Predictors of Treatment Response in Depressed Mothers Receiving In-Home Cognitive-Behavioral Therapy and Concurrent Home Visiting
Home visiting is a child abuse prevention strategy that seeks to optimize child development by providing mothers with support, training, and parenting information. Research has consistently found high rates of depression in mothers participating in home visiting programs and low levels of obtaining mental health treatment in the community. Successful treatment of depressed mothers in home visiting programs holds the potential to improve maternal and child outcomes. In-Home Cognitive-Behavioral Therapy (IH-CBT) is an adapted treatment for depressed mothers, which is provided alongside home visiting and seeks to optimize engagement and impact through delivery in the home setting; a focus on issues important to young, low-income mothers; and a strong collaborative relationship between therapists and home visitors. This study examined predictors of depression status at posttreatment in 60 mothers who received IH-CBT and concurrent home visiting. Variables considered included demographics, illness history, severity, and numbers of treatment sessions and home visits. Results indicated that young maternal age, fewer episodes of major depressive disorder, lower depression severity at pretreatment, lower levels of symptoms of personality disorders, and more treatment sessions and home visits predicted asymptomatic status at posttreatment.
36(4) 462 –481
© The Author(s) 2012
Reprints and permission:
20Ammerman et al.Behavior Modification
1Cincinnati Children’s Hospital Medical Center and University of Cincinnati College of
Medicine, OH, USA
Robert T. Ammerman, PhD, Cincinnati Children’s Hospital Medical Center,
3333 Burnet Avenue, Cincinnati, OH 45229, USA
Predictors of Treatment
Response in Depressed
Mothers Receiving In-
Therapy and Concurrent
Ammerman1, James L. Peugh1, Frank
W. Putnam1, and Judith B.
Home visiting is a child abuse prevention strategy that seeks to optimize
child development by providing mothers with support, training, and parent-
ing information. Research has consistently found high rates of depression in
mothers participating in home visiting programs and low levels of obtaining
mental health treatment in the community. Successful treatment of depressed
mothers in home visiting programs holds the potential to improve maternal
and child outcomes. In-Home Cognitive-Behavioral Therapy (IH-CBT) is an
adapted treatment for depressed mothers, which is provided alongside home
visiting and seeks to optimize engagement and impact through delivery in the
home setting; a focus on issues important to young, low-income mothers;
and a strong collaborative relationship between therapists and home visi-
tors. This study examined predictors of depression status at posttreatment
in 60 mothers who received IH-CBT and concurrent home visiting. Variables
considered included demographics, illness history, severity, and numbers of
treatment sessions and home visits. Results indicated that young maternal
Ammerman et al. 463
age, fewer episodes of major depressive disorder, lower depression sever-
ity at pretreatment, lower levels of symptoms of personality disorders, and
more treatment sessions and home visits predicted asymptomatic status at
home visiting, maternal depression, cognitive-behavioral therapy, adapted
Home visiting is a child abuse prevention strategy that targets new mothers
and their children who are at demographic risk for adverse parenting out-
comes. Such programs typically seek to enroll mothers prenatally or during
the first months after the child’s birth, provide a trained visitor who provides
services in the home setting, deliver a developmentally aligned curriculum
that focuses on positive child development and meeting maternal and child
needs, and link families to other resources in the community (Boller &
Strong, 2010). The overarching objective of home visiting is to optimize
child development and maternal life course by providing intensive services
and support during the critical early years of the child’s life.
By seeing mothers in the home, a frequent barrier to services is overcome.
As a result, home visiting programs are an ideal setting in which to identify
and intervene with depressed mothers who might otherwise fail to receive
adequate care (National Research Council and Institute of Medicine, 2009).
Depression is commonly reported in mothers enrolled in home visiting pro-
grams. Indeed, most mothers in home visiting programs display characteris-
tics associated with elevated risk for depression, including being unmarried,
socially isolated, young, living in poverty, and victims of violence (Segre,
O’Hara, Arndt, & Stuart, 2007). For example, Ammerman et al. (2009) found
that 45.3% of mothers participating in a regional home visiting program had
elevated levels of depressive symptoms using the Beck Depression Inventory–
II (BDI-II) at two time points in the 1st year of service. This study also reported
that 74.1% of mothers experienced childhood trauma and violence prior to
enrollment. Depression in mothers in home visiting is associated with greater
impairments in parenting, increased parenting stress, and social isolation rela-
tive to nondepressed mothers (Ammerman, Shenk, et al., 2011). Research has
consistently demonstrated that (a) depressed mothers in home visiting rarely
obtain mental health treatment in the community and (b) home visiting by
itself yields little benefit to maternal depressive symptoms over the course of
464 Behavior Modiﬁcation 36(4)
services (Ammerman, Putnam, Bosse, Teeters, & Van Ginkel, 2010). Because
maternal depression can mitigate the benefits of home visiting and prolonged
exposure to depressed mothers can contribute to poor outcomes in children
(Hay, Pawlby, Waters, Perra, & Sharp, 2010), it is important to provide effec-
tive treatment to this population.
In response to this need, Ammerman, Bodley, et al. (2007) systematically
adapted cognitive-behavioral therapy (CBT) for use with depressed mothers
receiving home visitation. Implemented by trained therapists and delivered
concurrently with home visiting, In-Home Cognitive-Behavioral Therapy
(IH-CBT) combines the principles and strategies of CBT (J. S. Beck, 2011)
with a set of procedures and approaches that promotes engagement, makes
content relevant to the needs of young mothers, permits implementation in the
home setting, and forges a strong collaborative relationship between the thera-
pist and home visitor to smoothly coordinate services and optimize outcomes.
