Citation
Latent factor structure of the BDI-II at 4 months postpartum

Material Information

Title:
Latent factor structure of the BDI-II at 4 months postpartum
Creator:
Asherin, Ryan M. ( author )
Place of Publication:
Denver, CO
Publisher:
University of Colorado Denver
Publication Date:
Language:
English
Physical Description:
1 electrnic file (46 pages). : ;

Thesis/Dissertation Information

Degree:
Master's ( Master of Arts)
Degree Grantor:
University of Colorado Denver
Degree Divisions:
Department of Psychology, CU Denver
Degree Disciplines:
Clinical health psychology

Subjects

Subjects / Keywords:
Postpartum depression ( lcsh )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis (M.A.)--University of Colorado Denver. Clinical health psychology
Bibliography:
Includes bibliographic references.
System Details:
System requirements: Adobe Reader.
General Note:
Department of Psychology
Statement of Responsibility:
by Ryan M. Asherin.

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:
898044025 ( OCLC )
ocn898044025

Downloads

This item has the following downloads:


Full Text
LATENT FACTOR STRUCTURE OF THE BDI-II AT 4 MONTHS POSTPARTUM
by
RYAN M. ASHERIN
B.A., Metropolitan State University of Denver, 2004
A thesis submitted to the Faculty of the Graduate School of the University of Colorado in partial fulfillment of the requirements for the degree of Master of Arts
Clinical Health Psychology Program
2014


This thesis for the Masters of Arts degree by Ryan M. Asherin has been approved for the Clinical Health Psychology Program by
Kevin Everhart, Chair
Peter Kaplan Krista Ranby


Asherin, Ryan M. (M. A., Clinical Health Psychology)
Latent Factor Structure of the BDI-II at 4 Months Postpartum Thesis directed by Professor Kevin Everhart.
ABSTRACT
BACKGROUND: Depression in the postpartum period (DPP) is the most common psychiatric disorder associated with childbirth, negatively impacting the mother-infant dyad. The manifestation of DPP symptoms has been attributed to a combination of biopsychosocial stressors. Current standard of care recommends routine screening for maternal mood disorders at infant well-child checks through the first four months of life, and the Beck Depression Inventory, 2nd Edition (BDI-II) has been validated for use in the postpartum period to assess maternal depressive symptoms. The present study tests a theoretically driven model emphasizing the role of intrapersonal cognitions (cognitive-appraisal) on the report of DPP as compared to somatic symptoms (somatic-exhaustion) and interpersonal (impairment) factors.
METHODS: A sample (N=100) of mothers four-months postpartum were assessed for depressive symptoms using the BDI-II and a Structured Clinical Interview for DSM-IV Axis I Disorders (SCID). Confirmatory factor analysis was conducted to identify the latent factor structure of the BDI-II, and then a path analysis was conducted to test the hypotheses that a cognitive-appraisal latent factor on the BDI-II drives elevated scores and correlates with a SCID diagnosis of depression.
RESULTS: A confirmatory factor analysis supported a three-factor model of the BDI-II with cognitive-appraisal, somatic-exhaustion, and impairment latent factors identified (x2(i86)=211.386, ns; CFI=0.947; RMSEA=0.037). A path analysis supports the proposed hypothesis based on good model fit estimates (x2(204)=231.9 1 7, ns; CFI=0.945; RMSEA=0.037), and the proposed influence of cognitive-appraisal on somatic-exhaustion (P = .754) and impairment (P = .481) was supported. Additionally, cognitive-appraisal demonstrated the highest correlation (r = .663) with SCID diagnosis, compared to either somatic-exhaustion (r = .458) or impairment (r = .012).
CONCLUSIONS: This study supports a theoretical model emphasizing the role of cognitive-appraisal in the development of DPP, suggesting that elevated responses to cognitive-appraisal related items on the BDI-II might warrant clinical assessment of DPP.
The form and content of this abstract are approved. I recommend its publication.
m
Approved: Kevin Everhart


DEDICATION
I dedicate this work to my wife and parents. Jess: you are my biggest supporter and best friend. Your love, strength, and support have given me the motivation to continue to pursue my career goals while building a beautiful life together. Thank you for your unconditional love and understanding over the years. My graduate school experience has been so much more fulfilling because I get to share it with you. Mom and Dad: I have spent a lot of time and energy throughout my academic career trying to gain a better understanding of infant and childhood development. I still have a lot to learn, but I am certain that none of the achievements throughout my life would have been possible without your love, support, and prayers. Thank you for all that you have given me, especially the confidence to believe in myself and never stop pursuing my dreams.
IV


ACKNOWLEDGMENTS
I would like to thank the Infant Lab at the University of Colorado Denver, especially my mentors, Drs. Kevin Everhart and Peter Kaplan. I have benefitted from having two primary advisors that are each able to nurture my intellectual growth while offering constructive criticism and praise. Within the Infant Lab I would like to acknowledge the support of my fellow graduate students, Jo Vogelli and Shiva Fekri. I appreciate all the time we have spent together pursuing shared interests and ambitions. I especially appreciate the support they offered in the conceptual development of my thesis work. I am looking forward to several more productive years of us all working together in the Infant Lab.
I would also like to thank Dr. Krista Ranby. I am grateful for her contributions to our graduate training program, and especially for the guidance she provided in the conceptual and statistical development of my thesis.
Last, but not least, I would like to thank all the original members of my PhD cohort, Kellie Martins, Megan Grigsby, Stephanie Hooker, Carissa Kinman, and Jenn Altman. Each has been an inspiration to me over the years, and I am truly thankful to be able to call them my friends. In particular, I would like to thank Kellie, Megan, and Stephanie for making my time within the program one that has been filled with as much laughter as hard work.
v


TABLE OF CONTENTS
CHAPTER
I BACKGROUND.................................................................1
Characterizing postpartum mood disorders................................1
Significance of depression in the postpartum period.....................3
Biopsychosocial pathways for developing depression in the postpartum period....4
Theoretical models for developing depression in the postpartum period...6
Theoretical approach for developing depression in the postpartum period.6
Hypothesis and specific aims............................................9
II METHODS...................................................................10
Sample.................................................................10
Participants...........................................................10
Procedure..............................................................11
Measures...............................................................12
Statistical analyses...................................................13
Power..................................................................14
Data-cleaning & descriptive statistics.................................14
Aim 1: Confirmatory factor analysis....................................15
Aim 2: Path analysis...................................................16
III RESULTS...................................................................18
Aim 1: Confirmatory factor analysis....................................18
Aim 2: Path analysis...................................................22
IV DISCUSSION................................................................24
REFERENCES....................................................................29
vi


LIST OF TABLES
Table
1. Maternal demographic and diagnostic data...................................11
2. Correlations for CFA.......................................................20
3. Standardized and Unstandardized Coefficients for CFA.......................21
vii


LIST OF FIGURES
Figure
1. Hypothesized Confirmatory Factor Analysis: BDI-II 3 Factor Model...........16
2. Hypothesized Path Analysis: BDI-II & SCID..................................17
3. Confirmatory Factor Analysis: BDI-II 3 Factor Model........................22
4. Path Analysis: BDI-II & SCID...............................................23
viii


LIST OF ABBREVIATIONS
BDI-II DPP PPD SC ID
SEM
Beck Depression Inventory, 2nd Edition Depression in the postpartum period Postpartum Depression Structured Clinical Interview for Diagnosis Structural Equation Modeling


CHAPTER I
BACKGROUND
The postpartum period is often characterized as a joyful time in the lives of mothers and their families. However, incidence and prevalence rates identify this period as a major risk factor for the development of a mood disorder. Postpartum depression is commonly used as a catchall phrase to describe any sort of depressive symptom associated with childbirth (Jones & Venis, 2001). However, this definition is an overgeneralization of a variety of symptoms that can vary in severity, duration, and onset. Categorically, mood disorders occurring in the postpartum period can be classified as baby blues, postpartum depression, and postpartum psychosis (C. T. Beck, 2006).
Characterizing postpartum mood disorders
The term baby blues describes a transient increase in depressive symptoms
shortly, usually within 10 days, after childbirth. Baby blues are often considered to be normal as they affect up to 85% of new mothers (Berggren-Clive, 1998; Dennis & Dowswell, 2013; OHara, Zekoski, Philipps, & Wright, 1990; M. Steiner, 1998) and generally resolve without intervention within 2 weeks of onset (OHara, Schlechte,
Lewis, & Wright, 1991a). Postpartum psychosis, on the other hand, is a relatively rare condition affecting just 1 in every 1,000 mothers after having a live birth, and is characterized by the presence of hallucinations or delusions (Spinelli, 2004). Postpartum psychosis typically manifests within 2 weeks of childbirth and should be considered a psychiatric emergency that requires immediate intervention due to the association between postpartum psychosis and suicide or neonaticide/infanticide (Spinelli, 2004). Statistically, postpartum depression (PPD) tends to fall somewhere between blues and
1


psychosis, in that PPD affects up to 15% of new mothers with depressive symptoms ranging from mild to severe (Gavin & Gaynes, 2005; OHara, Neunaber, & Zekoski, 1984). PPD is considered a non-psychotic depressive episode that either begins in or extends into the postpartum period. In Colorado, depression affects approximately 14% of mothers within the first year after giving birth (Centers for Disease Control and Prevention, 2008).
In research and healthcare settings PPD is often assessed using a self-report measure of depression. Among the most common self-report screening assessments used today are the Edinburgh Postnatal Depression Scale (EPDS), the Beck Depression Inventory, 2nd Edition (BDI-II), and the Postpartum Depression Screening Scale (PDSS). A recent study found that all three self-report measures performed equally well at detecting PPD, however traditional cutoff scores used with each measure could be adjusted to increase sensitivity and specificity of clinical diagnosis (Chaudron et al., 2010). Clinically, PPD is diagnosed using a postpartum onset specifier added to a current major depressive episode of major depressive disorder, biopolar I disorder, or biopolar II disorder (American Psychiatric Association, 2000, 2013). The Diagnostic Manual of Mental Disorders, Fifth Edition (DSM-V) offers the postpartum onset specifier for episodes of a mood disorder that occur within the first four weeks after childbirth (American Psychiatric Association, 2013). However empirical research suggests that PPD can manifest throughout the first year postpartum with peak incidence of PPD occurring within the first four months of childbirth (Gavin & Gaynes, 2005; OHara et al., 1990). Due to the inconsistency in the use of PPD terminology, the term depression in the postpartum period (DPP) will be used throughout the remainder of this paper.
2


Significance of depression in the postpartum period
The effects of DPP are well documented, as it is the most common psychiatric
comorbidity associated with childbirth and has been found to have negative effects for both members of the mother-infant dyad (Sharma & Sharma, 2012; Weissman et al., 2006). Symptoms of DPP tend to persist for quite some time, averging 5 months in the general population (Chaudron, 2003), and mothers are at increased risk of developing subsequent major depressive episodes throughout their lifetime (Josefsson & Sydsjo, 2007; Lynne Murray, Cooper, Wilson, & Romaniuk, 2003; Nylen et al., 2010). Additionally, depressed mothers tend to experience increased negative emotionality (Dietz, Jennings, Kelley, & Marshal, 2009; Lundy et al., 1996), which may be partly due to cognitive appraisal deficits that heighten self-focus (Salmela-Aro, Nurmi, Saisto, & Halmesmaki, 2001; Stein et al., 2010; Teti & Gelfand, 1991). Potentially even more concerning is the effects DPP can have on parenting behaviors and, thus, child development. For the past twenty years theorists have attributed deficits in infant development and cognition to disruptions in the depressed caregivers ability to support their infants state and behavior, which in turn contribute to deficits in infants abilities to attend to and extract information about environmental contingencies (Hay, 1997). This framework can be understood as a dynamic relationship between parenting behaviors that either reinforce or undermine an infants social learning abilities. Many studies have demonstrated that DPP is often accompanied by an impoverishment in the quantity and quality of pro-social parenting behaviors. Absence of pro-social parenting behaviors are linked with poorer cognitive development in infants, as well as delayed school readiness, and decreased IQ scores later in childhood (Cicchetti, Rogosch, & Toth, 2000; Hay,
1997; Milgrom, Westley, & Gemmill, 2004; Murray, 1992; NICHD, 1999; Stanley,
3


Murray, & Stein, 2004). Additional studies have demonstrated that the frequency of mother-child interactions are not only reduced if the mother suffers from DPP, but in some cases their reactions to their child can be characterized as hostile or unresponsive (Beebe et al., 2007; Dietz et al., 2009; Field, 2010; Flykt, Kanninen, Sinkkonen, & Punamaki, 2010). Taken together, we begin to see the deleterious effects of DPP for both the mother and child, as research has identified increased negative emotionality in caregivers and decreased attention to infants as predictive factors for cognitive deficits and behavior problems during childhood development (Lawson & Ruff, 2004).
Biopsychosocial pathways for developing depression in the postpartum period
Conceptually, the manifestation of DPP symptoms are attributed to a combination
of biopsychosocial stressors associated with childbirth and caregiving (Berggren-Clive,
1998). Independently, neither physical nor psychosocial stress has been shown to cause
DPP. To date no evidence has found a direct hormonal imbalance as the determinant for
DPP (OHara & McCabe, 2013), however several studies have implicated physiological
changes in estradiol and progesterone levels as contributing factors in the development of
mood dysregulation in the postpartum period (Bloch et al., 2000, 2005; Bloch, Daly, &
Rubinow, 2003). Even if estradiol and progesterone levels can be correlated with DPP,
this explanation may be overly simplistic as estradiol and progesterone interact with
neurotransmitter systems that directly affect serotonin and dopamine uptake. Serotonin
has been found to regulate fundamental physiologic responses such as sleep and appetite,
and a global decrease in serontinergic function has been associated with major depression
(Gu et al., 2003; Mann, 1999). Additionally, animal studies have demonstrated that a
sharp reduction in the functioning of dopaminergic pathways precede depressive
symptoms (Byrnes, Byrnes, & Bridges, 2001). However, despite these findings, there has
4


been little evidence to suggest that levels of estrogen, progesterone, or cortisol vastly differentiate between depressed and nondepressed mothers in the postpartum period (Bloch et al., 2003; Workman, Barha, & Galea, 2012).
From a psychosocial perspective, disadvantaged social conditions as well as stress accompanying the birth of a child have been associated with the development of DPP. Demographic markers (i.e. race/ethnicity, education status, and younger age) related to poverty levels have consistently been found to be risk factors for developing DPP (Howell, Mora, Horowitz, & Leventhal, 2005; Rich-Edwards et al., 2006; Segre, OHara, Arndt, & Stuart, 2007). These findings also hold true for mothers in Colorado, as maternal age (< 20), race/ethnicity (non-white), marital status (unmarried), education (<12 years), and insurance status (not insured) have all been found to be statistically significant predictors of self-reported symptoms of DPP (Centers for Disease Control and Prevention, 2008). Additional studies have looked at health and social risk factors occurring either perinatally or postpartum. Perinatal health complications, including preeclampsia, perinatal hospitalization, and emergency caesarean delivery, have been identified as risk factors for developing DPP, even while controlling for social economic status and pre-existing conditions (Blom et al., 2010). Relevant factors occurring during the postpartum period, such as infant health and temperament have also been identified as risk factors for developing DPP (Blom et al., 2010; Vik et al., 2009). Similarly, data from Colorado has suggested that pregnancy related risk factors, such as tobacco use, physical abuse, and financial, partner-related, or traumatic stressors, affect the development of postpartum depressive symptoms (Centers for Disease Control and Prevention, 2008).
5


Theoretical models for developing depression in the postpartum period
Theoretically, several models have been explored in prior research attempting to
explain the onset of DPP. An early cognitive behavioral model associated negative attributions about ones self and other negative psychological vulnerabilities to the development of DPP (OHara, Rehm, & Campbell, 1982). Expanding the cognitive behavioral model to include psychosocial stress has been found to better predict the development of DPP (OHara, Schlechte, Lewis, & Wright, 1991b), infant temperament, and interpersonal support (Cutrona & Troutman, 1986). One of the more recent models predicted DPP by analyzing cognitive dissonance between expected and actual social support received by mothers postpartum (C. T. Beck, 2002). Although several models have demonstrated conceptual and theoretical support for various biopsychosocial aspects to contribute to the development of DPP, recent literature has called for additional exploration into new model development and empirical testing with well defined populations to further the current understanding of why DPP manifests in -15% of new mothers (OHara & McCabe, 2013).
Theoretical approach for developing depression in the postpartum period
The current proposal aims to test specific factors that may influence the
development of DPP in a sample of mothers at four months postpartum. Collecting data on mothers who are all four months postpartum provides insight into unique challenges faced during this timepoint. Furthermore, conducting analyses on the depressive symptoms of mothers at this timepoint may be particularly useful as it falls within the recommended time frame for maternal depression screening put forth by the American Academy of Pediatrics (Earls, 2010), as well as near the peak incidence of new onset cases of depression occurring in the postpartum period (Chaudron, 2003). The conceptual
6


framework and theoretical model influencing this study was largely shaped by the biopsychosocial model, which emphasizes the dynamic interaction between biological, psychological, and social variables that may manifest as categorically defined psychological conditions (Engel, 1977; Fava & Sonino, 2008). As discussed above, recent literature has demonstrated that numerous biological and psychosocial factors are predictive or associated with DPP, yet brief self-report questionnaires are commonly used to assess PPD.
To better understand the underlying factors associated with self-reported DPP symptoms a recent study conducted exploratory and confirmatory factor analyses to identify the latent factor structure of the BDI-II. Findings from a sample of mothers three-months postpartum found a three-factor structure (i.e. Cognitive-Affective,
Somatic-Anxiety, and Guilt) to be the most parsimonious (Carvalho Bos et al., 2009). A replication study conducted on data obtained by the Infant Lab at the University of Colorado Denver also supported a three-factor structure for mothers four months postpartum. An exploratory factor analysis demonstrated that the most parsimonious fit indicated that Cognitive-Appraisal, Somatic-Exhaustion, and Impairment factors explained the majority of total variance of the item responses obtained on the BDI-II (Cudmore, 2011). The most recent study looking at the factor structure of the BDI-II included a large sample of postpartum women and also found a three factor solution (i.e. Cognitive, Appraisal, and Somatic) to provide the best fit for the data (Manian, Schmidt, Bomstein, & Martinez, 2013). These three studies identified factor structure constructs that are, at least nominally, related to biological (Somatic-Anxiety; Somatic-Exhaustion; & Somatic, respectively) and psychosocial (Cognitive-Affective; Cognitive-Appraisal; Cognitive, respectively) symptoms of depression categorized in the DSM (American
7


