Women with Comorbid Obesity and Depression

A Randomized Controlled Trial of Behavioral Weight Loss
Treatment Versus Combined Weight Loss/Depression Treatment
Among Women with Comorbid Obesity and Depression
Jennifer A. Linde, Ph.D.,
Division of Epidemiology and Community Health, School of Public Health, University of
Minnesota, Minneapolis, MN, USA
Gregory E. Simon, M.D., M.P.H.,
Group Health Research Institute, Group Health Cooperative, Seattle, WA, USA
Evette J. Ludman, Ph.D.,
Group Health Research Institute, Group Health Cooperative, Seattle, WA, USA
Laura E. Ichikawa, M.S.,
Group Health Research Institute, Group Health Cooperative, Seattle, WA, USA
Belinda H. Operskalski, M.P.H.,
Group Health Research Institute, Group Health Cooperative, Seattle, WA, USA
David Arterburn, M.D.,
Group Health Research Institute, Group Health Cooperative, Seattle, WA, USA
Paul Rohde, Ph.D.,
Oregon Research Institute, Eugene, OR, USA
Emily A. Finch, M.A., and
Indiana University School of Medicine, Indianapolis, IN, USA
Robert W. Jeffery, Ph.D.
Division of Epidemiology and Community Health, School of Public Health, University of
Minnesota, Minneapolis, MN, USA
Jennifer A. Linde: linde074@umn.edu
Abstract
Background—Obesity is associated with clinical depression among women. However,
depressed women are often excluded from weight loss trials.
Purpose—This study examined treatment outcomes among women with comorbid obesity and
depression.
Methods—Two hundred three (203) women were randomized to behavioral weight loss (n=102)
or behavioral weight loss combined with cognitive-behavioral depression management (n=101).
Results—Average participant age was 52 years; mean baseline body mass index was 39 kg/m2.
Mean Patient Health Questionnaire and Hopkins Symptom Checklist (SCL-20) scores indicated
© The Society of Behavioral Medicine 2010
Correspondence to: Jennifer A. Linde, linde074@umn.edu.
Conflict of Interest Statement The authors have no conflicts of interest to disclose.

Published in final edited form as:
Ann Behav Med. 2011 February ; 41(1): 119–130. doi:10.1007/s12160-010-9232-2.
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript
moderate to severe baseline depression. Weight loss and SCL-20 changes did not differ between
groups at 6 or 12 months in intent-to-treat analyses (p=0.26 and 0.55 for weight, p=0.70 and 0.25
for depressive symptoms).
Conclusions—Depressed obese women lost weight and demonstrated improved mood in both
treatment programs. Future weight loss trials are encouraged to enroll depressed women.
Keywords
Obesity; Depression; Comorbidity; Intervention; Women
Introduction
Obesity is associated with clinically significant depressive symptoms, particularly among
women [1,2]. Women are more likely to seek weight loss treatment on their own or be
referred by their physicians for specialist weight loss treatment than men [3]. In addition,
adults who seek weight loss treatment present with higher rates of depression than those
seeking other medical treatment [4]. Adding to the burden of comorbidity between the two
conditions, depression is associated with attrition from weight loss programs [5], poorer
weight loss outcomes [6], and weight regain after successful losses [7]. Even though rates of
clinically significant depression exceed 25% in obese women [2], data regarding treatment
of co-occurring obesity and depression are limited. Women with significant depression are
often excluded from trials of weight loss intervention because of untested assumptions that
standard weight loss interventions would be ineffective or inappropriate [8,9].
Experts in the fields of weight control and depression management have concluded
independently that cognitive-behavioral treatment approaches that address behavior
activation, problem solving, and reassessment of condition-specific beliefs are optimal ways
to address each area of concern [10–14]. For both conditions, these treatments are effective
when delivered in either individual or group formats [12,14]. However, group cognitivebehavioral approaches that address both obesity and depression simultaneously have not
been examined. To our knowledge, the only related study is a pending trial to treat
depressive symptoms with a combination of individual and group visits prior to group
weight loss treatment [15].
The purpose of the current study was to examine the effects on weight loss and depressive
symptoms of standard behavior therapy for obesity with specialized treatment addressing
both weight loss and depression simultaneously among women who were obese and had
comorbid clinical depression. Women were randomly assigned either to behavioral weight
loss treatment or to a combination of behavioral weight loss and cognitive-behavioral
therapy for depression. To observe whether depression had a negative effect on motivation
for study participation, participants were recruited for the study based solely on weight and
depression status.
Methods
Study Recruitment
Recruitment took place from April 2003 to April 2005. Eligible individuals were enrolled in
the Group Health Cooperative, a mixed-model prepaid health plan serving over 500,000
members in Washington and northern Idaho. The study was approved by the Group Health
Cooperative Institutional Review Board.
Linde et al. Page 2
Ann Behav Med. Author manuscript; available in PMC 2012 February 1.
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript
Participants were recruited from a cohort of women in eight Group Health primary care
clinics between the ages of 45 and 65 who had completed a population-based survey
delivered over the telephone by trained study staff and designed to establish the prevalence
of clinical depression and obesity [16]. Existing data from computerized administrative
records and the Group Health Breast Cancer Screening Program were used to identify three
strata for survey sampling based on self-reported height and weight, from which body mass
index (BMI) status was calculated [BMI<30, BMI≥30, or BMI unknown (did not participate
in screening program)]. Computerized administrative data were used to exclude those who
were diagnosed with bipolar disorder or psychotic disorder in the last 2 years. Details of
recruitment and procedures are described elsewhere [16].