Adaptations include identifying suitable areas for conducting therapy in
homes that are often small, crowded, and unkempt; addressing the unique
issues of socially isolated, low-income mothers who are raising a new baby;
and creating close working relationships between home visitors and thera-
pists. IH-CBT is a focused treatment that emphasizes the reduction of mater-
nal depressive symptoms and recovery from major depressive disorder
(MDD), thereby allowing home visitors to attend to issues related to parent-
ing, maternal functioning, and child development. IH-CBT is delivered in 15
weekly sessions and a 1-month booster that are in addition to visits provided
by home visitors. There is regular contact between the therapist and home visi-
tor over the course of treatment, and home visitors attend the 15th session to
review treatment impacts and facilitate home visitor support of mothers fol-
lowing termination (Ammerman, Putnam, Stevens, et al., 2011).
Preliminary evidence suggests that IH-CBT is efficacious in the treatment
of depressed mothers in home visiting. Ammerman, Putnam, Stevens, et al.
(2011) provided IH-CBT to 64 mothers aged 18 years and older who were
participating in a home visiting program, obtained scores of ≥20 on the
BDI-II, and met criteria for MDD using the Primary Care Evaluation of
Mental Disorders (PRIME-MD; Spitzer et al., 1994). Pre–post comparisons
revealed that 46.9% of mothers were partially remitted in terms of MDD
diagnosis and 32.8% were fully remitted at the end of treatment. Mothers also
reported decreased depressive symptoms, and improvements in social sup-
port, relationships, coping, and satisfaction with motherhood. A subsequent
comparison of the 64 mothers to 241 mothers who had similarly elevated
BDI-II scores at enrollment in home visiting showed a large reduction (M
difference = 7.8 points, p < .01) in the treatment group relative to untreated
Ammerman et al. 465
mothers. Group comparisons were not moderated by maternal age, race, edu-
cation, or home visitation model, suggesting that the adaptations made in
IH-CBT were successful in impacting the diverse populations served by
home visiting. Yet, it is not known what characteristics or other variables are
most predictive of improvement of maternal depression in those receiving
Identifying predictors of outcome among individuals who receive treatment
is a useful way to determine who is most likely to benefit from a specific inter-
vention. Such prognostic information can be used to refine treatments to better
address the needs of those who respond less optimally (Driessen & Hollon,
2010), and to determine at the start of treatment to what extent individuals
might benefit based on pretreatment characteristics. Several reviews have sum-
marized research examining predictors of outcomes in CBT for adult depres-
sion (Driessen & Hollon, 2010; Haby, Donnelly, Corry, & Vos, 2006; Hamilton
& Dobson, 2002). These efforts have identified demographic, clinical, and
treatment process variables that are associated with favorable or poor out-
comes. In terms of demographics, unmarried status and older age have been
found to predict poorer response to CBT in depressed adults (e.g., Fournier
et al., 2009). There is also a consistent and robust relationship between illness
severity and outcomes. Chronic depression, pretreatment depression severity,
earlier onset of MDD, and comorbid personality disorders have been found to
be predictive of poorer response. Duration of CBT is also important, with better
outcomes associated with more sessions. Such information is unavailable for
IH-CBT and depressed mothers in home visiting.
Delineating pretreatment variables that predict response to IH-CBT is an
important area for research for two reasons. First, depressed mothers in home
visiting display many of the characteristics associated with poorer response,
including unmarried status and significant morbidity (Ammerman, Putnam,
Chard, Stevens, & Van Ginkel, 2011). Second, such information can guide
modifications to IH-CBT that will lead to more effective treatment for this
high-risk population. The unique features of IH-CBT (provided in home set-
ting, integral collaboration between therapists and home visitors) may yield
different predictors than what has been found in studies that rely primarily on
adult clients of varied age, include males, report on treatments provided in
center-based settings, and without regard to parenting status.
The purpose of this study was to determine predictors of depression
outcome in a sample of depressed mothers who received concurrent home
visiting and IH-CBT. Mothers were enrolled in treatment between 2 and
10 months postpartum and assessed at pretreatment and posttreatment.
At posttreatment, mothers were categorized as asymptomatic or still
466 Behavior Modiﬁcation 36(4)
symptomatic based on obtaining a score of ≤8 on the BDI-II (Keller, 2003).
These groups were contrasted on demographic, clinical, and treatment pro-
cess variables. As the theoretical foundation of IH-CBT relies heavily on
the close relationship between therapy and home visiting, number of home
visits was examined to determine its potential contribution to predicting
outcome. It was hypothesized that, relative to mothers who are symptom-
atic at posttreatment, asymptomatic mothers would be younger, have less
severe clinical indicators of depression at pretreatment, and receive more
IH-CBT sessions and home visits.
Participants consisted of 60 new mothers aged 16 or older who participated in
a home visiting program, were diagnosed with MDD, enrolled between 2 and
10 months postpartum, and received IH-CBT as part of a clinical trial
(Ammerman, Putnam, Altaye, Stevens, & Van Ginkel, 2012). Mothers were
enrolled in Every Child Succeeds, a regional home visiting program serving
new mothers in Southwestern Ohio and Northern Kentucky. The geographic
area covered by the program included urban and rural areas. Two national
models of home visitation were utilized: Nurse–Family Partnership (NFP;
Olds, 2010) and Healthy Families America (HFA; Holton & Harding, 2007).