Psychiatric Association, 2000, 2013). Additionally, cognitive, somatic, and affective components are consistently reported for the BDI-II when tested with a variety of populations (Grothe et al., 2005; Kneipp, Kairalla, Stacciarini, & Pereira, 2009; Siegert, Walkey, & Tumer-Stokes, 2009; Storch, Roberti, & Roth, 2004; Viljoen, Iverson, Griffiths, & Woodward, 2003; Ward, 2006).
Based on the literature conducting factor analysis on the BDI-II on samples of mothers in the postpartum, and on the literature offering support for both biological and psychosocial factors that contribute to the development of DPP symptoms, I hypothesize that factors related to cognitive-appraisal drive elevated scores on a measurement of DPP symptoms. Theoretically, I posit that cognitive-appraisal related items on the BDI-II are endorsed by individuals that have developed psychological vulnerabilities (e.g. negative attributions about oneself, cognitive distortions, or increased self-focus), which lead to elevated responses on somatic-exhaustion and impairment related items. Although difficult to empirically test, assumptions are made that all women in the postpartum period undergo a rapid hormonal change that represents a biological stressor, and that these changes are partially captured by responses to somatic-related questions on the BDI-II. Furthermore, although all new mothers experience somatic changes during the postpartum period, it is the cognitive-appraisal aspect of depression that distinguishes mothers that develop DPP. Again, difficult to empirically test, but an assumption is made that elevated item responses on the BDI-II that are related to cognitive-appraisal are shaped by psychosocial factors that increase a mothers self-focus. Increased self-focus leads to elevated scores on the BDI-II as mothers are more likely to judge symptoms more severely than mothers not experiencing cognitive-appraisal deficits. Despite a lack of empirical research directly relating items on the BDI-II to specific biological or
8


psychosocial stressors, the proposed model offers a biopsychosocial pathway for understanding the development of DPP.
The present study attempts to replicate the factor structure of the BDI-II previously reported in the literature through a confirmatory factor analysis, while also designing and testing a path analysis emphasizing the role of cognitive-appraisal in the development of DPP by four months postpartum. To date, no publications have been reported in the extant literature testing a path analysis for DPP utilizing the BDI-II12.
Hypothesis and specific aims
The overarching hypothesis of this study is that the distinct factor structure of the BDI-II, as reported by a sample of mothers at four months postpartum, can offer a better understanding of the categorically related symptoms of DPP while providing support for a structural equation model (SEM) illustrating cognitive-appraisals role in driving elevated scores for DPP. This hypothesis will be explored through two distinct aims:
1. Test the hypothesized three-factor structure (i.e. cognitive-appraisal, somatic-exhaustion, and impairment) on the BDI-II using confirmatory factor analysis.
2. Test whether a conceptual model emphasizing the influence of cognitive-appraisal on the development of DPP as measured by the BDI-II and a Structured Clinical Interview for DSM-IV Axis I Disorders (SCID) is supported using path analysis.
1 A search was conducted on July 6, 2013 in Ovid using the following search criteria: [MeSH]: Statistical, models, [MeSH]: Depression, postpartum, and [Keyword]: BDI-II or [Keyword]: Beck Depression Inventory. Eight results were identified, however no results were found that specifically included a SEM.
2 A search was conducted on July 6, 2013 in CINAHL using the following search criteria: [MH]: Structural Equation Modeling, and [MH]: Depression, Postpartum. No results were identified.
9


CHAPTER II
METHODS
Sample
Data for the present study was collected as part of a larger study within the Infant Lab at the University of Colorado Denver. The larger study aimed to assess infants attention to infant directed speech and cognitive development based on the depressive status of caregivers (Kaplan, Danko, Diaz, & Kalinka, 2011; Kaplan, Danko, Kalinka, & Cejka, 2012). The present dataset consists of a sample of one hundred and eight mothers of four-month-old mother-infant dyads. Of the one hundred and eight mothers at four-months postpartum participating in the above referenced study, data was missing from the measures of DPP for eight of the mothers and, thus, these persons were excluded from analysis. One hundred mothers with complete data on all measures of DPP were included in the final analyses.
Participants
Table 1 presents demographic and diagnostic information for the full sample of mothers and infants (N = 100). No differences were found as a function of BDI-II category or depression diagnosis on maternal age, infant age, infant gender, maternal education, family income, ethnicity, or marital status. The number of children, for mothers with elevated compared to non-elevated BDI-II scores, was higher, but the number of children per caregiver did not differentiate clinically depressed from non-depressed mothers.
10


Table 1. Maternal demographic and diagnostic data.
Variable Non- Elevated (BDI-II) Elevated (BDI-II) NDEP (DSM-IV) DEP (DSM-IV)
N = 100 56 44 88 12
Infant Gender (girls/boys) 28/28 20/24 43/45 5/7
Age of mother (years) 30.1 (4.6) 28.9 (5.5) 30.0 (5.2) 29.4 (4.2)
Age of Infant (days) 131 (21.5) 143 (23.8) 135 (23.3) 143 (21.9)
Ethnicity White 33 (58.9%) 21(47.7%) 48 (54.5%) 6 (50.0%)
Latina 14 (25.0%) 13 (29.5%) 24 (27.3%) 3 (25.0%)
African-American 5 (8.9%) 6(13.6%) 8 (9.1%) 3 (25.0%)
Asian 3 (5.4%) 2 (4.6%) 5 (5.7%) 0 (0.0%)
Native American 1 (1.8%) 2 (4.6%) 3 (3.4%) 0 (0.0%)
Marital Status Married 45 (80.4%) 30 (68.2%) 66 (75.0%) 9 (75.0%)
Unmarried 11 (19.6%) 14 (31.8%) 22 (25.0%) 3 (25.0%)
Mothers Education 5.7 (1.5) 5.1 (1.1) 5.5 (1.4) 5.3 (1.2)
Family Income 6.1 (2.4) 5.7 (2.4) 5.8 (2.5) 6.6 (1.6)
Number of children 1.8 (l.l)a 2.3 (1.2)b 2.0 (1.2) 2.1 (1.0)
BDI-II score 7.8 (3.2)a 22.3 (6.8)b 12.5 (7.6)a 26.3 (8.0)b
GAF rating 79.4 (6.9)a 69.3 (9.6)b 77.3 (7.7)a 62.1 (9.4)b
Columns are defined based on BDI-II scores and SCID interviews at 4 months.
Numbers in parentheses are standard deviations. GAF = global assessment of functioning, as estimated from SCID interviews. Means with a superscripts significantly differed from means with b superscripts (p = .05).
Procedure
Mothers responding to an advertisement in a local parenting magazine were invited to participate in a study regarding the effects of depression on infant attention and learning. Mother-infant dyads were greeted by research assistants and were asked to complete a study consent form approved by our institutional review board. The mothers then completed a demographic questionnaire and a self-report measure of depressive symptoms. With assistance from research personnel, infants completed an auditory attention paradigm before taking part in an assessment of cognitive development. Mothers were then administered a structured clinical interview by a Ph.D, M.A, or graduate level student in clinical psychology to assess for a clinical diagnosis of
11


depression. The structured clinical interview took approximately one hour to complete, and total participation time lasted about two and a half hours. Participants were compensated fifty dollars for their time.
Measures
Participants were administered a Structured Clinical Interview for DSM-IV Axis I Disorders (SCID), Patient Edition (SCID-I/P) and were asked to complete a BDI-II.
The SCID is a diagnostic interview intended to specify patients that meet criteria for major mental disorders according to the DSM-IV-TR (First, Spitzer, Gibbon, & Williams, 2002). Including all versions of the SCID (i.e. nonpatient, patient, patient with psychotic screen, research, clinical trials, axis II disorders) well over seven hundred publications can be found in peer-reviewed journals where the SCID was utilized for diagnostic evaluation. Several studies have demonstrated that the SCID has good reliability with Kappa values ranging from 0.66 0.93 for diagnosis of major depressive disorder (Lobbestael, Leurgans, & Arntz, 2010; Skre, Onstad, Torgersen, & Kringlen, 1991; Williams et al., 1992; Zanarini & Frankenburg, 2001; Zanarini et al., 2000). The validity of the SCID is difficult to ascertain, as the diagnosis of mental illness is largely subjective in nature, depending on the symptoms endorsed by the patient and potentially confirmatory observations made by the clinician or witness. Additionally, the SCID was developed in an attempt to standardize the process of clinical diagnosis made ideographically by trained professionals, and, thus, the SCID is often referenced as the gold standard as to which diagnostic procedures are compared (Shear et al., 2000; J. L. Steiner, Tebes, Sledge, & Walker, 1995).
The Beck Depression Inventory, 2nd Edition (BDI-II) is a proprietary self-report assessment developed as a revision to the original Beck Depression Inventory (BDI)
12


published by Psychological Corporation (A. T. Beck, Steer, & Brown, 1996). The BDI-II, like its predecessor the BDI, is a twenty-one item self-report measure developed to detect symptoms of depression. The BDI-II revision made changes reflecting revised diagnostic criteria included in the DSM-IV for major depressive disorder, i.e. changes made to item location, wording, and time frame (Carvalho Bos et al., 2009; Dozois, Dobson, & Ahnberg, 1998; Ward, 2006). Empirical literature consistently demonstrates that the BDI-II is an accurate measure of depression, occurring in the postpartum period or otherwise (Boyd, Le, & Somberg, 2005; Chaudron et al., 2010; Dozois et al., 1998; Smarr &
Keefer, 2011). A review of self-report instruments used for PPD found the BDI-II to have excellent internal consistency, excellent specificity, acceptable sensitivity, excellent positive predictive value, and good concurrent validity (Boyd et al., 2005).
Statistical analyses
Considering the factor structures that have been previously published using the BDI-II as a measure of DPP, SEM will be employed to test current research questions. The first aim was tested through a confirmatory factor analysis on the sample consisting of one hundred mothers at four months postpartum. The second aim was tested through a proposed path analysis that identified whether latent factors, as measured by the twenty-one self-report questions of the BDI-II, could support a conceptual framework that symptoms related to cognitive-appraisal influence the development of DPP; as measured by the BDI-II and SCID. Both confirmatory factor analysis and path analysis are statistical tests that operate under the SEM umbrella term used to describe procedures that combine components of factor analysis and multiple regression (Tabachnick &
Fidell, 2012). The strength of SEM is seen as its unique ability to test proposed theoretical models simultaneously based on observed or latent factors. Although the
13


proposed model could be tested in one SEM, a distinct first step is being taken to test a three-factor solution to item responses on the BDI-II through a confirmatory factor analysis, before a path analysis of the latent factors affecting one another, and their correlation with a clinical diagnosis of depression, is applied.
Power
Statistical power refers to the probability of detecting an effect if a real effect exists. Given the limited sample size available in the retrospective dataset, concern was warranted. However, the planned statistical procedures used for data analysis are considered fairly robust and SEM is preferable for testing smaller sample sizes due to the tests ability to test multiple models simultaneously, as opposed to running multiple individual regression analyses and increasing the probability of a committing a Type I error. Despite the relatively small sample size, the proposed analysis was consistent with current literature establishing a precedent for factor analysis to be used with samples ranging from one hundred to two hundred if the model being tested consists of well determined factors (MacCallum, Widaman, Zhang, & Hong, 1999).
Data-cleaning & descriptive statistics
The full dataset of one hundred eight mothers at four months postpartum with any
information collected on the BDI-II or SCID was entered into IBM SPSS statistical
software (IBM Corp., 2012). Assumptions for CFA and path analysis were assessed. Data
with missing values were identified and a value of 999 was inserted to clearly mark
these cases. Descriptive statistics (i.e. variable specific mean, median, mode, standard
deviation, and frequency distribution charts) were then performed in SPSS to make sure
that there were not any apparent data entry errors in the dataset before continuing with
the analyses. Normality and linearity were assessed through the examination of frequency
14


histograms computed for each observed variable. Particular attention was paid to indices of skew and kurtosis, and an examination of the covariance matrix for the observed variables was performed to assess for multicolinarity and singularity.
Aim 1: Confirmatory factor analysis
CFA was first introduced in 1969 (Joreskog, 1969) and is widely used in the
social sciences as an acceptable statistical method for researchers to test applied research questions (Brown, 2006). To test the first aim of the proposed study, a CFA was conducted using MPlus statistical software (Muthen & Muthen, 2011) to test the hypothesis that a three-factor solution is acceptable for a sample of one hundred mothers at four months postpartum. The hypothesized three-factor solution tested on the BDI-II is shown in Figure 1. Acceptable model fit was determined using several goodness-of-fit indices, including the chi-square statistic (x2), comparative fit index (CFI), and root mean square error of approximation (RMSEA). The x2 tests the hypotheses that the variables are unrelated; within the context of a path analysis a non-significant value is desired (Tabachnick & Fidell, 2012). The CFI provides an estimate of how well the data fit the model, relative to other models. The CFI is well equipped to handle relatively small sample sizes and it is standardized on a scale of 0 to 1, with values greater than 0.95 indicative of a good model fit (Bentler, 1988; Hu & Bender, 1999). The RMSE compares the hypothesized model to a saturated, or perfect, model for the available data. This estimation is made to analyze the lack of appropriate model fit. When interpreting the RMSE, values less than 0.06 are considered indicative of a good model fit (Hu & Bentler, 1999), whereas values greater than 0.1 indicate poor fitting models (Browne & Cudeck, 1992). However, interpretation of the RMSEA value may be challenging with small samples given this tests tendency to over-reject the true model (Hu & Bentler, 1999).
15


Path Analysis: BDI-II (3 Factor Model at 4 months Postpartum) & SCID
Figure 1. Hypothesized Confirmatory Factor Analysis: BDI-II 3 Factor Model Aim 2: Path analysis
Path analysis is a statistical procedure that has been in use since its development in the 1920s (Wright, 1923, 1934), and has been widely applied to the social sciences (Stage, Carter, & Nora, 2004). To test the second aim, a path analysis was conducted with Mplus statistical software (Muthen & Muthen, 2011) to test the proposed theoretical model that emphasizes cognitive-appraisals influence on responses to the BDI-II and SCID in relation to factors of somatic-exhaustion and impairment. The hypothesized three-factor solution tested on the BDI-II and correlated with SCID diagnosis of DPP is shown in Figure 2. Acceptable model fit was determined using a chi-square statistic (x2), comparative fit index (CFI), and root mean square error of approximation (RMSEA).
Additionally, the correlation between each latent factor and SCID diagnosis was calculated using a Pearson Product-Moment Correlation Coefficient (r). The correlation
16


coefficient measures the degree to which a linear relationship exists between two variables, and can range in value from -1 to +1 (Tabachnick & Fidell, 2012). Values closer to +1 indicate a positive correlation, whereas values approaching 0 indicate no correlation, and values closer to -1 indicate a negative correlation.
Figure 2. Hypothesized Path Analysis: BDI-II & SCID
17


CHAPTER III
RESULTS
Aim 1: Confirmatory factor analysis
To test the first aim, a confirmatory factor analysis (CFA) was conducted on the twenty-one BDI-II item responses for 100 mothers at four months postpartum. Item responses were normally distributed and a CFA was conducted in MPlus 7.0. A correlation table with means and standard deviations is shown in Table 2; the theoretical model is presented in Figure 1. The hypothesized three-factor model (i.e. cognitive-appraisal, somatic-exhaustion, and impairment) is confirmed in the measurement portion of the model. Assumptions of multivariate normality and linerarity were evaluated through SPSS 21.0 using box plots and Mahalnobis distance. No univariate or multivariate outliers were identified; however item 9 on the BDI-II (i.e. Suicidal Thoughts or Wishes) only received responses of a zero (i.e. I dont have any thoughts of killing myself) or a one (i.e. I have thoughts of killing myself, but I would not carry them out) out of the four possible response options, and thus was declared as a categorical variable during Mplus analysis. The nonsignificance of the chi-square (x2(i86)=211.386; ns) shows that the covariance matrix and mean vector in the population are equal to the model-implied covariance matrix and mean vector. The CFI value (CFI=0.947) indicates that the proposed model fits better than an independent model in which the variables are not related. The low RMSE value (RMSE=0.037) also indicates that the model fit the data well. Additionally, all factor loadings were significant (p < .000). Those values indicate a good fit between the model and observed data. Standardized paramerter estimates are provided in Figure 3; unstandardized estimates are shown in Table 3. No
18


post-hoc modifications were indicated from the analysis because of the good-fit indexes, and the residual analysis did not indicate any problems.
19


