stabirizatjdri'is-achieved andVddesnot-moyelo'nex^ intervention until infant is ready^STvc'AsVy ill ig
and Predictability, ? rpi-tSr'C i--=r
13. demonstrates awareness of infant's likes and dislikes and infant's responses to aspects of caregiving and the environment
. T4Mitn'rts caregivers.to aTewprediciabjeindividuals who will; :: intersct With infant Mnsistently^sSfyTv^'SvvSvU^ T'S'ySSyy , /..-.v'.yV:-
15. provides consistency across caregivers Dy navmg them use similar strategies and approaches
'16, develops and uses a;consisterrt daily.schedule based on the Infants needsarid availability'-C/V St;'-
17. establishes and uses caregiving routines to enhance consistency
n8! advocates:for adherepce.to.schedule and appropriate: .interactions with infant.SS-S-S'- rv-/'S, Sv-: Y SStS mo
19. prepares infant for interactions by using voice or touch prior to intervening
li^fe^liSS^Supports^lPRegulatSoriSSfeg^SwSS: 1
20. verbalizes and/or notices infant's efforts to self-requlate
21;-accurately assesses infants efforts atself-requlation,-.-- "
22. understands environmental influences on infants self- regulatory behaviors
.23. notices fhat.fatique decreases infant's ability to use self-! ; | ' S.'S
86


Study ID
Domain 1 1 2 3 4 I 5 NA
reoulation strategies j'y-, i i: >V ; -1,-;:
24. positions infant to promote infants use of self-regulation strategies Dy using caregiver's body and environmental supports
25. effectively uses body and.environmentafaids.to support infant'sself-requlafibrr ": 1 . .V - 7 i'ir
-7:7;;r'SuppbrbVeTnovcrTienti ha nd liiiq/a ndposifumfriq.:,
26. maintains infant in a tucKea position wtth arms and legs close to body during movement, handling and careaiving
27. movesinfant slowfy-and^smoothly-without.abmpt;-;\^ movemenisychanges position slowly
28. moves infant as a whole with booy, nead ana extremities alipned
.29. uses own body._or blankets to^upporlinfant during ;y .movement r/ Jlv-~ C j'!
30. allows infant time to respond to changes in movement position.
31..moves in rby^m with the infant allowing the.infant Urne to be activer;^':-r^^;rt- :--2 yil
32. positions infant on back or in sidelying position with support to infants back when lying down and/or supports infants head when sitting up and/or legs and arms positioned forward and to midline
a^a^SuRports,StaMPraamMW^^-i^S.Tl b
33. verbalizes or notices infants level of alertness including transitions between states
34.\beginsjnleradions wFth;(hfanf anly.when.appropnate state., :is v
35. modifies environment, positioning, and interactions to support infants transitions between states
36. interads appropnatefy with infantduring caregiving.- ; >7 depending: on infarTts ayailability rather tbanron-caregi^fs: ;^--- maddmr availability:.:^;^ ;My; ^:0y
37. verbalizes infant competencies or engages in conversation with infant
38. playfully interacts with infanLand/or.routinely smiles at on1- in response to infant- - 7J:;
39. recognizes and responds to infant distress, pain, irritability by changing interaction to meet infanf's needs
40. offers visual input according to infants slate
41. engages with infant when infant is alert and demostrales availability for interaction.
42. notices signs of baby disiress/becoming overwhelmed and attempts to return baby to an organized state.
87


APPENDIX C
CAREGIVER CONTRIBUTIONS OBSERVATIONAL DETAILS
i
!
Caregiver Contributions to Infant Organization
Observation Details
Ayelet Talmt. Ph D & Joy V Browne. Ph D RN
(Adapted from FIRSTS. Browne, MacLeod, & Smith-Sharp. 19d(>)
Instructions. Use tins table to record observations while viewing videotapes. Observations include irifanl signs ol oigani/atiun and
disorganization in the physiologic, motor, and slate domains, caregiver behaviors, and specific caiecjiviny contributions Please* complete
the Coding Sheet and the Domain Item Coding Sheet for each observation
Study ID Today s Date______________Coder Initials ______________Filename and Path
Baby Behaviors
Physiologic
Caregiver Behaviors
Caregiving Contributions
Adjusts Enviionmenl
Pacing and 1 irning
Continuity and Predict.ihilily
j Suppnds Sell Regulation
j Uses Suppotlive Movement, Positioning, and
! Handling
; Supports State Oigani.alii >i
88