Participating mothers had at least one of four sociodemographic risk character-
istics needed for eligibility: unmarried, low income, age < 18 years, and inad-
equate prenatal care. Mothers were enrolled in home visiting prior to 28 weeks
gestation in NFP as per model parameters and from 20 weeks gestation through
the child reaching 3 months of age for HFA. In the NFP, home visits were
provided by nurses, whereas in HFA, home visits were provided by social
workers, related professionals, and paraprofessionals. The goals of home visi-
tation are to (a) improve pregnancy outcomes through nutrition education and
substance use reduction; (b) support parents in providing children with a safe,
nurturing, and stimulating home environment; (c) optimize child health and
development; (d) link families to health care and other needed services; and (e)
promote economic self-sufficiency (Ammerman, Putnam, et al., 2007).
Table 1 presents the demographic characteristics of the treated group. For
the sample as a whole, mothers were young (M = 22.4, SD = 5.0), predomi-
nantly Caucasian (61.6%) and African American (33.3%), unmarried
(86.7%), and low income (75.0% < US$20,000 annual household income). In
terms of home visitation model, 54 mothers were in the HFA model and 6 in
Ammerman et al. 467
the NFP model. Fifty-three mothers were primiparous and 7 (HFA only) had
more than one child.
Using the Structured Clinical Interview for DSM-IV (SCID) at pretreat-
ment assessment, the breakdown of MDD severity was mild = 20.0%, moder-
ate = 50.0%, and severe = 30.0%. Onset of the presenting episode during
postpartum (within 6 weeks after birth) was found in 26.7% of participants.
Recurrent depression was reported by 78.3% of mothers (M number of epi-
sodes = 2.9 [SD = 1.7], M age of first episode = 14.7 years [SD = 5.7]).
Comorbidity was high, with 76.3% of participants meeting criteria for other
Table 1. Demographic Characteristics of Sample (n = 60)
Variable M (SD) or n (%)
Mother age (years) 22.4 (5.0)
White 37 (61.6%)
African American 20 (33.3%)
Native American 1 (1.7%)
Native Hawaiian or other Pacific Islander 1 (1.7%)
Biracial 1 (1.7%)
Latina 4 (6.7%)
None 56 (93.3%)
Single, never married 52 (86.7%)
Married 8 (13.3%)
Education (years) 11.5 (1.9)
Income (in US$)
0-9,999 33 (55.0%)
10,000-19,999 12 (20.0%)
20,000-29,999 11 (18.3%)
30,000-39,999 3 (5.0%)
40,000-49,999 0 (0.0%)
>50,000 1 (1.7%)
Child’s age (days) 152.0 (73.0)
Male 28 (46.7%)
Female 32 (53.3%)
468 Behavior Modiﬁcation 36(4)
Mothers were recruited in a two-step process as part of a clinical trial of
IH-CBT. First, home visitors administered the Edinburgh Postnatal
Depression Scale (Cox, Holden, & Sagovsky, 1987) to mothers at 3 months
postpartum. Mothers who screened positive were approached to participate
in the trial. Interested mothers then received a pretreatment and eligibility
assessment. Inclusion criteria were age ≥16 years and current diagnosis of
MDD. Exclusion criteria were bipolar disorder, current substance depen-
dence, psychosis, mental retardation, suicidality or homicidality requiring
acute intervention, or current use of psychotropic medications or psycho-
therapy. Data are presented from the pretreatment and posttreatment
assessments. The sample included mothers originally assigned to IH-CBT
(n = 47) and those assigned to the control condition who subsequently
crossed over to IH-CBT (n = 13) at posttreatment based on continued diag-
nosis of MDD. Nine mothers were lost to posttreatment follow-up but were
included in analyses.
Home Visiting and IH-CBT
All mothers received ongoing services from home visitors as per the HFA
and NFP model directives while participating in IH-CBT. Both models call
for regular home visits during the intervals covered during the trial, and
home visitors are given discretion to increase frequency of visits if needed.
Curricula for both models are distinct, but each addresses child health and
development, nurturing mother–child relationship, maternal health and self-
sufficiency, and linkage to other community services.
IH-CBT was delivered in the home by two licensed, master’s level social
workers. Treatment consisted of 15 sessions that were scheduled weekly and
lasted 60 min plus a booster session 1 month posttreatment. Adaptations to
CBT were made to address setting, population, and context. These adapta-
tions were made based on a review of the literature, consultation with home
visitors, and input from mothers in home visitation. First, IH-CBT was deliv-
ered in the home. Creative solutions and accommodations were made to
ensure treatment delivery in home environments where privacy was some-
times difficult to ensure, the child was present, and unexpected interruptions
occurred. However, providing treatment in the home offered advantages in
that many of the clinical issues that were addressed in treatment occurred in
the home setting, and the therapist was able to observe elements of the home
Ammerman et al. 469
that may have been contributory to the clinical presentation. The second
adaptation involved addressing the primary concerns of young, low income,
new mothers who were socially isolated (Levy & O’Hara, 2010). Treatment
content focused on issues relevant to this population, such as transition to
adult roles, stress management, parenting challenges, and family relation-
ships. The third adaptation sought to facilitate close collaboration
(Zwarenstein, Goldman, & Reeves, 2009) with home visitors. Collaboration
occurred through weekly written communication between therapist and home
visitor utilizing a shared web-based documentation system, and/or telephone
contact as needed. In addition, the home visitor attended the 15th session
with the mother and therapist. Weekly supervision was provided by two doc-
toral level clinicians.
SCID Axis I Disorders (SCID-I). The SCID (January 2007 version; Spitzer,
Williams, Gibbon, & First, 1992) is a semistructured psychiatric interview that
is widely used in research and clinical practice. It is used to diagnose 14 com-
mon psychiatric disorders, including MDD. Interrater reliability is generally
high, ranging from 0.57 to 1.00 (Zanarini & Frankenburg, 2001), and validity
is well established (Shear et al., 2000). Interviews were audio-recorded and
25% were rated by a second rater yielding a kappa coefficient = .89. The SCID
was used to determine MDD diagnosis, comorbidities, and clinical features of
MDD, including age of onset and number of episodes.