Table 2. Correlations for CFA
Observed Variable i 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
1. Sadness .254 - - - - - - - - - - - - - - - - - - - -
2. Pessimism .305 .347 - - - - - - - - - - - - - - - - - - -
3. Past Failure .401 .438 .670 - - - - - - - - - - - - - - - - - -
4. Loss of Pleasure .528 .306 .483 .422 - - - - - - - - - - - - - - - - -
5. Guilty Feelings .606 .333 .479 .507 .592 - - - - - - - - - - - - - - - -
6. Punishment Feelings .292 .247 .496 .395 .421 .372 - - - - - - - - - - - - - - -
7. Self-Dislike .535 .510 .552 .519 .557 .562 .646 - - - - - - - - - - - - - -
8. Self-Criticalness .476 .387 .295 .404 .550 .230 .548 .666 - - - - - - - - - - - - -
9. Suicidal Thoughts .603 .177 .561 .520 .551 .485 .542 .468 - - - - - - - - - - - - -
10. Crying .404 .374 .378 .313 .274 .224 .438 .294 .283 .793 - - - - - - - - - - -
11. Agitation .200 .321 .225 .461 .391 .303 .359 .456 .237 .284 .644 - - - - - - - - - -
12. Loss of Interest .485 .272 .347 .614 .445 .365 .421 .365 .607 .352 .301 .405 - - - - - - - - -
13. Indeciseveness .288 .173 .167 .243 .373 .131 .283 .405 .357 .104 .247 .385 .502 - - - - - - - -
14. Worthlessness .363 .325 .515 .435 .457 .392 .517 .415 .430 .308 .321 .389 .333 .413 - - - - - - -
15. Loss of Energy .339 .328 .341 .450 .307 .215 .433 .362 .323 .273 .318 .401 .224 .184 .358 - - - - - -
16. Changes in Sleep .364 .249 .200 .486 .289 .179 .229 .239 .198 .315 .367 .376 .279 .136 .334 .712 - - - - -
17. Irritability .378 .435 .267 .491 .349 .331 .437 .394 .309 .255 .519 .410 .265 .361 .408 .394 .586 - - - -
18. Chages in Appetite .231 .089 .215 .245 .243 .010 .270 .277 .342 .186 .286 .343 .315 .155 .380 .190 .245 .606 - - -
19. Concentration Difficulty .257 .248 .245 .289 .301 .193 .368 .346 .438 .227 .420 .282 .351 .299 .309 .320 .348 .343 .432 - -
20. Tiredness or Fatigue .279 .251 .213 .304 .230 .219 .328 .202 .362 .211 .361 .308 .195 .172 .508 .386 .293 .345 .381 .458 -
21. Loss of Interest in Sex .262 .229 .044 .244 .306 .188 .181 .238 .221 -.010 .233 .265 .576 .269 .154 .361 .149 .121 .206 .252 1.0
N=100; M=0; SD=1
20


Table 3. Standardized and Unstandardized Coefficients for CFA
Observed variable______________B______(3_____SE p valule
Cognitive-Appraisal
1. Sadness
2. Pessimism
3. Past Failure
4. Loss of Pleasure
5. Guilty Feelings
6. Punishment Feelings
7. Self-Dislike
8. Self-Criticalness
9. Suicidal Thoughts
10. Crying
12. Loss of Interest
14. Worthlessness Somatic-Exhaustion
11. Agitation
15. Loss of Energy
16. Changes in Sleep
17. Irritiability
18. Changes in Appetite
19. Concentration Difficulties
20. Tiredness or Fatigue Impairment
13. Indecisiveness
21. Loss of Interest in Sex
.541 .675 .082 .000
.496 .529 .077 .000
.802 .617 .078 .000
.757 .733 .057 .000
.883 .722 .046 .000
.529 .546 .074 .000
1.00 .783 .055 .000
.845 .652 .062 .000
1.19 .750 .090 .000
.697 .493 .078 .000
.695 .687 .067 .000
.610 .598 .084 .000
1.37 .642 .066 .000
1.00 .633 .057 .000
1.24 .557 .082 .000
1.39 .684 .065 .000
.967 .467 .077 .000
1.03 .595 .080 .000
1.00 .556 .085 .000
1.00
1.07
.869
.663
.097
.096
.000
.000
B = unstandardized coefficient; (3 = standardized coefficient; SE = standardized standard error; p value = standardized p value
21


Path Analysis: BDI-II (3 Factor Model at 4 months Postpartum) & SCID
Figure 3. Confirmatory Factor Analysis: BDI-II 3 Factor Model. x2(i86)=21 1.386, ns; CFI=0.947; RMSE=0.037.
Aim 2: Path analysis
The second aim used a path analysis designed to demonstrate the influence that a cognitive-appraisal latent factor has on elevated scores on the BDI-II, as well as test the correlation between the individual latent factors in the proposed 3-factor model (figure 2) and a clinical diagnosis, as indicated by the SCID, of a major depressive episode in the the postpartum period. For a sample of mothers at four months postpartum (N=100), the proposed model (figure 2) demonstrated a good fit, as seen in figure 4. The nonsignificance of the chi-square (x2(204)=231.9 1 7; ns) showed that the covariance matrix and mean vector in the population were equal to the model-implied covariance matrix and mean vector. The CFI value (CFI=0.945) indicated that the proposed model fit better
22


than an independent model in which the variables are not related. The low RMSE value (RMSE=0.037) also indicated that the model fit the data well. Additionally, all factor loading were significant (p < .000). No post-hoc modifications were indicated from the analysis due to the good-fit indexes, and residual analysis did not indicate any problems.
The path analysis supported the proposed influence of cognitive-appraisal on somatic-exhaustion (P = .754) and impairment (P = .481). Moreover, the proposed path analysis tested the correlation between each latent factor and the presence of a SCID diagnosis of major depressive episode in the postpartum period. Cognitive-appraisal demonstrated the highest correlation (r = .663) with SCID diagnosis, compared to either somatic-exhaustion (r = .458) or impairment (r = .012).
Path Analysis: BDI-II (3 Factor Model at 4 months Postpartum) & SCID
Figure 4. Path Analysis: BDI-II & SCID. x2(2o4)=231.917, ns; CFI=0.945; RMSE=0.037.
23


CHAPTER IV
DISCUSSION
The current study proposed and tested a confirmatory factor analysis and a path analysis model based on item responses on the BDI-II and clinical diagnosis of major depressive episode in the postpartum period on the SCID. Both proposed models were formulated based on literature emphasizing the effects of biopsychosocial stressors on the development of DPP. Although the influences of both biological and psychosocial stressors have independently been associated with DPP, neither have been determined to cause DPP. Similarly, the current study is not able to offer any causal associations between biological or psychosocial factors on DPP. However, the present study does offer support for the influence of a cognitive-appraisal factor on elevation of scores on the BDI-II and subsequent diagnosis of a major depressive episode during the postpartum period.
The proposed models were specifically tested in a sample of mothers 4 months postpartum for several important reasons. Realizing that there has been a disconnect between onset timing used in the DSM criteria for PPD and empirical studies of PPD, it was determined that a study focused on latent factors associated with DPP would be best served by testing a sample at a specific timepoint. Although it is believed that hormonal levels are quite individual, a specific timepoint allows for a standardized amount of time to pass for the entire sample. Furthermore, as the focus of the present study was on the latent factor structure of the BDI-II, it was determined that results of the current study may have the most external validity if the sample was within the recommended
24


guidelines for screening for PPD put forth by the American Accademy of Pediatrics (Earls, 2010).
Aim 1 of the current study proposed and tested a CFA of a 3-factor model for the BDI-II completed by mothers at 4 months postpartum. Findings from this studys CFA support the identification of cognitive/appraisal, somatic, and impairment related latent factors consistently reported in prior literature assessing the BDI-II factor structure during the postpartum period (Carvalho Bos et al., 2009; Cudmore, 2011; Manian et al., 2013). Testing the proposed CFA, prior to testing the proposed path analysis, was important to determine whether the latent factor structure of the BDI-II, as measured by the current studys sample, was representative and consistent with previously published literature on the topic, before testing the proposed theoretical framework emphasizing an influence by the cognitive-appraisal latent factor on elevated BDI-II scores and SCID diagnosis via a path analysis. Results of the current studys CFA demonstrated good model fit and offers additional support for a 3-factor model representing cognitive-appraisal, somatic-exhaustion, and impairment latent factors of the BDI-II in a postpartum population.
Aim 2 of the current study tested a proposed path analysis suggesting that the cognitive-appraisal latent factor influences elevated scores on the BDI-II at 4 months postpartum. Results of the path analysis offered support to a theoretical model emphasizing the influence of a cognitive-appraisal factor on DPP. As seen in figure 4, significant path estimates of cognitive appraisal on somatic-exhaustion and impairment provides evidence that cognitive-appraisal influences item responses on the BDI-II, and thus elevated scores, at 4 months postpartum for the studys sample. Within SEM, the
25


path estimates between cognitive-appraisal and somatic-exhaustion and impairment latent factors essentially test a hypothesized causal relationship. Even though results of the path analysis supported hypothesized casual relationships between cognitive-appraisal and somatic-exhaustion and impairment factors, a causal relationship could not be truly measured given the study design and the fact that a single self-report measure was used to assess for DPP. However, additional support was generated for the proposed theoretical model by identifying the correlations between each of the latent factors and clinical diagnosis of a major depressive episode for the included sample. As shown in Figure 4, cognitive-appraisal provided the highest correlation of the 3-factor model of the BDI-II with SCID diagnosis, suggesting that elevated scores on the BDI-II items that are most highly correlated with the cognitive-appraisal latent factor are also most highly correlated with clinical diagnosis of major depressieve episode in the postpartum period.
Taken together, aims 1 and 2 of the current study offer important contributions to the understanding of DPP manifestation at 4 months postpartum as measured by the BDI-II. In addition to supporting previously published literature on the factor structure of the BDI-II in postpartum populations, the path analysis offers support for a theoretical model that suggests cognitive-appraisal influences symptoms and potentially the diagnosis of DPP. This finding could provide insight into future research studies designed to identify causal pathways for the development of PPD. Additionally, findings from the current study may provide clinicians with increased knowledge about the importance of cognitive-appraisal related variables measured through self-report measures. Although somatic-exhaustion related symptoms were endorsed more frequently on the BDI-II by mothers at 4 months postpartum in the current sample than cognitive-appraisal related
26


symptoms, the significant path estimates demonstrate the influence that cognitive-appraisal may exert on BDI-II item responses. Given the pediatric guidelines for assessing symptoms of DPP at infant well child checks throughout the first four months postpartum, clinicians may want to pay particular attention to endorsed symptoms that are related to cognitive-appraisal, as they may be more indicative of the manifestation of DPP.
Despite the promising findings reported in this study that offer support for the influence of a cognitive-appraisal factor on DPP, limitations of the current study and statistical methods must be considered. As previously discussed, the path analysis tests a hypothesized causal model, but does not prove causality between the endorsement of cognitive-appraisal related symptoms and DPP. Furthermore, within the proposed model the directionality of the paths between cognitive-appraisal, somatic-exhaustion, and impairment latent factors could be reversed without changing the path estimates. Due to this fact, correlations between the latent factors and a SCID diagnosis were included to offer additional evidence of cognitive-appraisals influence on DPP. Additionally, the current studys sample size was relatively small and the number of mothers with a clinical diagnosis of major depressive episode were even fewer, however the percentage of mothers receiving a clinical diagnosis of major depressive episode during the postpartum period are equivalent to larger studies identifying the prevelance of PPD. Although SEM offers advantages for testing hypothesized causal models and may even be considered preferential for analyzing small samples, caution is advised when considering the generalizability of the findings.
27


Future research studies should look to include larger samples, additional timepoints, and additional measures of depression in an attempt to determine if symptoms related to cognitive-appraisal are a primary cause of DPP.
28


REFERENCES
American Psychiatric Association. (2000). Diagnostic and statistical manual of mental disorders (Revised 4th ed.). Washington DC author. American Psychiatric Association. Retrieved from
http://psychiatry online.com/resourceTOC.aspx?resourceID=l
American Psychiatric Association. (2013). Diagnostic and Statistical Manual of Mental Disorders (Fifth ed.) (5th Editio.). Arlington, VA: American Psychiatric Publishing, Inc.
Beck, A. T., Steer, R. A., & Brown, G. (1996). Manual for the Beck Depression Inventory-II. San Antonio, TX: Psychological Corporation.
Beck, C. T. (2002). Theoretical perspectives of postpartum depression and their treatment implications. The American Journal of Maternal Child Nursing, 27(5), 282-7. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/12209059
Beck, C. T. (2006). Postpartum depression: it isnt just the blues. The American Journal of Nursing, 106(5), 40-50; quiz 50-1. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/16639243
Beebe, B., Jaffe, J., Buck, K., Chen, H., Cohen, P., Blatt, S., ... Andrews, H. (2007). Six-week postpartum maternal self-criticism and dependency and 4 month mother-infant self- and interactive contingencies. Developmental Psychology, 43(6), 1360-76. doi:10.1037/0012-1649.43.6.1360
Bentler, P. M. (1988). Comparative fit indexes in structural models. Psychological Bulletin, (107), 238-246.
Berggren-Clive, K. (1998). Out of the darkness and into the light: womens experiences with depression after childbirth. Canadian Journal of Community Mental Health = Revue Canadienne de Sante Mentale Communautaire, 17(1), 103-20. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/10351169
Bloch, M., Daly, R. C., & Rubinow, D. R. (2003). Endocrine factors in the etiology of postpartum depression. Comprehensive Psychiatry, 44(3), 234-46. doi: 10.1016/S0010-440X(03)00034-8
Bloch, M., Rubinow, D. R., Schmidt, P. J., Lotsikas, A., Chrousos, G. P., & Cizza, G. (2005). Cortisol response to ovine corticotropin-releasing hormone in a model of pregnancy and parturition in euthymic women with and without a history of postpartum depression. The Journal of Clinical Endocrinology and Metabolism, 90(2), 695-9. doi:10.1210/jc.2004-1388
29


Bloch, M., Schmidt, P. J., Danaceau, M., Murphy, J., Nieman, L., & Rubinow, D. R. (2000). Effects of gonadal steroids in women with a history of postpartum depression. The American Journal of Psychiatry, 757(6), 924-30. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/10831472
Blom, E. A., Jansen, P. W., Verhulst, F. C., Hofman, A., Raat, H., Jaddoe, V. W. V, ... Tiemeier, H. (2010). Perinatal complications increase the risk of postpartum depression. The Generation R Study. BJOG: An International Journal of Obstetrics and Gynaecology, 777(11), 1390-8. doi:10.1111/j.l471-0528.2010.02660.x
Boyd, R. C., Le, H. N., & Somberg, R. (2005). Review of screening instruments for postpartum depression. Archives of Womens Mental Health, 5(3), 141-53. doi: 10.1007/s00737-005-0096-6
Brown, T. A. (2006). Confirmatory Factor Analysis for Applied Research (p. 475). New York: Guilford Publications. Retrieved from http://books. google. com/books?id=QTasbvQCZQcC
Browne, M. W., & Cudeck, R. (1992). Alternative ways of assessing model fit. Sociological Methods & Research, (21), 230-258.
Byrnes, E. M., Byrnes, J. J., & Bridges, R. S. (2001). Increased sensitivity of dopamine systems following reproductive experience in rats. Pharmacology, Biochemistry, and Behavior, 68(3), 481-9. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/11325402
Carvalho Bos, S., Pereira, A. T., Marques, M., Maia, B., Soares, M. J., Valente, J., ... Azevedo, M. H. (2009). The BDI-II factor structure in pregnancy and postpartum: Two or three factors? European Psychiatry: The Journal of the Association of European Psychiatrists, 24(5), 334-40. doi:10.1016/j.eurpsy.2008.10.003
Centers for Disease Control and Prevention, C. (2008). Prevalence of Self-Reported Postpartum Depressive Symptoms: 17 States, 2004-2005. MMWR. Morbidity and Mortality Weekly Reports, 57(14), 361-366. Retrieved from http://www.cdc.gov/mmWR/preview/mmwrhtml/mm5714al.htm
Chaudron, L. H. (2003). Postpartum depression: what pediatricians need to know.
Pediatrics in Review / American Academy of Pediatrics, 24(5), 154-61. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/12728188
Chaudron, L. EL, Szilagyi, P. G., Tang, W., Anson, E., Talbot, N. L., Wadkins, H. I. M.,
... Wisner, K. L. (2010). Accuracy of depression screening tools for identifying postpartum depression among urban mothers. Pediatrics, 125(3), e609-e617. Retrieved from
http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=3030186&tool=pmcentr ez&rendertype=abstract
30