BDI-II. The BDI-II (A. T. Beck, Steer, & Brown, 1996) is one of the most
widely used self-report screens of depressive symptomatology, with strong
reliability and validity properties. It consists of 21 items in which mothers
indicated presence and severity of depressive symptoms over the past 2
weeks by endorsing one of four statements reflecting degree of severity,
yielding a total score. The BDI-II was administered at pretreatment and post-
treatment. Participants were categorized as asymptomatic (BDI-II ≤ 8) or
symptomatic (BDI-II ≥ 9) at posttreatment (Keller, 2003).
Childhood Trauma Questionnaire (CTQ). The CTQ (Bernstein et al., 2003) is a
28-item version of the larger CTQ. Items describe maltreatment experiences
in childhood and are endorsed on a 5-item Likert-type scale reflecting how
true they are. Scores reflect physical abuse, sexual abuse, emotional abuse,
physical neglect, and emotional neglect. The CTQ has excellent internal reli-
ability (α = .61-.94 in clinical and community samples) and correlates highly
470 Behavior Modiﬁcation 36(4)
with determinations of maltreatment (Scher, Stein, Asmundson, McCreary, &
Forde, 2001). The CTQ was analyzed using the total raw score.
Iowa Personality Disorder Screen (IPDS). The IPDS (Langbehn et al., 1999)
is a brief, 11-item screen for symptoms of personality disorders. It is admin-
istered as an interview, and each item is endorsed as yes/no based on reports
consistent with symptoms suggestive of personality disorders (e.g., “feels
uncomfortable in situations where he or she is not the center of attention”). A
total score is derived reflecting the number of items rated as “yes.” The IPDS
has acceptable psychometric characteristics, including adequate sensitivity
and specificity (Trull & Amdur, 2001).
Overview of Analyses
Based on the number of mothers available for posttreatment assessment,
asymptomatic (n = 33, 64.7%) and symptomatic (n = 18, 35.3%) groups were
formed using the BDI-II ≤ 8 as the criterion. Missing data were handled via
maximum likelihood estimation with a saturated correlates model (Enders,
2010). Handling missing data via maximum likelihood estimation assumes
that these data are missing at random (MAR), meaning the missing data are
related to the variables included in the analysis model. There is no formal test
for MAR, but increasing the number of variables in the analysis model tends
to make the MAR assumption more plausible. A saturated correlates model
allows additional variables to be added to the analysis model without biasing
parameter estimates of interest. Accordingly, selected demographics were
added to the analysis model as additional correlates of missing data.
Asymptomatic and symptomatic groups were contrasted using MANOVA
and Wald Z values with accompanying Cohen’s d effect size estimates. Next,
a hierarchical linear regression was conducted to determine the relative con-
tributions of previously identified determinants of endpoint symptom status
to reduction in BDI-II scores from pre- to posttreatment. All analyses were
conducted using Mplus version 6.12 (Muthen & Muthen, 2010).
Table 2 presents contrasts for the asymptomatic and symptomatic groups
using a maximum likelihood estimation with saturated correlates model for
the full sample (n = 60). Maternal demographics, pretreatment and historical
clinical characteristics, and intervention parameters (number of IH-CBT ses-
sions and number of home visits) were examined. Race, income, and marital
status were added as missing data correlates to improve the accuracy of the
Table 2. Means and SDs for Asymptomatic and Symptomatic Groups and Results From Comparisons Using MANOVA
Variable Asymptomatic Symptomatic M difference SE Wald Z95% CI Cohen’s d
Mother age 22.30 (4.72) 23.50 (4.85) −1.20 0.72 −1.67* [−2.06, −0.15] 0.53
Mother education 11.61 (3.41) 11.56 (3.40) 0.05 0.19 0.27 [−0.21, 0.29] 0.03
CTQ 59.18 (7.69) 61.56 (7.65) −2.37 3.60 −0.66 [−7.90, 1.94] 0.50
Age 1st episode 15.06 (3.88) 14.80 (3.85) 0.27 1.44 0.19 [−1.04, 2.87] 0.11
# MDD episodes 2.74 (1.65) 3.18 (1.78) −0.44 0.16 −2.77** [−0.39, −0.01] 0.33
# Diagnoses 1.85 (1.36) 2.00 (1.41) −0.15 0.13 −1.19 [−0.23, 0.12] 0.12
BDI-II pre-treatment 33.46 (5.78) 36.39 (6.03) −2.93 1.44 −2.04* [−5.03, −1.56] 0.95
IPDS 4.67 (2.16) 5.83 (2.42) −1.17 0.43 −2.73** [−1.80, −0.75] 0.76
# IH-CBT sessions 13.12 (3.62) 10.94 (3.31) 2.18 0.77 2.83** [1.42, 3.28] 1.04
# Home visits 13.94 (3.73) 8.83 (3.97) 5.11 0.55 9.37** [4.39, 5.94] 2.27
Note: CI = confidence interval; CTQ = Childhood Trauma Questionnaire; MDD = major depressive disorder ; BDI-II = Beck Depression Inven-
tory–II; IPDS = Iowa Personality Disorder Screen; IH-CBT = In-Home Cognitive-Behavioral Therapy.
*p < .05. **p < .01 based on Wald Z values > ±1.645; 95% CI’s based on 1,000 bootstrap samples.