Cicchetti, D., Rogosch, F. a, & Toth, S. L. (2000). The efficacy of toddler-parent psychotherapy for fostering cognitive development in offspring of depressed mothers. Journal of Abnormal Child Psychology, 28(2), 135-48. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/10834766
Cudmore, Z. (2011). The Beck Depression Inventory-II Factor Structure in Postpartum and Implications for Maternal Sensitivity (Honors Thesis). University of Colorado Denver.
Cutrona, C. E., & Troutman, B. R. (1986). Social support, infant temperament, and parenting self-efficacy: a mediational model of postpartum depression. Child Development, 57(6), 1507-18. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/3802975
Dennis, C.-L., & Dowswell, T. (2013). Psychosocial and psychological interventions for preventing postpartum depression. Cochrane Database of Systematic Reviews (Online), 2, CD001134. doi:10.1002/14651858.CD001134.pub3
Dietz, L. J., Jennings, K. D., Kelley, S. A., & Marshal, M. (2009). Maternal depression, paternal psychopathology, and toddlers behavior problems. Journal of Clinical Child and Adolescent Psychology: The Official Journal for the Society of Clinical Child and Adolescent Psychology, American Psychological Association, Division 53, 38(1), 48-61. doi:10.1080/15374410802575362
Dozois, D. J. A., Dobson, K. S., & Ahnberg, J. L. (1998). A psychometric evaluation of the Beck Depression Inventory-II. Psychological Assessment, 10(2), 83-89. doi:10.1037/1040-3590.10.2.83
Earls, M. F. (2010). Incorporating recognition and management of perinatal and postpartum depression into pediatric practice. Pediatrics, 126(5), 1032-1039. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/20974776
Engel, G. L. (1977). The need for a new medical model: a challenge for biomedicine. Science (New York, N.Y.), 796(4286), 129-36. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/847460
Fava, G. A., & Sonino, N. (2008). The biopsychosocial model thirty years later. Psychotherapy andPsychosomatics, 77(1), 1-2. doi: 10.1159/000110052
Field, T. (2010). Postpartum depression effects on early interactions, parenting, and safety practices: a review. Infant Behavior & Development, 33(1), 1-6. doi: 10.1016/j.infbeh.2009.10.005
First, M., Spitzer, R. L., Gibbon, M., & Williams, J. B. (2002). Structured Clinical Interview for DSM-IV-TR Axis I Disorders, Research Version, Patient Edition. (SCID-I/P). New York: Biometrics Research, New York State Psychiatric Institute.
31


Flykt, M., Kanninen, K., Sinkkonen, J., & Punamaki, R. (2010). Maternal depression and dyadic interaction: the role of maternal attachment style. Infant Child Development, (19), 530-550.
Gavin, N., & Gaynes, B. (2005). Perinatal depression: a systematic review of prevalence and incidence. Obstetrics & ..., 106(5 Pt 1), 1071-83. doi: 10.1097/0 l.AOG. 0000183 597.31630. db
Grothe, K. B., Dutton, G. R., Jones, G. N., Bodenlos, J., Ancona, M., & Brantley, P. J. (2005). Validation of the Beck Depression Inventory-II in a low-income African American sample of medical outpatients. Psychological Assessment, 17(1), 110-4. doi:10.1037/1040-3590.17.1.110
Gu, H., Hu, D., Hong, X., Mao, J., Cui, Y., Hui, N., & Sha, J. (2003). [Changes and significance of orphanin and serotonin in patients with postpartum depression], Zhonghua Fu Chan Ke Za Zhi, 35(12), 727-8. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/14728841
Hay, D. F. (1997). Postpartum depression and cognitive development. In L. Murray & P. J. Cooper (Eds.), Postpartum depression and child development (pp. 85-110). New York: The Guilford Press.
Howell, E. A., Mora, P. A., Horowitz, C. R., & Leventhal, H. (2005). Racial and ethnic differences in factors associated with early postpartum depressive symptoms. Obstetrics and Gynecology, 105(6), 1442-50. doi: 10.1097/0 l.AOG. 0000164050.34126.37
Hu, L.-T., & Bentler, P. M. (1999). Cutoff criteria for fit indices in covariance structure analysis: conventional criteria versus new alternatives. Structural Equation Modeling, 6( 1), 1-55.
IBM Corp. (2012). IBM SPSS Statistics for Mac, Version 21.0. Armonk, NY.
Jones, H. W., & Venis, J. A. (2001). Identification and classification of postpartum
psychiatric disorders. Journal of Psychosocial Nursing and Mental Health Services, 39(12), 23-30. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/11775292
Joreskog, K. G. (1969). A general approach to confirmatory maximum likelihood factor analysis. Psychometrika, 34(2), 183-202. doi:10.1007/BF02289343
Josefsson, A., & Sydsjo, G. (2007). A follow-up study of postpartum depressed women: recurrent maternal depressive symptoms and child behavior after four years. Archives of Womens Mental Health, 10(4), 141-5. doi:10.1007/s00737-007-0185-9
32


Kaplan, P. S., Danko, C. M., Diaz, A., & Kalinka, C. J. (2011). An associative learning deficit in 1-year-old infants of depressed mothers: role of depression duration. Infant Behavior & Development, 34( 1), 35-44. doi:10.1016/j.infbeh.2010.07.014
Kaplan, P. S., Danko, C. M., Kalinka, C. J., & Cejka, A. M. (2012). A developmental decline in the learning-promoting effects of infant-directed speech for infants of mothers with chronically elevated symptoms of depression. Infant Behavior & Development, 35(3), 369-379. doi:10.1016/j.infbeh.2012.02.009
Kneipp, S. M., Kairalla, J. A., Stacciarini, J., & Pereira, D. (2009). The Beck Depression Inventory II factor structure among low-income women. Nursing Research, 58(6), 400-9. doi:10.1097/NNR.0b013e3181bee5aa
Lawson, K. R., & Ruff, H. A. (2004). Early attention and negative emotionality predict later cognitive and behavioural function. International Journal of Behavioral Development, 28 (2 ), 157-165. doi:10.1080/01650250344000361
Lobbestael, J., Leurgans, M., & Arntz, A. (2010). Inter-rater reliability of the Structured Clinical Interview for DSM-IV Axis I Disorders (SCID I) and Axis II Disorders (SCID II). Clinical Psychology & Psychotherapy, 75(1), 75-9. doi: 10.1002/cpp. 693
Lundy, B. L., Field, T., Cuadra, A., Nearing, G., Cigales, M., & Hashimoto, M. (1996). Mothers with Depressive Symptoms Touching Newborns. Early Dev. Parent, (5), 129-134.
MacCallum, R. C., Widaman, K. F., Zhang, S., & Hong, S. (1999). Sample size in factor analysis. Psychological Methods, (4), 84-99.
Manian, N., Schmidt, E., Bomstein, M. H., & Martinez, P. (2013). Factor structure and clinical utility of BDI-II factor scores in postpartum women. Journal of Affective Disorders, 149(1-3), 259-68. doi: 10.1016/j jad.2013.01.039
Mann, J. J. (1999). Role of the serotonergic system in the pathogenesis of major
depression and suicidal behavior. Neuropsychopharmacology: Official Publication of the American College of Neuropsychopharmacology, 21(2 Suppl), 99S-105S. doi: 10.1016/S0893 -13 3X(99)00040-8
Milgrom, J., Westley, D. T., & Gemmill, A. W. (2004). The mediating role of maternal responsiveness in some longer term effects of postnatal depression on infant development. Infant Behavior and Development, 27(4), 443-454. doi:10.1016/j.infbeh.2004.03.003
Murray, L. (1992). The impact of postnatal depression on infant development. Journal of Child Psychology and Psychiatry, and Allied Disciplines, 33(3), 543-61. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/1577898
33


Murray, L., Cooper, P. J., Wilson, A., & Romaniuk, H. (2003). Controlled trial of the short- and long-term effect of psychological treatment of post-partum depression: 2. Impact on the mother-child relationship and child outcome. The British Journal of Psychiatry: The Journal of Mental Science, 182, 420-7. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/12724245
Muthen, L. K., & Muthen, B. O. (2011). Mplus Users Guide (6th Edition) (Sixth Edit.). Los Angeles, CA: Muthen & Muthen.
NICHD Early Child Care Research Network. (1999). Chronicity of maternal depressive symptoms, maternal sensitivity, and child functioning at 36 months. NICHD Early Child Care Research Network. Developmental Psychology, 35(5), 1297-310. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/10493655
Nylen, K. J., OHara, M. W., Brock, R., Moel, J., Gorman, L., & Stuart, S. (2010). Predictors of the longitudinal course of postpartum depression following interpersonal psychotherapy. Journal of Consulting and Clinical Psychology, 75(5), 757-63. doi: 10.1037/a0020623
OHara, M. W., & McCabe, J. E. (2013). Postpartum depression: current status and future directions. Annual Review of Clinical Psychology, 9, 379-407. doi: 10.1146/annurev-clinpsy-050212-185612
OHara, M. W., Neunaber, D. J., & Zekoski, E. M. (1984). Prospective study of postpartum depression: prevalence, course, and predictive factors. Journal of Abnormal Psychology, 93(2), 158-71. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/6725749
OHara, M. W., Rehm, L. P., & Campbell, S. B. (1982). Predicting depressive
symptomatology: cognitive-behavioral models and postpartum depression. Journal of Abnormal Psychology, 91(6), 457-61. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/7153419
OHara, M. W., Schlechte, J. A., Lewis, D. A., & Wright, E. J. (1991a). Prospective study of postpartum blues. Biologic and psychosocial factors. Archives of General Psychiatry, 48(9), 801-6. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/1929770
OHara, M. W., Schlechte, J. A., Lewis, D. A., & Wright, E. J. (1991b). Prospective study of postpartum blues. Biologic and psychosocial factors. Archives of General Psychiatry, 48(9), 801-6. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/1929770
OHara, M. W., Zekoski, E. M., Philipps, L. H., & Wright, E. J. (1990). Controlled prospective study of postpartum mood disorders: comparison of childbearing and
34


nonchildbearing women. Journal of Abnormal Psychology, 99(1), 3-15. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/2307763
Rich-Edwards, J. W., Kleinman, K., Abrams, A., Harlow, B. L., McLaughlin, T. J., Joffe, H., & Gillman, M. W. (2006). Sociodemographic predictors of antenatal and postpartum depressive symptoms among women in a medical group practice. Journal of Epidemiology and Community Health, 60(3), 221-7. doi: 10.1136/jech.2005.039370
Salmela-Aro, K., Nurmi, J. E., Saisto, T., & Halmesmaki, E. (2001). Goal reconstruction and depressive symptoms during the transition to motherhood: evidence from two cross-lagged longitudinal studies. Journal of Personality and Social Psychology, 81(6), 1144-59. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/11761314
Segre, L. S., OHara, M. W., Arndt, S., & Stuart, S. (2007). The prevalence of
postpartum depression: the relative significance of three social status indices. Social Psychiatry and Psychiatric Epidemiology, 42(4), 316-21. doi : 10.1007/s00127-007-0168-1
Sharma, V., & Sharma, P. (2012). Postpartum depression: diagnostic and treatment issues. Journal of Obstetrics and Gynaecology Canada: JOGC = Journal D obstetrique et Gynecologie Du Canada: JOGC, 34(5), 436-42. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/22555136
Shear, M. K., Greeno, C., Kang, J., Ludewig, D., Frank, E., Swartz, H. A., & Hanekamp, M. (2000). Diagnosis of nonpsychotic patients in community clinics. The American Journal of Psychiatry, 157(4), 581-7. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/10739417
Siegert, R. J., Walkey, F. H., & Turner-Stokes, L. (2009). An examination of the factor structure of the Beck Depression Inventory-II in a neurorehabilitation inpatient sample. Journal of the International Neuropsychological Society: JINS, 75(1), 142-7. doi:10.1017/S1355617708090048
Skre, I., Onstad, S., Torgersen, S., & Kringlen, E. (1991). High interrater reliability for the Structured Clinical Interview for DSM-III-R Axis I (SCID-I). Acta Psychiatrica Scandinavica, 84(2), 167-73. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/1950612
Smarr, K. L., & Keefer, A. L. (2011). Measures of depression and depressive symptoms: Beck Depression Inventory-II (BDI-II), Center for Epidemiologic Studies Depression Scale (CES-D), Geriatric Depression Scale (GDS), Hospital Anxiety and Depression Scale (HADS), and Patient Health Questionna. Arthritis Care & Research, 63 Suppl 1, S454-66. doi:10.1002/acr.20556
35


Spinelli, M. G. (2004). Maternal infanticide associated with mental illness: prevention and the promise of saved lives. The American Journal of Psychiatry, 161(9), 1548 57. doi:10.1176/appi.ajp. 161.9.1548
Stage, F. K., Carter, H. C., & Nora, A. (2004). Path Analysis: An Introduction and
Analysis of a Decade of Research. The Journal of Educational Research, 98( 1), 5-13. doi:10.3200/JOER.98.1.5-13
Stanley, C., Murray, L., & Stein, A. (2004). The effect of postnatal depression on mother-infant interaction, infant response to the Still-face perturbation, and performance on an Instrumental Learning task. Development and Psychopathology, 16(1), 1-18. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/15115062
Stein, A., Arteche, A., Lehtonen, A., Craske, M., Harvey, A., Counsell, N., & Murray, L. (2010). Interpretation of infant facial expression in the context of maternal postnatal depression. Infant Behavior & Development, 33(3), 273-8. doi:10.1016/j.infbeh.2010.03.002
Steiner, J. L., Tebes, J. K., Sledge, W. H., & Walker, M. L. (1995). A comparison of the structured clinical interview for DSM-III-R and clinical diagnoses. The Journal of Nervous and Mental Disease, 183(6), 365-9. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/7798084
Steiner, M. (1998). Perinatal mood disorders: position paper. Psychopharmacology Bulletin, 34(3), 301-6. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/9803759
Storch, E. A., Roberti, J. W., & Roth, D. A. (2004). Factor structure, concurrent validity, and internal consistency of the Beck Depression Inventory-Second Edition in a sample of college students. Depression and Anxiety, 19(3), 187-9. doi: 10.1002/da.20002
Tabachnick, B. G., & Fidell, L. S. (2012). Using Multivariate Statistics. Pearson Education, Limited. Retrieved from http: //b ooks. googl e. com/b ooks?i d=ucj 1 y g A AC A A J
Teti, D. M., & Gelfand, D. M. (1991). Behavioral competence among mothers of infants in the first year: the mediational role of maternal self-efficacy. Child Development, 62(5), 918-29. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/1756667
Vik, T., Grote, V., Escribano, J., Socha, J., Verduci, E., Fritsch, M., ... Koletzko, B. (2009). Infantile colic, prolonged crying and maternal postnatal depression. Acta Paediatrica (Oslo, Norway: 1992), 98(8), 1344-8. doi: 10.1111/j.1651-2227.2009.01317.x
36


Viljoen, J. L., Iverson, G. L., Griffiths, S., & Woodward, T. S. (2003). Factor Structure of the Beck Depression InventoryII in a Medical Outpatient Sample. Journal of Clinical Psychology in Medical Settings, 10(4), 289-291.
Ward, L. C. (2006). Comparison of factor structure models for the Beck Depression InventoryIF Psychological Assessment, 18( 1), 81-8. doi: 10.1037/1040-3590.18.1.81
Weissman, M. M., Pilowsky, D. J., Wickramaratne, P. J., Talati, A., Wisniewski, S. R., Fava, M., ... Rush, A. J. (2006). Remissions in maternal depression and child psychopathology: a STAR*D-child report. JAMA : The Journal of the American Medical Association, 295(12), 1389-98. doi:10.1001/jama.295.12.1389
Williams, J. B., Gibbon, M., First, M. B., Spitzer, R. L., Davies, M., Borus, J., ...
Rounsaville, B. (1992). The Structured Clinical Interview for DSM-III-R (SCID). II. Multisite test-retest reliability. Archives of General Psychiatry, 49(8), 630-6. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/1637253
Workman, J. L., Barha, C. K., & Galea, L. A. M. (2012). Endocrine substrates of cognitive and affective changes during pregnancy and postpartum. Behavioral Neuroscience, 126(1), 54-72. doi:10.1037/a0025538
Wright, S. (1923). The Theory of Path Coefficients a Reply to Niless Criticism.
Genetics, 5(3), 239-55. Retrieved from
http://www.pubmedcentral.nih. gov/articlerender.fcgi?artid=1200747&tool=pmcentr ez&rendertype=abstract
Wright, S. (1934). The method of path coefficients. Ann. Math. Statist., (5), 161-215.
Zanarini, M. C., & Frankenburg, F. R. (2001). Attainment and maintenance of reliability of axis I and II disorders over the course of a longitudinal study. Comprehensive Psychiatry, 42(5), 369-74. doi: 10.1053/comp.2001.24556
Zanarini, M. C., Skodol, A. E., Bender, D., Dolan, R., Sanislow, C., Schaefer, E., ... Gunderson, J. G. (2000). The Collaborative Longitudinal Personality Disorders Study: reliability of axis I and II diagnoses. Journal of Personality Disorders, 14(4), 291-9. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/11213787
37


Full Text

PAGE 1

LATENT FACTOR STRUCTURE OF THE BDI II AT 4 MONTHS POSTPARTUM by RYAN M. ASHERIN B.A., Metropolitan State University of Denver, 2004 A thesis submitted to the Faculty of the Graduate School of the University of Colorado in partial fulfillment of the requirements for the degree of Master of Arts Clinical Health Psychology Program 2014

PAGE 2

ii This thesis for the Master's of Arts degree by Ryan M. Asherin has been approved for the Clinical Health Psychology Program by Kevin Everhart Chair Pete r Kaplan Krista Ranby January 29 2014

PAGE 3

iii Asherin, Ryan M. (M.A., Clinical Health Psychology) Latent Factor Structure of the BDI II at 4 Months Postpartum Thesis directed by Professor Kevin Everhart ABSTRACT BACKGROUND: Depression in the postpartum perio d (DPP) is the most common psychiatric disorder associated with childbirth, negatively impacting the mother infant dyad. The manifestation of DPP symptoms has been attributed to a combination of biopsychosocial stressors. Current standard of care recommend s routine screening for maternal mood disorders at infant well child checks through the first four months of life, and the Beck Depression Inventory, 2 nd Edition (BDI II) has been validated for use in the postpartum period to assess maternal depressive sym ptoms. The present study tests a theoretically driven model emphasizing the role of intrapersonal cognitions ( cognitive appraisal ) on the report of DPP as compared to somatic symptoms ( somatic exhaustion ) and interpersonal ( impairment ) factors. METHODS: A sample (N=100) of mothers four months postpartum were assessed for depressive symptoms using the BDI II and a Structured Clinical Interview for DSM IV Axis I Disorders (SCID). Confirmatory factor analysis was conducted to identify the latent factor struct ure of the BDI II, and then a path analysis was conducted to test the hypothese s that a cognitive appraisal latent factor on the BDI II drives elevated scores and correlates with a SCID diagnosis of depression. RESULTS: A confirmatory factor analysis supp orted a three factor model of the BDI II with cognitive appraisal, somatic exhaustion, and impairment latent factors identified ( x 2 (186) =211.386, ns ; CFI=0.947; RMSE A =0.037) A path analysis supports the proposed hypothesis based on good model fit estimate s ( x 2 (204) =231.917, ns ; CFI=0.945; RMSE A =0.037), and the proposed influence of cognitive appraisal on somatic exhaustion ( = .754) and impairment ( = .481) was supported. Additionally, cognitive appraisal demonstrated the highest correlation (r = .663) with SCID diagnosis, compared to either somatic exhaustion (r = .458) or impairment (r = .012). CONCLUSIONS: This study su pports a theoretical model emphasizing the role of cognitive appraisal in the development of DPP, suggesting that elevated responses to cognitive appraisal related items on the BDI II might warrant clinical assessment of DPP. The form and content of this abstract are approved. I recommend its publication. Approved: Kevin Everhart

PAGE 4

iv DEDICATION I dedicate this work to my wife and parents Jess: you are my biggest supporter and best friend. Your love, strength, and support have given me the motivation to c ontinue to pursue my career goals while building a beautiful life together. Thank you for your unconditional love and understanding over the years. My graduate school experience has been so much more fulfilling because I get to share it with you. Mom and D ad: I have spent a lot of time and energy throughout my academic career trying to gain a better understanding of infant and childhood development. I still have a lot to learn, but I am certain that none of the achievements throughout my life would have bee n possible without your love, support, and prayers. Thank you for all that you have given me, especially the confidence to believe in myself and never stop pursuing my dreams.

PAGE 5

v ACKNOWLEDGMENTS I would like to thank the Infant Lab at the University of Color ado Denver, especially my mentors, Drs. Kevin Everhart and Peter Kaplan. I have benefitted from having two primary advisors that are each able to nurture my intellectual growth while offering constructive criticism and praise. Within the Infant Lab I would like to acknowledge the support of my fellow graduate students, Jo Vogelli and Shiva Fekri. I appreciate all the time we have spent together pursuing shared interests and ambitions. I especially appreciate the support they offered in the conceptual d evelo pment of my thesis work. I a m looking forward to several more productive years of us all working together in the Infant Lab. I would also like to thank Dr. Krista Ranby. I am grateful for her contributions to our graduate training program, and especially f or the guidance she provided in the conceptual and statistical development of my thesis. Last, but not least, I would like to thank all the original members of my PhD cohort, Kellie Martins, Megan Grigsby, Stephanie Hooker, Carissa Kinman, and Jenn Altma n. Each has been an inspiration to me over the years, and I am truly thankful to be able to call them my friends. In particular, I would like to thank Kellie, Megan, and Stephanie for making my time within the program one that has been filled with as much laughter as hard work.

PAGE 6

vi TABLE OF CONTENTS CHAPTER I BACKGROUND ................................ ................................ ................................ .......... 1 Characterizing postpartum mood disorders ................................ ............................. 1 Significance of depression in the postpartum period ................................ .............. 3 Biopsychosocial pathways for developing depression in the postpartum period .... 4 Theoretical models for developing depression in the postpartum period ................ 6 T heoretical approach fo r developing depression in the postpartum period ............. 6 Hypothesis and specific aims ................................ ................................ .................. 9 II METHODS ................................ ................................ ................................ ................. 10 Sample ................................ ................................ ................................ ................... 10 Participants ................................ ................................ ................................ ............ 10 Procedure ................................ ................................ ................................ ............... 11 Measures ................................ ................................ ................................ ................ 12 Statistical analyses ................................ ................................ ................................ 13 Power ................................ ................................ ................................ ..................... 14 Data cleaning & descriptive statistics ................................ ................................ ... 14 Aim 1: Confirmatory factor analysis ................................ ................................ ..... 15 Aim 2: Path analysis ................................ ................................ .............................. 16 III RESULTS ................................ ................................ ................................ ................... 18 Aim 1: Confirmatory factor analysis ................................ ................................ ..... 18 Aim 2: Pa th analysis ................................ ................................ .............................. 22 IV DISCUSSION ................................ ................................ ................................ ............ 24 REFERENCES ................................ ................................ ................................ .................. 29

PAGE 7

vii LIST OF TABLES T able 1. Maternal demographic and diagnostic data ................................ ................................ ... 11 2. Correlations for CFA ................................ ................................ ................................ ..... 20 3. Standardized and Unstandardized Coefficients for CFA ................................ .............. 21

PAGE 8

viii LIST OF FIGURES Figure 1. Hypothesized Confirmatory Factor Analysis: BDI II 3 Factor Model ......................... 16 2. Hypothesized Path Analysis: BDI II & SCID ................................ ............................... 17 3. Confirmatory Factor Analysis: BDI II 3 Factor Model ................................ ............... 22 4. Path Analysis: BDI II & SCID ................................ ................................ ...................... 23

PAGE 9

ix LIST OF ABBREVIATIONS BDI II Beck Depression Inventory, 2nd Edition DPP Depression in the postpartum period PPD Postpartum Depression SCID Structured Clinical Interview for Diagnosis SEM Struct ural Equation Modeling

PAGE 10

1 CHAPTER I BACKGROUND T he postpartum period is often characterized as a joyful time in the l ives of mothers and their families However, incidence and prevalence rates identify this period as a major risk factor for the develop ment of a mood disorder. Postpartum depression is commonly used as a catchall phrase to describe any sort of depressive symptom associated with childbirth (Jones & Venis, 2001) However, this definition is an overgeneralization of a variety of symptoms that can vary in severity, duration, and onset. Categorically, mood disorders occurring in the postpartum period can be classified as baby blues, postpartum depression, and postpartum psychosis (C. T. Beck, 2006) Characterizing postpartum mood d isorders The term "baby blues" describe s a transient increase in depressive symp toms shortly, usually within 10 days, after childbirth. Baby blues are often considered to be "normal" as they affect up to 85% of new mothers (Berggren Clive, 1998; Dennis & Dowswell, 2013; O'Hara, Zekoski, Philipps, & Wright, 1990; M. Steiner, 1998) and generally resolve without intervention within 2 weeks of onset (O'Hara, Schlechte, Lewis, & Wright, 1991a) Postpartum psychosis, on the other hand, is a relatively rare condition affecting just 1 in every 1,000 mothe rs after having a live birth and is characterized by the presence of hallucinations or delusions (Spinelli, 2004) Postpartum psychosis typically manifests within 2 weeks of childbirth an d should be considered a psychiatric emergency that requires immediate intervention due to the association between postpartum psychosis and suicide or neonaticide/infanticide (Spinelli, 200 4) Statistically, postpartum depression (PPD) tends to fall somewhere between blues and

PAGE 11

2 psychosis, in that PPD affects up to 15% of new mothers with depressive symptoms ranging from mild to severe (Gavin & Gaynes, 2005; O'Hara, Neunaber, & Zekoski, 1984) PPD is considered a non psychotic depressive episode that either begins in or extends into the postpartum period. In Colorado, de pression affects approximately 14% of mothers within the first year after giving birth (Centers for Disease Control and Pre vention, 2008) In research and healthcare settings PPD is often assessed using a self repor t measure of depression. Among the most common self report screening assessments used today are the Edinburgh Postnatal Depression Scale (EPDS), the Beck Depressio n Inventory, 2 nd Edition (BDI II), and the Postpartum Depression Screening Scale (PDSS). A recent study found that all three self report measures performed equally well at detecting PPD, however traditional cutoff scores used with each measure could be adj usted to increase sensitivity and specificity of clinical diagnosis (Chaudron et al., 2010) Clinically, PPD is diagnosed using a postpartum onset specifier added to a current major depressive episode of major depressive disorder, biopolar I disorder, or biopolar II disorder (American Psychiatric Association, 2000, 2013) The Diagnostic Manual of Mental Disorders, Fifth Edition (DSM V) offers the postpartum onset specifier for episodes of a mood disorder that occur within the first four we eks after childbirth (American Psychiatric Association, 2013) H owever empirical research suggest s that PPD can manifest throughout the first year postpartum with peak inciden ce of PPD occurring within the first four months of childbirth (Gavin & Gaynes, 2005; O'Hara et al., 1990) Due to the inconsistency in the use of PPD terminology, the term depression in the postpartum period ( DPP ) will be used throughout the remainder of this paper.

PAGE 12

3 Significance of depression in the postpartum period The effects of DPP are well documented, as it is the most common psychiatric comorbidity associated with childbirth and has been found to have negative effects for both members of the mother infant dyad (Sharma & Sharma, 2012; Weissman et al., 2006) Symptoms of DPP tend to persist for quite some time averging 5 months in the general population (Chaudron, 2003) and mothers are at increased risk of developing subsequent major depressive episodes throughout their lifetime (Josefsson & Sydsjš, 2007; Lynne Murray, Cooper, Wilson, & Romaniuk, 2003; Nylen et al., 2010) Additionally, depressed mothers tend to experience increased negative emotional ity (Dietz, Jennings, Kelley, & Marshal, 2009; Lundy et al., 1996) which may be partly due to cognitive appraisal deficits that heighten self focus (Salmela Aro, Nurmi, Saisto, & Halmesmaki, 2001; Stein et al., 2010; Teti & Gelfand, 1991) Potentially even more concerning is the effects DPP can have on parenting behaviors and, thus, child development. For the past twenty years theorists have attributed deficits in infant development and cognition to disruptions in the depressed caregivers' ability to support their infant's state and behavior, which in turn contribute to deficits in infants' abilities to attend to and extract information about environmental contingencies (Hay, 1997) This framework can be understood as a dynamic relationship between parenting behaviors that either reinforce or undermine an infant's social learning abilities. Many studies have demonstrated that DPP is often accompanied by an impoverishment in the quantity and quality of pro social parenting behaviors Absence of pro social parenting behaviors are linked with poorer cognitive development in infants, as well as delayed school readiness, and decreased I Q scores later in childhood (Cicchetti, Rogosch, & Toth, 2000; Hay, 1997; Milgrom, Westley, & Gemmill, 2004; Murray, 1992; NICHD, 1999; Stanley,

PAGE 13

4 Murray, & St ein, 2004) Additional studies have demonstrated that the frequency of mother child interactions are not only reduced if the mother suffers from DPP but in some cases their reactions to their child can be characterized as hostile or unresponsive (Beebe et al., 2007; Dietz et al., 2009; Field, 2010; Flykt, Kanninen, Sinkkonen, & Punamaki, 2010) Taken together, we begin to see the deleterious effects of DPP for bot h the mother and child, as research has identified increased negative emotionality in caregivers and decreased attention to infants as predictive factors for cognitive deficits and behavior problems during childhood development (Lawson & Ruff, 2004) Biopsychosocial pathways for developing depression in the postpartum period Conceptually, the manifestation of DPP symptoms are attributed to a combination of biopsychosocial stressors associated with childbirth and caregiving (Berggren Clive, 1998) Independently, neither physical nor psychosocial stress has been shown to cause DPP To date no evidence has found a direct hormonal imbalance as the determinant for DPP (O'Hara & McCabe, 2013) however several studies have implicated physiological changes in estradiol and progesterone levels as contributing factors in the development of mood dysregulation in the postpartum period (Bloch et al., 2000, 2005; Bloch, Daly, & Rubinow, 2003) Even if estradiol an d progesterone levels can be correlated with DPP this explanation may be overly simplistic as estradiol and progesterone interact with neurotransmitter systems that directly affect serotonin and dopamine uptake. Serotonin has been found to regulate fundam ental physiologic responses such as sleep and appetite, and a global decrease in serontinergic function has been associated with major depression (Gu et al., 2003; Mann, 1999) Additionally, animal studies have demonstrated that a sharp reduction in the functioning of dopaminergic pathways precede depressive symptoms (Byrnes, Byrnes, & Bridges, 2001) However, despite these findings, there has

PAGE 14

5 been little evidence to suggest that levels of estrogen, progesterone, or cortisol vastly differentiate between depressed and nondepressed mothers in the postpart um period (Bloch et al., 2003; Workman, Barha, & Galea, 2012) From a psychosocial perspective, disadvantaged social conditions as well as stress accompanying the birth of a child have been associated with the development of DPP Demographic markers (i.e. race/eth nicity, education status, and younger age) related to poverty levels have consistently been found to be risk factor s for developing DPP (Howell, Mora, Horowitz, & Leventhal, 2005; Rich Edwards et al., 2006; Segre, O'Hara, Arndt, & Stuart, 2007) These findings also hold true for mothers in Colorado, as maternal age (< 20), race/ethnicity (non white), marital status (unmarried), education (<12 years), and insurance status (not insured) have all been fo und to be statistically significant predictors of self reported symptoms of DPP (Centers for Disease Control and Prevention 2008) Additional studies have looked at health and social risk factors occurring either perinatally or postpartum. Perinatal health complications, including preeclampsia, perinatal hospitalization, and emergency caesarean delivery, have been identified as risk factors for developing DPP even while controlling for social economic status and pre existing conditions (Blom et al., 2010) Relevant factors occurring during the postpartum period, such as infant health and temperament have also been identified as risk factors for developing DPP (Blom et al. 2010; Vik et al., 2009) Similarly, data from Colorado has suggested that pregnancy related risk factors, such as tobacco use, physical abuse, and financial, partner related, or traumatic stressors, affect the development of postpartum depressive sympto ms (Centers for Disease Control and Prevention, 2008)

PAGE 15

6 Theoretical models for developing depression in the postpartum per iod Theoretically, several models have been explored in prior research attempting to explain the onset of DPP An early cognitive behavioral model associated negative attributions about one's self and other negative psychological vulnerabilities to the dev elopment of DPP (O'Hara, Rehm, & Campbell, 1982) Expanding the cognitive behavioral model to include psy chosocial stress has been found to better predict the development of DPP (O'Hara, Schlech te, Lewis, & Wright, 1991b) infant temperament and interpersonal support (Cutrona & Troutman, 1986) One of the more recent models predicted DPP by analyzing cognitive dissonance between expected and actual social support received by mothers postpartum (C. T. Beck, 2002) Although several models have demonstrated conceptual and theoretical support for various biopsychosocial aspects to contribute to the development of DPP recent literature has called for additional exploration into new model development and empirical testing with w ell defined populations to further the current understanding of why DPP manifests in ~15% of new mothers (O'Hara & McCabe, 2013) Theoretical approach for developing depression in the postpartum period The current proposal aims to test specific factors tha t may influence the development of DPP in a sample of mothers at four months postpartum. Collecting data on mothers who are all four months postpartum provides insigh t into unique challenges faced during this time point. Furthermore, conducting analyses on the depressive s ymptoms of mothers at this time point may be particularly useful as it falls within the recommended time frame for maternal depression screening put forth by the American Academy of Pediatrics (Earls, 2010) as well as near the peak incidence of new onset cases of depression occurring in the postpartum period (Chaudron, 2003) The conceptual