472 Behavior Modiﬁcation 36(4)
estimates. Marital status was not considered given that most of the sample was
unmarried. MANOVA revealed six variables that were statistically significant
and differentiated asymptomatic and symptomatic groups. In contrast with
mothers who still had clinically elevated levels of depression at posttreatment,
asymptomatic mothers were younger, had fewer lifetime episodes of MDD,
had lower BDI-II scores at pretreatment, had lower levels of symptoms sug-
gesting a personality disorder, and received more IH-CBT sessions and home
visits during the treatment interval. Effect sizes were medium to large in size,
with particularly noteworthy effect sizes emerging for number of treatment
sessions and number of home visits. Taken together, a variety of participant
characteristics and service parameters separately contributed to predicting a
low level of depressive symptoms at posttreatment.
Regarding number of home visits, follow-up analyses indicated that this
effect was particularly evident during the first half of treatment. Although
asymptomatic mothers received more home visits than their symptomatic
counterparts during the first half (t = 4.32, p < .001) and second half (t = 2.11,
p < .05) of treatment, the gap was most evident during the first half (M home
visits = 8.20 [SD = 3.44] vs. 4.54 [SD = 2.41]; Cohen’s d effect size estimate
= 2.08) than the second half (M home visits = 6.52 [SD = 3.28] vs. 4.58
[SD = 3.16]; Cohen’s d effect size estimate = 1.08). Thus, although increased
home visits were observed throughout the treatment interval, mothers who
were asymptomatic at posttreatment received almost twice as many home
visits in the first half of treatment than those who ended in the symptomatic
category. The inverse relationship between number of home visits and post-
treatment BDI-II scores is further elucidated in Figure 1 which shows the
scatterplot with fitted trend line for number of home visits and endpoint
A hierarchical linear regression was next conducted to identify predictors
of posttreatment BDI-II scores while considering the demographic, clinical,
and intervention process variables that emerged from the group contrasts.
Pretreatment BDI-II scores and number of home visits received prior to com-
mencement of IH-CBT were also included. Table 3 presents results of the
regression (F = 5.87, p < .001, R2 = .446). Two variables emerged as statisti-
cally significant (p < .05) predictors in the full model: number of home visits
and symptoms of personality disorders as measured by the IPDS. Number of
IH-CBT sessions approached (p = .08) but did not reach statistical signifi-
cance. Despite this, results from the regression indicate that number of
IH-CBT sessions and home visits contribute uniquely to positive depression
Ammerman et al. 473
Table 3. Results of Hierarchical Linear Regression Predicting Posttreatment BDI-II
b SE p value 95% CI (b)
BDI-II pre-treatment 0.21 0.22 0.17 [−0.23, 0.62]
Pre-txt home visits −0.15 0.21 0.22 [−0.57, 0.24]
Mother age 0.34 0.39 0.20 [−0.58, 0.99]
IPDS 1.91 0.88 0.02* [0.11, 3.47]
# MDD episodes 1.10 1.41 0.21 [−2.62, 2.95]
# IH-CBT sessions −0.90 0.63 0.08 [−2.19, 0.27]
# Home visits −0.78 0.42 0.03* [−1.52, 0.18]
Note: CI = confidence interval; BDI-II = Beck Depression Inventory–II; IPDS = Iowa Personality
Disorder Screen; MDD = major depressive disorder; IH-DBT = In-Home Cognitive-Behavioral
*p < .05. 95% CIs computed based on 1,000 bootstrap samples.
Figure 1. Scatterplot showing association between number of home visits and
posttreatment BDI-II scores
Note: BDI-II = Beck Depression Inventory–II. R2 = .19.
474 Behavior Modiﬁcation 36(4)
This study identified predictors of treatment outcome in depressed mothers
who received IH-CBT and concurrent home visiting. IH-CBT is unique in
that it is an adapted treatment that was designed to be engaging, relevant, and
effective for new mothers participating in home visiting programs. Results
showed that 64.7% of treated mothers obtained a posttreatment BDI-II score
of ≤8 indicative of asymptomatic status, a pattern consistent with other effec-
tive treatments for perinatal depression (Sockol, Epperson, & Barber, 2011).
Delineating predictors of reaching a clinical threshold indicative of positive
treatment response is important to refining and improving IH-CBT. In addi-
tion, predictors can be used a priori to develop treatment plans for mothers
who may be expected to respond more or less robustly to IH-CBT. As
hypothesized, a number of variables emerged as predictive of treatment out-
come involving demographic, clinical, and intervention process factors.
Consistent with hypotheses, number of IH-CBT sessions was predictive
of better depression outcomes. Yet, the most robust predictor was number
of home visits. Mothers who received increased home visits during treat-
ment were more likely to be asymptomatic at posttreatment than those with
fewer home visits even when controlling for frequency of home visits prior
to treatment and other predictor variables (including number of IH-CBT
sessions). IH-CBT is distinctive in that it is designed to be implemented
alongside home visiting. A close working relationship between therapists
and home visitors is posited to be essential to optimal outcomes. Number of
home visits is a reasonable proxy of the strength of this collaborative effort.
As such, results suggest that the active involvement of home visitors with
mothers during the treatment interval, and particularly during the first half
of treatment, propels mothers toward lower depression severity at posttreat-
ment. It is possible that the increased frequency of home visits works syn-
ergistically with treatment to bring about better outcomes, which would be
consistent with the theoretical foundation of IH-CBT. An alternative expla-
nation is that mothers who are on a trajectory toward improvement are more
engaged in home visiting and seek additional contact with home visitors.