PAGE 16

7 framework and theoretical model influencing this study was largely shaped by the biopsychosocial model, which emphasizes the dynamic interaction between biological, psychological, and social variables that may manifest as categorically defined psychological conditions (Engel, 1977; Fava & Sonin o, 2008) As discussed above, recent literature has demonstrated that numerous biological and psychosocial factors are predictive or associated with DPP yet brief self report questionnaires are commonly used to assess PPD. To better understand the under lying factors associated with self reported DPP symptoms a recent study conducted exploratory and confirmatory factor analyses to identify the latent factor structure of the BDI II. Findings from a sampl e of mothers three months post partum found a three fa ctor structure (i.e. Cognitive Affective, Somatic Anxiety, and Guilt) to be the most parsimonious (Carvalho Bos et al., 2009) A replication study conducted on data obtained by the Infant Lab at the Universit y of Colorado Denver also supported a three factor structure for mothers four months postpartum. An exploratory factor analysis demonstrated that the most parsimonious fit indicated that Cognitive Appraisal, Somatic Exhaustion, and Impairment factors explained the majority of total variance of the item responses obtained on the BDI II (Cudmore, 2011) The most recent study looking at the factor structure of the BDI II included a large sample of postpartum women and also found a three factor solution (i.e. Cognitive, Appraisal, and Somatic) to provide the best fit for th e data (Manian, Schmidt, Bornstein, & Martinez, 2013) These three studies identified factor structure constructs that are, at least nominally related to biological (Somatic Anxiety; Somatic Exhaustion; & Somatic, respectively) and psychosocial (Cognitive Affective; Cognitive Appraisal; Cognitive, respectively) symptoms of depression categorized in the DSM (American

PAGE 17

8 Psychiatric Association, 2000, 2013) Additionally, cognitive, somatic, and affective components are consistently reported for the BDI II when tested with a variety of populations (Grothe et al., 2005; Kneipp, Kairalla, Stacciarini, & Pereira, 2009; Siegert, Walkey, & Turner St okes, 2009; Storch, Roberti, & Roth, 2004; Viljoen, Iverson, Griffiths, & Woodward, 2003; Ward, 2006) Based on the literature conducting factor analysis on the BDI II on samples of mothers in the postpartum and on the literature offering support for b oth biological and psychosocial factors that contribute to the development of DPP symptoms, I hypothesize that factors related to cognitive appraisal drive elevated scores on a measurement of DPP symptoms Theoretically, I posit that cognitive appraisal re lated items on the BDI II are endorsed by individuals that have developed psychological vulnerabilities (e.g. negative attributions about oneself, cognitive distortions, or increased self focus ), which lead to elevated responses o n somatic exhaustion and i mpairment related items. Although difficult to empirically test, assumptions are made that all women in the postpartum period undergo a rapid hormonal change that represents a biological stressor and that these changes are partially c aptured by responses to somatic related questions on the BDI II. Furthermore, although all new mothers experience somatic changes during the postpartum period, it is the cognitive appraisal aspect of depression that distinguishes mothers that develop DPP Again, difficult to e mpirically test, but an assumption is made that elevated item responses on the BDI II that are related to cognitive appraisal are shaped by psychosocial factors that increase a mother's self focus. Increased self focus leads to elevated scores on the BDI I I as mothers are more likely to judge symptoms more severely than mothers not experiencin g cognitive appraisal deficits. Despite a lack of empirical research directly relating items on the BDI II to specific biological or

PAGE 18

9 psychosocial stressors, the propos ed model offers a biopsychosocial pathway for understanding the development of DPP The present study attempts to replicate the factor structure of the BDI II previously reported in the literature through a confirmatory factor analysis, while also designin g and testing a path analysis emphasizing the role of cognitive appraisal in the development of DPP by four months postpartum. To date, no publications have been reported in the extant literature testing a path analysis for DPP utilizing the BDI II 1 2 Hyp othesis and specific aims The overarching hypothesis of this study is that the distinct factor structure of the BDI II, as reported by a sample of mothers at four months postpartum, can offer a better understanding of the categorically related symptoms of DPP while providing support for a structural equat ion model (SEM) illustrating cognitive appraisal 's role in driving elevated scores for DPP This hypothesis will be explored through two distinct aims: 1. Test the hypothesized three factor structure (i.e. cog nitive appraisal, somatic exhaustion, and impairment) on the BDI II using confirmatory factor analysis 2. Test whether a conceptual model emphasizing the influence of cognitive appraisal on the development of DPP as measured by the BDI II and a Structured Cl inical Interview for DSM IV Axis I Disorders (SCID) is supported using path analysis 1 A search was conducted on July 6, 2013 in Ovid using the followi ng search criteria: [MeSH]: Statistical, models, [MeSH]: Depression, postpartum, and [Keyword]: BDI II or [Keyword]: Beck Depression Inventory. Eight results were identified, however no results were found that specifically included a SEM. 2 A search was conducted on July 6, 2013 in CINAHL using the following search criteria: [MH]: Structural Equation Modeling, and [MH]: Depression, Postpartum. No results were identified.

PAGE 19

10 CHAPTER II METHODS Sample Data for the present study was collected as part of a larger study within the Infant Lab at the University of Colorado Denver. The larger study aimed to assess infant's attention to infant directed speech and cognitive development based on the depressive status of caregivers (Kaplan, Danko, Diaz, & Kalinka, 2011; Kaplan, Danko, Kalinka, & Cejka, 2012) The present dataset c onsists of a sample of one hundred and eight mothers of four month old mother infant dyads. Of the one hundred and eight mothers at four months postpartum participating in the above referenced study, data was missing from the measures of DPP for eight of t he mothers and, thus, these persons were excluded from analysis. One hundred mothers with complete data on all measure s of DPP were included in the final analyses. Participants Table 1 presents demographic and diagnostic information for the full sample of mothers and infants (N = 100 ). No differences were found as a function of BDI II category or depression diagnosis on maternal age, infant age, infant gender, maternal education, family income, ethnicity, or marital status. The number of children, for mothe rs with elevated compared to non elevated BDI II scores, was higher, but the number of children per caregiver did not differentiate clinically depressed from non depressed mothers.

PAGE 20

11 Table 1 Maternal demographic and diagnostic data. Variable Non Elevated (BDI II) Elevated (BDI II) NDEP (DSM IV) DEP (DSM IV) N = 100 56 4 4 8 8 12 Infant Gender (girls/boys) 28 /2 8 20/24 43 /4 5 5 / 7 Age of mother (years) 30. 1 (4. 6 ) 28.9 (5. 5 ) 30.0 (5. 2 ) 29.4 ( 4.2 ) Age of Infant (days) 131 (21. 5 ) 143 (23. 8 ) 135 ( 23.3 ) 143 (21.9) Ethnicity White Latina African American Asian Native American 33 (58.9%) 14 (25.0%) 5 (8.9%) 3 (5.4%) 1 (1.8%) 21( 47.7 %) 13 ( 29.5 %) 6 ( 13.6 %) 2 (4. 6 %) 2 (4. 6 %) 4 8 (5 4 .5%) 2 4 ( 27.3 %) 8 ( 9 1 %) 5 (5. 7 %) 3 (3. 4 %) 6 (50.0%) 3 (25.0%) 3 (25.0%) 0 (0.0%) 0 (0.0%) Marital Status Married Unmarried 45 (80.4%) 11 (19.6%) 30 (68. 2 %) 1 4 (31. 8 %) 66 (75.0%) 22 (2 5 0 %) 9 (75.0%) 3 (25.0%) Mother's Education 5. 7 (1.5) 5. 1 (1. 1 ) 5.5 (1.4) 5.3 (1.2) Family Income 6.1 (2. 4 ) 5. 7 (2.4) 5.8 ( 2.5 ) 6. 6 (1.6) Number of children 1.8 (1. 1 ) a 2.3 (1. 2 ) b 2.0 (1. 2 ) 2.1 (1.0) BDI II score 7. 8 ( 3.2 ) a 22. 3 ( 6.8 ) b 12. 5 (7. 6 ) a 26. 3 ( 8.0 ) b GAF rating 79. 4 (6. 9 ) a 69. 3 (9. 6 ) b 77.3 (7.7) a 62.1 (9.4) b Columns are defined based on BDI II scores and SCID interviews at 4 m onths. Numbers in parentheses are standard deviations. GAF = global assessment of functioning, as estimated from SCID interviews. Means with a s uper scripts significantly differed from means with b super scripts (p = .05) Procedure Mothers responding to an advertisement in a local parenting magazine were invited to participate in a study regarding the effects of depression on infant attention and learning. Mother infant dyads were greeted by research assistants and were asked to complete a study consent f orm approved by our institutional review board. The mothers then completed a demographic questionnaire and a self report measure of depr essive symptoms. With assistance from research personnel, infants completed an auditory attention paradigm before taking part in an assessment of cognitive development. Mothers were then administered a structured clinical interview by a Ph.D, M.A, or graduate level student in clinical psychology to assess for a clinical diagnosis of

PAGE 21

12 depression. The structured clinical inter view took approximately one hour to complete, and total participation time lasted about two and a half hours. Participants were compensated fifty dollars for their time. Measures Participants were administered a Structured Clinical Interview for DSM IV Axi s I Disorders (SCID), Patient Edition (SCID I/P) and were asked to complete a BDI II. The SCID is a diagnostic interview intended to specify patients that meet criteria for major mental disorders according to the DSM IV TR (First, Spitzer, Gibbon, & Williams, 2002) In cluding all versions of the SCID (i.e. nonpatient, patient, patient with psychotic screen, research, clinical trials, axis II disorders) well over seven hundred publications can be found in peer reviewed journals where the SCID was utilized for diagnostic evaluation. Several studies have demonstrated that the SCID has good reliability with Kappa values ranging from 0.66 0.93 for diagnosis of major depressive disorder (Lobbestael, Leurgans, & Arntz, 2010; Skre, Onstad, Torgersen, & Kringlen, 1991; Williams et al., 1992; Zanarini & Frankenburg, 2001; Zanarini et al., 2000) The validity of the SCID is difficult to ascertain, as the diagnosis of mental illness is largely subjective in nature, depending on the symptoms endorsed by the patient and potentially confirmatory observations made by the clinician or witness. Additionally, the SCID was developed in an attempt to standardize the process of clinical diagnosis made ideographically by trained professionals, and, thus, the SCID is often referenced as the "gold standard" as to which diagnostic procedures are compared (Shear et al., 2000; J. L. Steiner, Tebes, Sledge, & Walker, 1995) T he Beck Depression Inventory, 2 nd Edition (BDI II) is a proprietary self report assessment developed as a revision to the original Beck Depression Inventory (BDI)

PAGE 22

13 published by Psychological Corporation (A. T. Beck, Steer, & Brown, 1996) The BDI II, like its predecessor the BDI, is a twenty one item self report measure developed to detect symptoms of depression. The BDI II revision made changes reflecting revised diagnostic criteria included in the DSM IV for major depressive disorder, i.e. changes ma de to item location, wording, and time frame (Carvalho Bos et al., 2009; Dozois, Dobson, & Ahnberg, 1998; Ward, 2006) Empirical literature consistently demonstrates that the BDI II is an accurate measure of depression, occurring in the postpartum period or otherwise (Boyd, Le, & Somberg, 2005; Chaudron et al., 2010; Dozois et al., 1998; Smarr & Keefer, 2011) A review of self report instruments used for PPD found the BDI II to have excellent internal consis tency, excellent specificity, acceptable sensitivity, excellent positive predictive value, and good concurrent validity (Boyd et al., 2005) Statistical a nalyses Considering the factor structures that have been previously published using the BDI II as a measure of DPP SEM will be employed to test current research questions The first aim was tested through a confirmatory factor analysis on the sample consisting of one hundred mothers at four months postpartum. The second aim was tested through a propos ed path analysis that identified whether latent factors, as mea sured by the twenty one self report questions of the BDI II, could support a conceptual framework that symptoms related to cognitive appraisal influence the development of DPP as measured by the BDI II and SCID Both confirmatory factor analysis and path analysis are statistical tests that operate under the SEM umbrella term used to describe procedures that combine components of factor analysis and multiple regression (Tabachnick & Fidell, 2012) The strength of SEM is seen as its unique ability to test proposed theoretical models simultaneously based on observed or latent factors. Although the

PAGE 23

14 proposed model could be tested in one SEM, a distinct first ste p is being taken to test a three factor solution to item responses on the BDI II through a confirmatory factor analysis, before a path analysis of the latent factors affecting one another and their correlation with a clinical diagnosis of depression, is a pplied. Power Statistical power refers to the probability of detecting an effect if a real effect exists Given the limited sample size available in the retrospective dataset, concern was warranted. However, the planned statistical procedures used for data analysis are considered fairly robust and SEM is preferable for testing smaller sample sizes due to the tests ability to test multiple models simultaneously, as opposed to running multiple indiv idual regression analyses and increasing the probability of a committing a Type I error. Despite the relatively small sample size, the proposed analysis was consistent with current literature establishing a precedent for factor analysis to be used with samples ranging from one hundred to two hundred if the model bei ng tested consists of well determined factors (MacCallum, Widaman, Zhang, & Hong, 1999) Data cleaning & descriptive s tatistics The full dataset of one hundred eight mothers at four months postpartum with any information collected on the BDI II or SCID was entered into IBM SPSS statistical software (IBM Corp., 2012) Assumptions for CFA and path analysis were assessed. Data with missing values were identifie d and a value of "999" was inserted to clearly mark these cases. Descriptive statistics (i.e. variable specific mean, median, mode, standard deviation, and frequency distribution charts) were then performed in SPSS to make sure that there were not any appa rent data entry errors in the dataset before continuing with the analyses. Normality and linearity were assessed through the examination of frequency

PAGE 24

15 histograms computed for each observed variable Particular attention was paid to indices of skew and kurto sis, and an e xamination of the covariance matrix for the observed variables was performed to assess for multicolinarity and singularity. A im 1: Confirmatory factor a nalysis CFA was first introduced in 1969 (Jšreskog, 1969) and is widely used in the social sciences as an acceptable statistical method for researchers to test applied research questions (Brown, 2006) To test the first aim of the proposed study, a CFA was conducted using MPlus statistical software (MuthŽn & MuthŽn, 2011) to test the hypothesis that a three factor solution is acceptable for a sample of one hundred mothers at four months postpartum. The hypothesized three factor solution tested on the BDI II is shown in Figure 1. Accept able model fit was determined using several goodness of fit indices, including the chi square statistic ( x 2 ), comparative fit index (CFI), and root mean square error of approximation (RMSE A ). The x 2 tests the hypotheses that the variables are unrelated; wi thin the context of a path analysis a non significant value is desired (Tabachnick & Fidell, 2012) The CFI provides an estimate of how well the data fit the model, relative to other models. The CFI is well equipped to handle relatively small sample sizes and it is standardized on a scale of 0 to 1, with values greater than 0.95 indicative of a good model fit (Bentler, 1988; Hu & Bentler, 1999) The RMSE compares the hypothesized model to a saturated, or perfect, model for the available data. This estimation is made to analyze the lack of appropriate model fit. When interpreting the RMSE, values less than 0.06 are considered indicative of a good model fit (Hu & Bentler, 1999) whereas values greater than 0.1 indicate poor f itting models (Browne & Cudeck, 1992) However, interpretation of the RMSE A value may be challenging with small sample s given this test 's tendency to over reject the true model (Hu & Bentler, 1999)

PAGE 25

16 Figure 1 Hypothesized Confirmatory Factor Analysis: BDI II 3 Factor Model Aim 2: Path a nalysis Path analysis is a statistical procedure that has been in use since its development in the 1920 's (Wright, 1923, 1934) and has been widely applied to the social sciences (S tage, Carter, & Nora, 2004) To test the second aim, a path analysis was conducted with Mplus statistical software (MuthÂŽn & MuthÂŽn, 2011) to test the proposed theor etical model that emphasizes cognitive appraisal 's influence on responses to the BDI II and SCID in relation to factors of somatic exhaustion and impairment. The hypothesized three factor s olution tested on the BDI II and correlated with SCID diagnosis of DPP is shown in Figure 2. Acceptable model fit was determined using a chi square statistic ( x 2 ), comparative fit index (CFI), and root mean square error of approximation (RMSE A ). Additional ly, the correlation between each latent factor and SCID diagnosis was calculated using a Pearson Product Moment Correlation Coefficient (r). The correlation

PAGE 26

17 coefficient measures the degree to which a linear relationship exists between two variables, and ca n range in value from 1 to +1 (Tabachnick & Fidell, 2012) Values closer to +1 indicat e a positive correlation, where as values approaching 0 indicate no correlation, and values closer to 1 indicate a negative correlation. Figure 2 Hypothesized Path Analysis: BDI II & SCID

PAGE 27

18 CHAPTER III RESULTS A im 1: Confirmatory factor a nalysis To test the first aim a confirmatory facto r analysis (CFA) was conducted on the twenty one BDI II item responses for 100 mothers at four months postpartum. Item responses were normally distributed and a CFA was conducted in MPlus 7.0 A correlation table with means and standard deviations is shown in Table 2 ; the theoretical model is presented in Figure 1. The hypothesized three factor model (i.e. cognitive appraisal, somatic exhaustion, and impairment) is confirmed in the measurement portion of the model. Assumptions of multivariate normality and linerarity were evaluated through SPSS 21.0 using box plots and Mahalnobis distance. No univariate or multivariate outliers were identified; however item 9 on the BDI II (i.e. Suicidal Thoughts or Wishes) only received responses of a zero (i.e. I don't hav e any thoughts of killing myself) or a one (i.e. I have thoughts of killing myself, but I would not carry them out) out of the four possible response options, and thus was declared as a categorical variable during Mplus analysis The nonsignificance of the chi square (x 2 (186 ) = 211.386; ns ) shows that the covariance matrix and mean vector in the population are equal to the model implied covariance matrix and mean vector. The CFI value (CFI=0. 947 ) indicates that the proposed model fits better than an independe nt model in which the variables are not related. The low RMSE value (RMSE=0. 037 ) also indicates th at the model fit the data well. Additionally, all factor loadings were significant (p < .000). Those values indicate a good fit between the model and observed data. Standardized paramerter estimates are provided in Figure 3; unstandardized estimates are shown in Table 3. No

PAGE 28

19 post hoc modifications were indicated from the analysis b ecause of the good fit indexes, and the residual analysis did not indicate any pro blems.

PAGE 29

20 Table 2 Correlations for CFA Observed Variable 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 1. Sadness .254 2. Pessimism .305 .347 3. Past Fai lure .401 .438 .670 4. Loss of Pleasure .528 .306 .483 .422 5. Guilty Feelings .606 .333 .479 .507 .592 6. Punishment Feelings .292 .247 .496 .395 .421 .372 7. Self Dislike .535 .510 .552 .519 .557 .562 .646 8. Self Criticalness .476 .387 .295 .404 .550 .230 .548 .666 9. Suicidal Thoughts .603 .177 .561 .520 .551 .485 .542 .468 10. Crying .404 .374 .378 .313 .274 .224 .438 .294 .283 .793 11. Agitation .200 .321 .225 .461 .391 .303 .359 .456 .237 .284 .644 12. Loss of Interest .485 .272 .347 .614 .445 .365 .421 .365 .607 .352 .301 .405 13. Indeciseveness .288 .173 .167 .243 .373 .131 .283 .405 .357 .104 .247 .385 .502 14. Worthlessness .363 .325 .515 .435 .457 .392 .517 .415 .430 .308 .321 .389 .333 .413 15. Loss of Energy .339 .328 .341 .450 .307 .215 .433 .362 .323 .273 .318 .401 .224 .184 .358 16. Changes in Sleep .364 .249 .200 .486 .289 .179 .229 .239 .198 .315 .367 .376 .279 .136 .334 .712 17. Irritability .378 .435 .267 .491 .349 .331 .437 .394 .309 .255 .519 .410 .265 .3 61 .408 .394 .586 18. Chages in Appetite .231 .089 .215 .245 .243 .010 .270 .277 .342 .186 .286 .343 .315 .155 .380 .190 .245 .606 19. Concentration Difficulty .257 .248 .245 .289 .301 .193 .368 .346 .438 .227 .420 .282 .351 .299 .309 .320 .348 .34 3 .432 20. Tiredness or Fatigue .279 .251 .213 .304 .230 .219 .328 .202 .362 .211 .361 .308 .195 .172 .508 .386 .293 .345 .381 .458 21. Loss of Interest in Sex .262 .229 .044 .244 .306 .188 .181 .238 .221 .010 .233 .265 .576 .269 .154 .361 .149 .121 .206 .252 1.0 N=100; M=0; SD=1

PAGE 30

21 Table 3 Standardized and Unstandardized Coefficients for CFA Observed variable ! SE p valule Cognitive Appraisal 1. Sadness .541 .675 .082 .000 2. Pessimism .496 .529 .077 .000 3. Past Failure .802 .617 .078 .000 4. Loss of Pleasure .757 .733 .057 .000 5. Guilty Feelings .883 .722 .046 .000 6. Punishment Feelings .529 .5 46 .074 .000 7. Self Dislike 1.00 .783 .055 .000 8. Self Criticalness .845 .652 .062 .000 9. Suicidal Thoughts 1.19 .750 .090 .000 10. Crying .697 .493 .078 .000 12. Loss of Interest .695 .687 .067 .000 14. Worthlessness .610 .598 .084 .000 Somatic Exhaustion 11. Agitation 1.37 .642 .066 .000 15. Loss of Energy 1.00 .633 .057 .000 16. Changes in Sleep 1.24 .557 .082 .000 17. Irritiability 1.39 .684 .065 .000 18. Changes in Appetite .967 .467 .077 .000 19. Concentration Difficulties 1.03 .59 5 .080 .000 20. Tiredness or Fatigue 1.00 .556 .085 .000 Impairment 13. Indecisiveness 1.00 .869 .097 .000 21. Loss of Interest in Sex 1.07 .663 .096 .000 = unstandardized coefficient; = standardized coefficient ; SE = standardized standard err or; p value = standardized p value

PAGE 31

22 Figure 3 Confirmatory Factor Analysis: BDI II 3 Factor Model x 2 (186) =211.386, ns ; CFI=0.947; RMSE=0.037. Aim 2: Path a nalysis The second aim used a path analysis designed to demonstrate th e influence that a cognitive appraisal latent factor has on elevated scores on the BDI II, as well as test the correlation between the individual latent factors in the proposed 3 factor model (figure 2 ) and a clinical diagnosis, as indicated by the SCID, o f a major depressive episode in the the postpartum period. For a sample of moth ers at four months postpartum (N =100), the proposed model (figure 2) demonstrated a good fit, as seen in figure 4. The nonsignificance of the chi square ( x 2 (204 ) = 231.917; ns ) sh owed that the covariance matrix and mean vector in the population were equal to the model implied covariance matrix and mean vector. The CFI value (CFI=0.945) indicated that the proposed model fit better

PAGE 32

23 than an independent model in which the variables are not related. The low RMSE value (RMSE=0.037) als o indicated that the model fit the data well. Additionally, all fac tor loading were significant (p < .000). No post hoc modifications were indicated from the analysis due to the good fit indexes, and residua l analysis did not indicate any problems. T he path analysis supported the proposed influence of cognitive appraisal on somatic exhaustion ( = .754) and impairment ( = .481). Moreover, the proposed path analysis tested the correlation between each latent factor and the presence of a SCID diagnosis of major depressive episode in the postpartum period. Cognitive appraisal demonstrated the high est correlation ( r = .663) with SCID diagnosis, compared to either somatic exhaustion ( r = .458) or impairment ( r = .012). Figure 4 Path Analysis: BDI II & SCID x 2 (204) =231.917, ns ; CFI=0.945; RMSE=0.037.

PAGE 33

24 CHAPTER IV DISCUSSION The current study proposed and tested a confirmatory factor analysis and a path analysis model based on item responses on the BDI II and clinical diagnosis of major depressive episode in the postpartum period on the SCID. Both proposed models were formula ted based on literature emphasizing the effects of biopsychosocial stressors on the development of DPP Although the influences of both biological and psych o social stressors have independ en tly been associated with DPP neither have been determined to cause DPP Similarly, the current study is not able to offer any causal association s between biological or psychosocial factors on DPP However, the present study does offer support for the influence of a cognitive appraisal factor on elevation of scores on the BDI II and subsequent diagnosis of a major depressive episode during the postpartum period. The proposed models were specifically tested in a sample of mothers 4 m onths postpar tum for several important reasons. R ealizing that there has been a disconnect between onset timing used in the DSM criteria for PPD and empirical studies of PPD, it was deter mined that a study focused on latent factors associated with DPP would be best served by testing a sample at a specific timepoint. Although it is believed that hormonal levels are quite individual a specific timepoint allows for a standardized amount of time to pass for the entire sample. Furthermore, as the focus of the present study was on the latent factor structure of the BDI II, it was determined that resul ts of the current study may have the most external validity if the sample was within the recommended

PAGE 34

25 guidelines for screening for PPD put forth by the American Accademy of Pediatrics (Earls, 2010) Aim 1 of the current study proposed and tested a CFA of a 3 factor model for the BDI II completed by mothers at 4 m onths postpartum Findings from this study's CFA support the identification of cognitive/appraisal somatic, and impairment related la t ent factors consistently reported in prior literature assessing the BDI II factor s tructure during the postpartum period (Carvalho Bos et al., 2009; Cudmore, 2011; Manian et al., 2013) Testing the proposed CFA, prior to testing the proposed path analysis, was important to determine w h ether the latent factor structure of the BDI II, as measured by the cur rent study' s sample, was representative and consistent with previously published literature on the topic, before testing the proposed theoretical framework emphasizing an influence by the cognitive appraisal latent factor on elevated BDI II scores and SCID diagnosis via a path analysis. Results of the current study's CFA demonstrated good model fit and offers additional support for a 3 factor model representing cognitive appraisal, somatic exhaustion, and impairment latent factors of the BDI II in a postpar tum population. Aim 2 of the current study tested a proposed path analysis suggesting that the cognitive appraisal latent factor influences elevated scores on the BDI II at 4 m onths postpartum. Results of the path analys is offered support to a theoretical model emphasizing the influence of a cognitive appraisal factor on DPP As seen in figure 4, significant path estimates of cognitive appraisal on somatic exhaustion and impairment provides evidence that cognitive appraisal influences item responses on the BDI II, and thus elevated scores, at 4 m onths postpartum for the study's sample Within SEM, the

PAGE 35

26 path estimates between cognitive appraisal and somatic exhaustion and impairment latent factors essentially test a hypothesi zed causal relationship. Even thou gh results of the path analysis supported hypothesized casual relationships between cognitive appraisal and somatic exhaustion and impairment factors, a causal relationship could not be truly measured given the study design and the fact that a single self report measure was used to assess for DPP However, additional support was g enerated for the proposed theoretical model by identifying the correlations between each of the latent factors and clinical diagnosis of a major depressive episode for th e included sample. As shown in F igure 4, cognitive appraisal provided the highest correlation of the 3 factor model o f the BDI II with SCID diagnosis, suggesting that elevated scores on the BDI II items that are most highly correlated with the cognitive appraisal la tent factor are als o most highly correlated with clinical diagnosis of major depressieve episode in the postpartum period. Taken together, aims 1 and 2 of the current study offer important contributions to the understanding of DPP manifestation at 4 m onth s postpartum as measured by the BDI II In addition to supporting previously published literature on the factor structure of the BDI II in postpartum populations, the path anal ysis offers support for a theoretical model that suggests cognitive appraisal in fluences symptoms and potentially the diagnosis of DPP This finding could provide insight into future research studies designed to identify causal pathways for the development of PPD. Additionally, findings from the current study may provide clinicians wi th increased knowledge about the importance of cognitive appraisal related variables measured through self report measures. Although somatic exhaustion related symptoms were endorsed more frequently on the BDI II by mothers at 4 m onths postpartum in the cu rrent sample than cognitive appraisal related

PAGE 36

27 symptoms, the significant path estimates demonstrate the influence that cognitive appraisal may exert on BDI II item responses. Given the pediatric guidelines for assessing symptoms of DPP at infant well child checks throughout the first four months postpartum, clinicians may want to pay particular attention to endorsed symptoms that are related to cognitive appraisal, as they may be more indicative of the manifestation of DPP Despite the promising findings re ported in this study that offer support for the influence of a cognitive appraisal factor on DPP limitations of the current study and statistical methods must be considered. As previously discussed, the path analysis tests a hypothesized causal model, but does not prove causality between the endorsement of cognitive appraisal related symptoms and DPP Furthermore, within the proposed model the directionality of the paths between cognitive appraisal, somatic exhaustion, and impairment latent factors could b e reversed without changing the path estimates. Due to this fact, correlations between the latent factors and a SCID diagnosis were included to offer additional evidence of cognitive appraisal's influence on DPP Additionally the current study's sample si ze was relatively small and the number of mothers with a clinical diagnosis of major depressive episode were even fewer, however the percentage of mothers receiving a clinical diagnosis of major depressive episode during the postpartum period are equiva len t to larger studies identif ying the prevelance of PPD. Although SEM offers advantages for testing hypothesized causal models and may even be considered preferential for analyzing small samples, caution is advised when considering the generalizability of th e findings.

PAGE 37

28 Future research studies should look to include larger samples, additional timepoints and additional measures of depression in an attempt to determine if symptoms related to cognitive appraisal are a primary cause of DPP

PAGE 38

29 REFERENCES American Psychiatric Association. (2000). Diagnostic and statistical manual of mental disorders (Revised 4th ed.) Washington DC author American Psychiatric Association. Retrieved from http://psychiatryonline.com/resourceTOC.aspx?resourceID=1 American Psychiatric Association. (2013). Diagnostic and Statistical Manual of Mental Disorders (Fifth ed.) (5th Editio.). Arlington, VA: American Psychiatric Publishing, Inc. Beck, A. T., Ste er, R. A., & Brown, G. (1996). Manual for the Beck Depression Inventory II San Antonio, TX: Psychological Corporation. Beck, C. T. (2002). Theoretical perspectives of postpartum depression and their treatment implications. The American Journal of Maternal Child Nursing 27 (5), 282 7. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/12209059 Beck, C. T. (2006). Postpartum depression: it isn't just the blues. The American Journal of Nursing 106 (5), 40 50; quiz 50 1. Retrieved from http://www.ncbi.nlm.nih.g ov/pubmed/16639243 Beebe, B., Jaffe, J., Buck, K., Chen, H., Cohen, P., Blatt, S., Andrews, H. (2007). Six week postpartum maternal self criticism and dependency and 4 m onth mother infant self and interactive contingencies. Developmental Psychology 43 ( 6), 1360 76. doi:10.1037/0012 1649.43.6.1360 Bentler, P. M. (1988). Comparative fit indexes in structural models. Psychological Bulletin (107), 238 246. Berggren Clive, K. (1998). Out of the darkness and into the light: women's experiences with depression after childbirth. Canadian Journal of Community Mental Health = Revue Canadienne de SantÂŽ Mentale Communautaire 17 (1), 103 20. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/10351169 Bloch, M., Daly, R. C., & Rubinow, D. R. (2003). Endocrine factors i n the etiology of postpartum depression. Comprehensive Psychiatry 44 (3), 234 46. doi:10.1016/S0010 440X(03)00034 8 Bloch, M., Rubinow, D. R., Schmidt, P. J., Lotsikas, A., Chrousos, G. P., & Cizza, G. (2005). Cortisol response to ovine corticotropin relea sing hormone in a model of pregnancy and parturition in euthymic women with and without a history of postpartum depression. The Journal of Clinical Endocrinology and Metabolism 90 (2), 695 9. doi:10.1210/jc.2004 1388

PAGE 39

30 Bloch, M., Schmidt, P. J., Danaceau, M. Murphy, J., Nieman, L., & Rubinow, D. R. (2000). Effects of gonadal steroids in women with a history of postpartum depression. The American Journal of Psychiatry 157 (6), 924 30. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/10831472 Blom, E. A., Jan sen, P. W., Verhulst, F. C., Hofman, A., Raat, H., Jaddoe, V. W. V, Tiemeier, H. (2010). Perinatal complications increase the risk of postpartum depression. The Generation R Study. BJOG : An International Journal of Obstetrics and Gynaecology 117 (11), 1 390 8. doi:10.1111/j.1471 0528.2010.02660.x Boyd, R. C., Le, H. N., & Somberg, R. (2005). Review of screening instruments for postpartum depression. Archives of Women's Mental Health 8 (3), 141 53. doi:10.1007/s00737 005 0096 6 Brown, T. A. (2006). Confirm atory Factor Analysis for Applied Research (p. 475). New York: Guilford Publications. Retrieved from http://books.google.com/books?id=QTasbvQCZQcC Browne, M. W., & Cudeck, R. (1992). Alternative ways of assessing model fit. Sociological Methods & Research (21), 230 258. Byrnes, E. M., Byrnes, J. J., & Bridges, R. S. (2001). Increased sensitivity of dopamine systems following reproductive experience in rats. Pharmacology, Biochemistry, and Behavior 68 (3), 481 9. Retrieved from http://www.ncbi.nlm.nih.gov/p ubmed/11325402 Carvalho Bos, S., Pereira, A. T., Marques, M., Maia, B., Soares, M. J., Valente, J., Azevedo, M. H. (2009). The BDI II factor structure in pregnancy and postpartum: Two or three factors? European Psychiatry : The Journal of the Association of European Psychiatrists 24 (5), 334 40. doi:10.1016/j.eurpsy.2008.10.003 Centers for Disease Control and Prevention, C. (2008). Prevalence of Self Reported Postpartum Depressive Symptoms: 17 States, 20 04 2005. MMWR. Morbidity and Mortality Weekly Reports 57 (14), 361 366. Retrieved from http://www.cdc.gov/mmWR/preview/mmwrhtml/mm5714a1.htm Chaudron, L. H. (2003). Postpartum depression: what pediatricians need to know. Pediatrics in Review / American Aca demy of Pediatrics 24 (5), 154 61. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/12728188 Chaudron, L. H., Szilagyi, P. G., Tang, W., Anson, E., Talbot, N. L., Wadkins, H. I. M., Wisner, K. L. (2010). Accuracy of depression screening tools for identi fying postpartum depression among urban mothers. Pediatrics 125 (3), e609 e617. Retrieved from http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=3030186&tool=pmcentr ez&rendertype=abstract

PAGE 40

31 Cicchetti, D., Rogosch, F. a, & Toth, S. L. (2000). The effi cacy of toddler parent psychotherapy for fostering cognitive development in offspring of depressed mothers. Journal of Abnormal Child Psychology 28 (2), 135 48. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/10834766 Cudmore, Z. (2011). The Beck Depress ion Inventory II Factor Structure in Postpartum and Implications for Maternal Sensitivity (Honor's Thesis) University of Colorado Denver. Cutrona, C. E., & Troutman, B. R. (1986). Social support, infant temperament, and parenting self efficacy: a mediatio nal model of postpartum depression. Child Development 57 (6), 1507 18. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/3802975 Dennis, C. L., & Dowswell, T. (2013). Psychosocial and psychological interventions for preventing postpartum depression. Cochra ne Database of Systematic Reviews (Online) 2 CD001134. doi:10.1002/14651858.CD001134.pub3 Dietz, L. J., Jennings, K. D., Kelley, S. A., & Marshal, M. (2009). Maternal depression, paternal psychopathology, and toddlers' behavior problems. Journal of Clini cal Child and Adolescen t Psychology : The Official Journal for the Society of Clinical Child and Adolescent Psychology, American Psychological Association, Division 53 38 (1), 48 61. doi:10.1080/15374410802575362 Dozois, D. J. A., Dobson, K. S., & Ahnberg, J. L. (1998). A psycho metric evaluation of the Beck Depression Inventory II. Psychological Assessment 10 (2), 83 89. doi:10.1037/1040 3590.10.2.83 Earls, M. F. (2010). Incorporating recognition and management of perinatal and postpartum depression into pediatric practice. Pedia trics 126 (5), 1032 1039. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/20974776 Engel, G. L. (1977). The need for a new medical model: a challenge for biomedicine. Science (New York, N.Y.) 196 (4286), 129 36. Retrieved from http://www.ncbi.nlm.nih.gov /pubmed/847460 Fava, G. A., & Sonino, N. (2008). The biopsychosocial model thirty years later. Psychotherapy and Psychosomatics 77 (1), 1 2. doi:10.1159/000110052 Field, T. (2010). Postpartum depression effects on early interactions, parenting, and safety practices: a review. Infant Behavior & Development 33 (1), 1 6. doi:10.1016/j.infbeh.2009.10.005 First, M., Spitzer, R. L., Gibbon, M., & Williams, J. B. (2002). Structured Clinical Interview for DSM IV TR Axis I Disorders, Research Version, Patient Editio n. (SCID I/P) New York: Biometrics Research, New York State Psychiatric Institute.