However, home visitors may be drawn to mothers who are emerging from
a depressive state. They may capitalize on mothers’ increased energy and
interest by increasing home visit frequency. There may also be an unknown
variable that is associated with increased home visits and IH-CBT treat-
ment sessions that contributes to better outcomes. Although home visitors
strive to follow a schedule of visits as dictated by home visiting models,
there is substantial variability in intensity of service due to maternal
Ammerman et al. 475
willingness and availability and perceived need. Findings from this study
suggest that, for depressed mothers who receive IH-CBT treatment, consis-
tent and regular contact by home visitors is associated with more robust
decreases in depressive symptoms.
Other variables also predicted posttreatment depression. Younger moth-
ers were more likely to be asymptomatic than their older counterparts. In
the larger literature on adults treated for MDD with psychotherapy (includ-
ing CBT) or medications, findings have been mixed (Driessen & Hollon,
2010). Most recently, Fournier et al. (2009) found that younger adults were
more responsive to CBT than older adults. Similarly, the Treatment for
Adolescents With Depression Study (TADS; Curry et al., 2006) found that
younger adolescents had superior outcomes following treatment than older
peers. It is not clear why younger age is associated with improved out-
comes, and this finding warrants continued research. It is possible that less
experience with depression improves responsiveness to IH-CBT treatment.
To this end, it is noteworthy that age was not a significant contributor to
posttreatment BDI-II scores when considered alongside other predictors,
such as number of MDD episodes.
Severity emerged as a predictor of posttreatment depression, at least in the
form of number of MDD episodes and symptoms of personality disorders.
Depression chronicity, reflected in part by number of episodes, has been con-
sistently found to be associated with less improvement in CBT treatment.
Chronicity may reflect a more recalcitrant form of depression. Likewise, there
is evidence that repeated episodes increase susceptibility to subsequent depres-
sion (Monroe & Harkness, 2011), which may in turn diminish treatment effec-
tiveness. Baseline severity was also associated with a less robust response to
treatment. For mothers who have higher pretreatment levels of depression and
more episodes, they may require additional treatment. Age of onset, a variable
reflecting illness severity that has been found to be related to treatment
outcome in other adult populations, was not a significant predictor of post-
treatment depressive levels in mothers receiving IH-CBT. In this population of
young, low income, new mothers it is not a useful indicator of treatment
response. In contrast, increased symptoms of personality disorders predicted
less improvement, and this variable was significant even after controlling for
other predictors in a regression model. Personality disorders, particularly
those reflecting a submissive personality style, are associated with chronicity
of depression (Cain et al., 2011). Although CBT treatment specifically focused
on adults with personality disorders has been found to be efficacious
(Matusiewicz, Hopwood, Banducci, & Lejuez, 2010), the co-occurrence of
MDD and personality disorders has been found to be detrimental to outcomes
476 Behavior Modiﬁcation 36(4)
in adults receiving depression-focused CBT (Fournier et al., 2008). As with
chronicity, additional IH-CBT sessions or a more direct focus on personality
disorder mechanisms may be needed to boost outcomes for mothers in home
visiting programs. Finally, in contrast, history of childhood trauma (which was
widely represented in the sample) did not predict differential levels of post-
This is the first study to examine predictors of outcomes in IH-CBT, and one
of the few efforts that focused on depressed mothers who were young, had low
income, and were new parents. The study had a number of strengths. First, an
array of theoretically meaningful predictors was examined reflecting demo-
graphic, clinical, and intervention process variables. Second, multiple measure-
ment strategies were used, including self-report and clinical interview. Third,
maximum likelihood estimation with a saturated correlates model was used to
address missing data and take advantage of the full sample. Fourth, two widely
disseminated models of home visitation were represented in the sample thereby
facilitating generalizability of findings.
There are also several limitations that warrant caution in interpreting find-
ings. First, the sample size was relatively small, limiting statistical power.
Second, determination of asymptomatic versus symptomatic status at posttreat-
ment is one of several ways that treatment response can be described. Others
include recovery or remission (Keller, 2003), both of which require an extended
interval during which depressive symptoms remain below a preestablished
level of clinical significance. The predictors found in the current study may not
necessarily emerge under different definitions of treatment response. Third,
although the IPDS is a psychometrically acceptable screen for symptoms of
personality disorders, it is not a substitute for more comprehensive clinical
interviews that more definitively diagnose specific personality disorders.
Fourth, although number of IH-CBT sessions and home visits were considered,
quality of services was not addressed.
Findings from this study have important implications for clinical practice
in the treatment depression in new mothers participating in home visiting
programs. A high rate of IH-CBT treatment sessions and home visits, admin-
istered concurrently, are likely to bring about the most robust reductions in
depressive symptoms. As IH-CBT explicitly promotes and facilitates collab-
oration between therapists and home visitors, it is better able to create this
synergistic effect than other treatments. Depression treatment received in the
community is less likely to produce the close working relationship between
clinicians and home visitors, given structural and logistical barriers. Several
pretreatment and historical features of depression (e.g., personality disorder
symptoms, number of MDD episodes) may indicate a need for additional
Ammerman et al. 477
treatment, although this is an empirical question that awaits additional
research. Results from this study underscore the need to conduct research on
new mothers participating in home visiting programs separately given that
findings from studies of other adult populations may not be generalizable to
Frank W. Putnam is now at the Department of Psychiatry, University of North
Carolina School of Medicine.
The authors acknowledge the participation and support of The Health Foundation of
Greater Cincinnati, United Way of Greater Cincinnati, Kentucky H.A.N.D.S., Ohio
Help Me Grow, and www.OhioCanDo4Kids.org.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research,
authorship, and/or publication of this article.
The author(s) disclosed receipt of the following financial support for the research,
authorship, and/or publication of this article: This study was supported by Grant
R34MH073867 from the National Institute of Mental Health.