PAGE 41

32 Flykt, M., Kanninen, K., Sinkkonen, J., & Punamaki, R. (2010). Maternal depression and dyadic interaction: the role of maternal attachment style. Infant Child Development (19), 530 550. Gavin, N., & Gaynes, B. (2005). Perinatal depression: a systematic review of prevalence and incidence. Obstetrics & 106 (5 Pt 1), 1071 83. doi:10.1097/01.AOG.0000183597.31630.db Grothe, K. B., Dutton, G. R., Jones, G. N., Bodenlos, J., Anc ona, M., & Brantley, P. J. (2005). Validation of the Beck Depression Inventory II in a low income African American sample of medical outpatients. Psychological Assessment 17 (1), 110 4. doi:10.1037/1040 3590.17.1.110 Gu, H., Hu, D., Hong, X., Mao, J., Cui, Y., Hui, N., & Sha, J. (2003). [Changes and significance of orphanin and serotonin in patients with postpartum depression]. Zhonghua Fu Chan Ke Za Zhi 38 (12), 727 8. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/14728841 Hay, D. F. (1997). Postpartum depression and cognitive development. In L. Murray & P. J. Cooper (Eds.), Postpartum depression and child development (pp. 85 110). New York: The Guilford Press. Howell, E. A., Mora, P. A., Horowitz, C. R., & Leventhal, H. (2005). Racial and ethnic differ ences in factors associated with early postpartum depressive symptoms. Obstetrics and Gynecology 105 (6), 1442 50. doi:10.1097/01.AOG.0000164050.34126.37 Hu, L. T., & Bentler, P. M. (1999). Cutoff criteria for fit indices in covariance structure analysis: conventional criteria versus new alternatives. Structural Equation Modeling 6 (1), 1 55. IBM Corp. (2012). IBM SPSS Statistics for Mac, Version 21.0. Armonk, NY. Jones, H. W., & Venis, J. A. (2001). Identification and classification of postpartum psychiatr ic disorders. Journal of Psychosocial Nursing and Mental Health Services 39 (12), 23 30. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/11775292 Jšreskog, K. G. (1969). A general approach to confirmatory maximum likelihood factor analysis. Psychometrika 34 (2), 183 202. doi:10.1007/BF02289343 Josefsson, A., & Sydsjš, G. (2007). A follow up study of postpartum depressed women: recurrent maternal depressive symptoms and child behavior after four years. Archives of Women's Mental Health 10 (4), 141 5. doi:1 0.1007/s00737 007 0185 9

PAGE 42

33 Kaplan, P. S., Danko, C. M., Diaz, A., & Kalinka, C. J. (2011). An associative learning deficit in 1 year old infants of depressed mothers: role of depression duration. Infant Behavior & Development 34 (1), 35 44. doi:10.1016/j.inf beh.2010.07.014 Kaplan, P. S., Danko, C. M., Kalinka, C. J., & Cejka, A. M. (2012). A developmental decline in the learning promoting effects of infant directed speech for infants of mothers with chronically elevated symptoms of depression. Infant Behavior & Development 35 (3), 369 379. doi:10.1016/j.infbeh.2012.02.009 Kneipp, S. M., Kairalla, J. A., Stacciarini, J., & Pereira, D. (2009). The Beck Depression Inventory II factor structure among low income women. Nursing Research 58 (6), 400 9. doi:10.1097/NN R.0b013e3181bee5aa Lawson, K. R., & Ruff, H. A. (2004). Early attention and negative emotionality predict later cognitive and behavioural function. International Journal of Behavioral Development 28 (2 ), 157 165. doi:10.1080/01650250344000361 Lobbestael J., Leurgans, M., & Arntz, A. (2010). Inter rater reliability of the Structured Clinical Interview for DSM IV Axis I Disorders (SCID I) and Axis II Disorders (SCID II). Clinical Psychology & Psychotherapy 18 (1), 75 9. doi:10.1002/cpp.693 Lundy, B. L., F ield, T., Cuadra, A., Nearing, G., Cigales, M., & Hashimoto, M. (1996). Mothers with Depressive Symptoms Touching Newborns. Early Dev. Parent (5), 129 134. MacCallum, R. C., Widaman, K. F., Zhang, S., & Hong, S. (1999). Sample size in factor analysis. Psy chological Methods (4), 84 99. Manian, N., Schmidt, E., Bornstein, M. H., & Martinez, P. (2013). Factor structure and clinical utility of BDI II factor scores in postpartum women. Journal of Affective Disorders 149 (1 3), 259 68. doi:10.1016/j.jad.2013.01 .039 Mann, J. J. (1999). Role of the serotonergic system in the pathogenesis of major depression and suicidal behavior. Neuropsychopharmacology : Official Publication of the American College of Neuropsychopharmacology 21 (2 Suppl), 99S 105S. doi:10.1016/S0 893 133X(99)00040 8 Milgrom, J., Westley, D. T., & Gemmill, A. W. (2004). The mediating role of maternal responsiveness in some longer term effects of postnatal depression on infant development. Infant Behavior and Development 27 (4), 443 454. doi:10.1016/ j.infbeh.2004.03.003 Murray, L. (1992). The impact of postnatal depression on infant development. Journal of Child Psychology and Psychiatry, and Allied Disciplines 33 (3), 543 61. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/1577898

PAGE 43

34 Murray, L., Coope r, P. J., Wilson, A., & Romaniuk, H. (2003). Controlled trial of the short and long term effect of psychological treatment of post partum depression: 2. Impact on the mother child relationship and child outcome. The British Journal of Psychiatry : The Jou rnal of Mental Science 182 420 7. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/12724245 MuthÂŽn, L. K., & MuthÂŽn, B. O. (2011). Mplus User's Guide (6th Edition) (Sixth Edit.). Los Angeles, CA: MuthÂŽn & MuthÂŽn. NICHD Early Child Care Research Network. (1999). Chronicity of maternal depressive symptoms, maternal sensitivity, and child functioning at 36 months. NICHD Early Child Care Research Network. Developmental Psychology 35 (5), 1297 310. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/10493655 Ny len, K. J., O'Hara, M. W., Brock, R., Moel, J., Gorman, L., & Stuart, S. (2010). Predictors of the longitudinal course of postpartum depression following interpersonal psychotherapy. Journal of Consulting and Clinical Psychology 78 (5), 757 63. doi:10.1037 /a0020623 O'Hara, M. W., & McCabe, J. E. (2013). Postpartum depression: current status and future directions. Annual Review of Clinical Psychology 9 379 407. doi:10.1146/annurev clinpsy 050212 185612 O'Hara, M. W., Neunaber, D. J., & Zekoski, E. M. (1984 ). Prospective study of postpartum depression: prevalence, course, and predictive factors. Journal of Abnormal Psychology 93 (2), 158 71. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/6725749 O'Hara, M. W., Rehm, L. P., & Campbell, S. B. (1982). Predic ting depressive symptomatology: cognitive behavioral models and postpartum depression. Journal of Abnormal Psychology 91 (6), 457 61. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/7153419 O'Hara, M. W., Schlechte, J. A., Lewis, D. A., & Wright, E. J. ( 1991a). Prospective study of postpartum blues. Biologic and psychosocial factors. Archives of General Psychiatry 48 (9), 801 6. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/1929770 O'Hara, M. W., Schlechte, J. A., Lewis, D. A., & Wright, E. J. (1991b) Prospective study of postpartum blues. Biologic and psychosocial factors. Archives of General Psychiatry 48 (9), 801 6. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/1929770 O'Hara, M. W., Zekoski, E. M., Philipps, L. H., & Wright, E. J. (1990). Cont rolled prospective study of postpartum mood disorders: comparison of childbearing and

PAGE 44

35 nonchildbearing women. Journal of Abnormal Psychology 99 (1), 3 15. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/2307763 Rich Edwards, J. W., Kleinman, K., Abrams, A ., Harlow, B. L., McLaughlin, T. J., Joffe, H., & Gillman, M. W. (2006). Sociodemographic predictors of antenatal and postpartum depressive symptoms among women in a medical group practice. Journal of Epidemiology and Community Health 60 (3), 221 7. doi:10 .1136/jech.2005.039370 Salmela Aro, K., Nurmi, J. E., Saisto, T., & Halmesmaki, E. (2001). Goal reconstruction and depressive symptoms during the transition to motherhood: evidence from two cross lagged longitudinal studies. Journal of Personality and Soci al Psychology 81 (6), 1144 59. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/11761314 Segre, L. S., O'Hara, M. W., Arndt, S., & Stuart, S. (2007). The prevalence of postpartum depression: the relative significance of three social status indices. Social Psychiatry and Psychiatric Epidemiology 42 (4), 316 21. doi:10.1007/s00127 007 0168 1 Sharma, V., & Sharma, P. (2012). Postpartum depression: diagnostic and treatment issues. Journal of Obstetrics and Gynaecology Canada : JOGC = Journal D'obstÂŽtrique et G ynÂŽcologie Du Canada : JOGC 34 (5), 436 42. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/22555136 Shear, M. K., Greeno, C., Kang, J., Ludewig, D., Frank, E., Swartz, H. A., & Hanekamp, M. (2000). Diagnosis of nonpsychotic patients in community clinics The American Journal of Psychiatry 157 (4), 581 7. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/10739417 Siegert, R. J., Walkey, F. H., & Turner Stokes, L. (2009). An examination of the factor structure of the Beck Depression Inventory II in a neuro rehabilitation inpatient sample. Journal of the International Neuropsychological Society : JINS 15 (1), 142 7. doi:10.1017/S1355617708090048 Skre, I., Onstad, S., Torgersen, S., & Kringlen, E. (1991). High interrater reliability for the Structured Clinical Interview for DSM III R Axis I ( SCID I). Acta Psychiatrica Scandinavica 84 (2), 167 73. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/1950612 Smarr, K. L., & Keefer, A. L. (2011). Measures of depression and depressive symptoms: Beck Depression Inventory II (BDI II), Center for Epidem iologic Studies Depression Scale (CES D), Geriatric Depression Scale (GDS), Hospital Anxiety and Depression Scale (HADS), and Patient Health Questionna. Arthritis Care & Research 63 Suppl 1 S454 66. doi:10.1002/acr.20556

PAGE 45

36 Spinelli, M. G. (2004). Maternal infanticide associated with mental illness: prevention and the promise of saved lives. The American Journal of Psychiatry 161 (9), 1548 57. doi:10.1176/appi.ajp.161.9.1548 Stage, F. K., Carter, H. C., & Nora, A. (2004). Path Analysis: An Introduction and A nalysis of a Decade of Research. The Journal of Educational Research 98 (1), 5 13. doi:10.3200/JOER.98.1.5 13 Stanley, C., Murray, L., & Stein, A. (2004). The effect of postnatal depression on mother infant interaction, infant response to the Still face pe rturbation, and performance on an Instrumental Learning task. Development and Psychopathology 16 (1), 1 18. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/15115062 Stein, A., Arteche, A., Lehtonen, A., Craske, M., Harvey, A., Counsell, N., & Murray, L. (2010). Interpretation of infant facial expression in the context of maternal postnatal depression. Infant Behavior & Development 33 (3), 273 8. doi:10.1016/j.infbeh.2010.03.002 Steiner, J. L., Tebes, J. K., Sledge, W. H., & Walker, M. L. (1995). A compari son of the structured clinical interview for DSM III R and clinical diagnoses. The Journal of Nervous and Mental Disease 183 (6), 365 9. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/7798084 Steiner, M. (1998). Perinatal mood disorders: position paper. Psychopharmacology Bulletin 34 (3), 301 6. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/9803759 Storch, E. A., Roberti, J. W., & Roth, D. A. (2004). Factor structure, concurrent validity, and internal consistency of the Beck Depression Inventory Second Edition in a sample of college students. Depression and Anxiety 19 (3), 187 9. doi:10.1002/da.20002 Tabachnick, B. G., & Fidell, L. S. (2012). Using Multivariate Statistics Pearson Education, Limited. Retrieved from http://books.google.com/books?id=ucj1ygAACAAJ Teti, D. M., & Gelfand, D. M. (1991). Behavioral competence among mothers of infants in the first year: the mediational role of maternal self efficacy. Child Development 62 (5), 918 29. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/1756667 Vik, T., Grote, V., Escribano, J., Socha, J., Verduci, E., Fritsch, M., Koletzko, B. (2009). Infantile colic, prolonged crying and maternal postnatal depression. Acta Paediatrica (Oslo, Norway : 1992) 98 (8), 1344 8. doi:10.1111/j.1651 2227.2009.01317.x

PAGE 46

37 Viljoen, J. L., Iverson, G. L., Griffiths, S., & Woodward, T. S. (2003). Factor Structure of the Beck Depression Inventory II in a Medical Outpatient Sample. Journa l of Clinical Psychology in Medical Settings 10 (4), 289 291. Ward, L. C. (2006). Comparison of factor structure models for the Beck Depression Inventory -II. Psychological Assessment 18 (1), 81 8. doi:10.1037/1040 3590.18.1.81 Weissman, M. M., Pilowsky, D J., Wickramaratne, P. J., Talati, A., Wisniewski, S. R., Fava, M., Rush, A. J. (2006). Remissions in maternal depression and child psychopathology: a STAR*D child report. JAMA : The Journal of the American Medical Association 295 (12), 1389 98. doi:10.1001/jama.295.12.1389 Williams, J. B., Gibbon, M., First, M. B., Spitzer, R. L., Davies, M., Borus, J., Rounsaville, B. (1992). The Structured Clinical Interview for DSM III R (SCID). II. Multisite test retest reliability. Archives of General Psychiatry 49 (8), 630 6. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/1637253 Workman, J. L., Barha, C. K., & Galea, L. A. M. (2012). Endocrine substrates of cognitive and affectiv e changes during pregnancy and postpartum. Behavioral Neuroscience 126 (1), 54 72. doi:10.1037/a0025538 Wright, S. (1923). The Theory of Path Coefficients a Reply to Niles's Criticism. Genetics 8 (3), 239 55. Retrieved from http://www.pubmedcentral.nih.gov /articlerender.fcgi?artid=1200747&tool=pmcentr ez&rendertype=abstract Wright, S. (1934). The method of path coefficients. Ann. Math. Statist. (5), 161 215. Zanarini, M. C., & Frankenburg, F. R. (2001). Attainment and maintenance of reliability of axis I an d II disorders over the course of a longitudinal study. Comprehensive Psychiatry 42 (5), 369 74. doi:10.1053/comp.2001.24556 Zanarini, M. C., Skodol, A. E., Bender, D., Dolan, R., Sanislow, C., Schaefer, E., Gunderson, J. G. (2000). The Collaborative Lon gitudinal Personality Disorders Study: reliability of axis I and II diagnoses. Journal of Personality Disorders 14 (4), 291 9. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/11213787