Ammerman, R. T., Bodley, A. L., Putnam, F. W., Lopez, W. L., Holleb, L. J.,
Stevens, J., & Van Ginkel, J. B. (2007). In-Home Cognitive Behavior Therapy for a
depressed mother in a home visitation program. Clinical Case Studies, 6, 161-180.
Ammerman, R. T., Putnam, F. W., Altaye, M., Chen, L., Holleb, L. J., Stevens,
J., . . . Van Ginkel, J. B. (2009). Changes in depressive symptoms in first time
mothers in home visitation. Child Abuse & Neglect, 33, 127-138. doi:10.1016/
Ammerman, R. T., Putnam, F. W., Altaye, M., Stevens, J., & Van Ginkel, J. B. (2012).
A clinical trial of In-Home CBT for depressed mothers in home visitation. Cincin-
nati, OH: Cincinnati Children’s Hospital Medical Center.
Ammerman, R. T., Putnam, F. W., Bosse, N. R., Teeters, A. R., & Van Ginkel, J. B.
(2010). Maternal depression in home visitation: A systematic review. Aggression
and Violent Behavior, 15, 191-200. doi:10.1016/j.avb.2009.12.002
478 Behavior Modiﬁcation 36(4)
Ammerman, R. T., Putnam, F. W., Chard, K. M., Stevens, J., & Van Ginkel, J. B.
(2011). PTSD in depressed mothers in home visitation. Psychological Trauma:
Theory, Research, Practice, and Policy. doi:10.1037/a0023062
Ammerman, R. T., Putnam, F. W., Kopke, J. E., Gannon, T. A., Short, J. A., Van Ginkel,
J. B., . . . Spector, A. R. (2007). Development and implementation of a quality assur-
ance infrastructure in a multisite home visitation program in Ohio and Kentucky.
Journal of Prevention and Intervention in the Community, 34, 89-107. doi:10.1300/
Ammerman, R. T., Putnam, F. W., Stevens, J., Bosse, N. R., Short, J. A., Bodley,
A. L., & Van Ginkel, J. B. (2011). An open trial of In-Home CBT for depressed
mothers in home visitation. Maternal and Child Health Journal, 15, 1333-1341.
Ammerman, R. T., Shenk, C. E., Teeters, A. R., Noll, J. G., Putnam, F. W., & Van
Ginkel, J. B. (2011). Impact of depression and childhood trauma in mothers
receiving home visitation. Journal of Child and Family Studies. doi:10.1007/
Beck, A. T., Steer, R. A., & Brown, G. K. (1996). BDI-II Manual. San Antonio, TX:
The Psychological Corporation.
Beck, J. S. (2011). Cognitive behavior therapy: Basics and beyond (2nd ed.). New
York, NY: Guilford.
Bernstein, D. P., Stein, J. A., Newcomb, M. D., Walker, E., Pogge, D., Ahluvalia, T.,
. . . Zule, W. (2003). Development and validation of a brief screening version
of the Childhood Trauma Questionnaire. Child Abuse & Neglect, 27, 169-190.
Boller, K., & Strong, D. A. (2010). Home visiting: Looking back and moving forward.
Zero to Three, 30(6), 4-9.
Cain, N. M., Ansell, E. B., Wright, A. G., Hopwood, C. J., Thomas, K. M., Pinto,
A., . . . Grilo, C. M. (2011). Interpersonal pathoplasticity in the course of major
depression. Journal of Consulting and Clinical Psychology, 24, 595-606.
Cox, J. L., Holden, J. M., & Sagovsky, R. (1987). Detection of postnatal depression.
Development of the 10-item Edinburgh Postnatal Depression Scale. British Journal
of Psychiatry, 150, 782-786.
Curry, J., Rohde, P., Simons, A., Silva, S., Vitiello, B., Kratochvil, C., . . . March,
J. (2006). Predictors and moderators of acute outcome in the Treatment for Adoles-
cents with Depression Study (TADS). Journal of the American Academy of Child
and Adolescent Psychiatry, 45, 1427-1439. doi:10.1097/01.chi.0000240838.78984.
e2 00004583-200612000-00005 [pii]
Driessen, E., & Hollon, S. D. (2010). Cognitive behavioral therapy for mood disor-
ders: Efficacy, moderators and mediators. Psychiatric Clinics of North America,
33, 537-555. doi:10.1016/j.psc.2010.04.005
Ammerman et al. 479
Enders, C. K. (2010). Applied missing data analysis. New York, NY: Guilford.
Fournier, J. C., DeRubeis, R. J., Shelton, R. C., Gallop, R., Amsterdam, J. D., &
Hollon, S. D. (2008). Antidepressant medications v. cognitive therapy in people
with depression with or without personality disorder. British Journal of Psychiatry,
192, 124-129. doi:192/2/124 [pii] 10.1192/bjp.bp.107.037234
Fournier, J. C., DeRubeis, R. J., Shelton, R. C., Hollon, S. D., Amsterdam, J. D., &
Gallop, R. (2009). Prediction of response to medication and cognitive therapy in
the treatment of moderate to severe depression. Journal of Consulting and Clini-
cal Psychology, 77, 775-787. doi:2009-11168-016 [pii] 10.1037/a0015401
Haby, M. M., Donnelly, M., Corry, J., & Vos, T. (2006). Cognitive behavioural therapy
for depression, panic disorder and generalized anxiety disorder: A meta-regres-
sion of factors that may predict outcome. Australian and New Zealand Journal of
Psychiatry, 40, 9-19. doi:ANP1736 [pii] 10.1111/j.1440-1614.2006.01736.x
Hamilton, K. E., & Dobson, K. S. (2002). Cognitive therapy of depression: Pretreat-
ment patient predictors of outcome. Clinical Psychology Review, 22, 875-893.
Hay, D. F., Pawlby, S., Waters, C. S., Perra, O., & Sharp, D. (2010). Mothers’ ante-
natal depression and their children’s antisocial outcomes. Child Development, 81,
Holton, J. K., & Harding, K. (2007). Healthy Families America: Ruminations
on implementing a home visitation program to prevent child maltreatment.
Journal of Prevention and Intervention in the Community, 34, 13-38. doi:10.1300/
Keller, M. B. (2003). Past, present, and future directions for defining optimal treatment
outcome in depression: Remission and beyond. Journal of the American Medical
Association, 289, 3152-3160. doi:10.1001/jama.289.23.3152
Langbehn, D. R., Pfohl, B. M., Reynolds, S., Clark, L. A., Battaglia, M., Bellodi,
L., . . . Links, P. (1999). The Iowa Personality Disorder Screen: Development
and preliminary validation of a brief screening interview. Journal of Personality
Disorders, 13, 75-89.
Levy, L. B., & O’Hara, M. W. (2010). Psychotherapeutic interventions for depressed,
low-income women: A review of the literature. Clinical Psychology Review, 30,
Matusiewicz, A. K., Hopwood, C. J., Banducci, A. N., & Lejuez, C. W. (2010).
The effectiveness of cognitive behavioral therapy for personality disorders.
Psychiatric Clinics of North America, 33, 657-685. doi:S0193-953X(10)00049-3
Monroe, S. M., & Harkness, K. L. (2011). Recurrence in major depression: A con-
ceptual analysis. Psychological Review, 118, 655-674. doi:2011-20042-001 [pii]
480 Behavior Modiﬁcation 36(4)
Muthen, L. K., & Muthen, B. O. (2010). Mplus User’s Guide (6th ed.). Los Angeles,
National Research Council and Institute of Medicine. (2009). Depression in par-
ents, parenting, and children: Opportunities to improve identification, treatment,
and prevention. Committee on Depression, Parenting Practices, and the Healthy
Development of Children, Board on Children, Youth, and Families, Division on
Behavioral and Social Sciences and Education. Washington, DC: The National
Olds, D. L. (2010). The Nurse-Family Partnership: From trials to practice. In A. J.
Reynolds, A. J. Rolnick, M. M. Englund & J. A. Temple (Eds.), Childhood pro-
grams and practices in the first decade of life: A human capital integration (pp.
49-75). New York, NY: Cambridge University Press.
Scher, C. D., Stein, M. B., Asmundson, G. J., McCreary, D. R., & Forde, D. R.
(2001). The childhood trauma questionnaire in a community sample: Psycho-
metric properties and normative data. Journal of Traumatic Stress, 14, 843-857.
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, 316-321. doi:10.1007/
Shear, M. K., Greeno, C., Kang, J., Ludewig, D., Frank, E., Swartz, H. A., &
Hanekamp, M. (2000). Diagnosis of nonpsychotic patients in community clin-
ics. American Journal of Psychiatry, 157, 581-587. doi:0.1176/appi.ajp.157.4.581
Sockol, L. E., Epperson, C. N., & Barber, J. P. (2011). A meta-analysis of treatments
for perinatal depression. Clinical Psychology Review, 31, 839-849. doi:10.1016/j.
cpr.2011.03.009 S0272-7358(11)00057-2 [pii]
Spitzer, R. L., Williams, J. B., Gibbon, M., & First, M. B. (1992). The Structured
Clinical Interview for DSM-III-R (SCID). I: History, rationale, and description.
Archives of General Psychiatry, 49, 624-629.
Spitzer, R. L., Williams, J. B., Kroenke, K., Linzer, M., deGruy, F. V., 3rd, Hahn, S.
R., . . . Johnson, J. G. (1994). Utility of a new procedure for diagnosing mental
disorders in primary care. The PRIME-MD 1000 study. Journal of the American
Medical Association, 272, 1749-1756.
Trull, T. J., & Amdur, M. (2001). Diagnostic efficiency of the Iowa Personality Dis-
order Screen items in a nonclinical sample. Journal of Personality Disorders, 15,
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, 369-374. doi:10.1053/comp.2001.24556
Ammerman et al. 481
Zwarenstein, M., Goldman, J., & Reeves, S. (2009). Interprofessional collaboration: Effect s
of practice-based interventions on professional practice and healthcare outcomes.
Cochrane Database of Systematic Reviews, 3, CD000072. doi:10.1002/14651858.
Robert T. Ammerman, PhD, is a professor of pediatrics in the Behavioral Medicine
and Clinical Psychology Division and scientific director of Every Child Succeeds at
Cincinnati Children’s Hospital Medical Center. His research interests are enhance-
ment of early prevention programs, treatment of maternal depression, and quality
improvement in prevention.
James L. Peugh, PhD, is an assistant professor of pediatrics in the Behavioral
Medicine and Clinical Psychology Division at Cincinnati Children’s Hospital
Medical Center. His research interests include Monte Carlo simulations in cross-
sectional, longitudinal, and multilevel mixture structural equation models, missing
data handling, and dyadic data analysis.
Frank W. Putnam, MD, is a professor of psychiatry at the University of North
Carolina. His research focuses on the implementation of scalable interventions for
prevention and treatment of child maltreatment in a variety of settings.
Judith B. Van Ginkel, PhD, is a professor of pediatrics and president of Every Child
Succeeds at Cincinnati Children’s Hospital Medical Center. Her research interests are
effectiveness of home visiting programs and early childhood policy.