Women with BMI≥30 and who met depressive disorder criteria from the survey sample
were invited to an in-person baseline interview to determine eligibility for the present study.
Depression status was indicated by scores in the moderate range or higher (≥10) on the
Patient Health Questionnaire-9 (PHQ-9) [17] administered as part of the epidemiological
survey described above, which assesses depressive disorder via self-reported items based on
criteria in the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSMIV) [18]. Of the 285 depressed women (based on survey screening) invited to these baseline
interviews, 68 declined the interview, 11 were ineligible for the treatment study, and three
declined participation in the treatment study, for a final sample of 203 depressed and obese
women randomized into this study (see Fig. 1). Randomization to the treatment conditions
described here was conducted based on PHQ-9 depression status assessed during the
epidemiological survey. All participants provided written informed consent prior to this
baseline interview, which was conducted in person by trained survey interviewers. Women
were paid $20 as compensation for completing the interview.
Measures
At baseline and 6 and 12 months, participants completed in-person interviews conducted by
trained study interviewers who were blinded to participants’ treatment assignment, for
collection of anthropometric and survey data. Each participant was encouraged to complete
6 and 12-month interviews regardless of her level of treatment participation. Participants
were compensated $20 for completion of each interview. Baseline interviews took place
between April 2003 and April 2005, 6-month interviews took place between October 2003
and November 2005, and 12-month interviews took place between May 2004 and April
2006. All measures are described as follows:
Demographic Characteristics—A brief questionnaire assessed age, race/ethnicity,
marital status, and educational attainment at baseline.
Height and Weight—Height was measured by study staff at baseline using a wallmounted stadiometer. Weight was assessed on a calibrated balance beam scale with women
wearing light street clothes without shoes at baseline and 6 and 12 months. BMI (kilograms/
m2) was calculated from measured height and weight.
Depression Diagnosis—The Structured Clinical Interview for DSM-IV or SCID [19]
mood module was used to assess current depression diagnosis at the baseline interview, after
randomization but before the start of treatment. The SCID is a semi-structured interview to
diagnose major DSM-IV mental disorders and has been used in studies as the gold standard
in determining the accuracy of clinical diagnoses [20]. A diagnosis of current major
depression was conferred if a participant reported a major depressive episode during the
previous 2 weeks (i.e., depressed mood for at least 2 weeks and/or loss of interest in
typically enjoyable activities, plus at least four additional symptoms of depression such as
Linde et al. Page 3
Ann Behav Med. Author manuscript; available in PMC 2012 February 1.
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript
feelings of worthlessness, fatigue, insomnia or hypersomnia, difficulty concentrating,
psychomotor changes, appetite changes, unintended weight changes, or suicidal ideation, as
specified in DSM-IV).
Depression Severity—Depressive symptom severity was assessed at baseline and 6 and
12 months using the Symptom Checklist-20 (SCL-20), a 20-item depression scale extracted
from the SCL-90, scored on a 0 to 4 scale [21]. The SCL has been found to have high
reliability and validity in multiple studies with medical patients and is sensitive to change in
depressed primary care patients [21]. A score of 1.72 on the SCL-20 has been shown to have
the highest positive predictive value for major depression [22]. Previous trials in primary
care that have used the SCL depression scale score as an outcome measure support the
following depression severity estimates: 0.75=remission, greater than 0.75 to 1.5=mild
depression, greater than 1.5 to 2.0=moderate depression, and greater than 2.0=severe
depression [23,24].
Physical Activity—The Paffenbarger Activity Questionnaire or PAQ [25] was
administered at baseline and 6 and 12 months to provide an estimate of calories expended
per week in leisure time physical activities. It assessed activities performed in the previous
week, including specific activities of stairs climbed and blocks walked and minutes of any
other activities using a free recall format. Open-ended responses are scored by categorizing
activities as light (5 kcal/min), medium (7.5 kcal/min), or high (10 kcal/min) intensity [26].
The PAQ has been shown to predict weight outcomes in previous weight loss intervention
trials [27].
Dietary Intake—Dietary intake was assessed at baseline and 6 and 12 months using a food
frequency questionnaire developed by the Nutrition Assessment Shared Resource of the
Fred Hutchinson Cancer Research Center [28,29]. Nutrient calculations were performed
using the Nutrient Data System for Research software version 2005, developed by the
Nutrition Coordinating Center at the University of Minnesota. Due to changes in the survey
protocol during the epidemiological survey from which the current study participants were
recruited [16], Nutrition Assessment Shared Resource dietary intake data are not available
for 27 participants; missing data are distributed equally between treatment groups.
Antidepressant Use—Computerized pharmacy refill records were used to examine use
of antidepressant medications during the 12-month intervention period.
Intervention Procedures
A computerized randomized selection process conducted by study programmers was
employed to assign participants to groups; this process used a “biased coin” strategy (in
which the randomization method applied a differential probability to the selection process
depending on divergence of numbers of cases between groups) to ensure that assignments
were equally distributed between the two groups [30,31]. Assignments were computer
generated after the baseline assessment; whereas participants and intervention staff were not
blinded to group assignment, interviewers conducting evaluations for the study (as described
in Measures above) were blinded to group assignment. Time between randomization and
intervention start date was not normally distributed; the mean duration between
randomization and group start date was 92 days for both intervention conditions (25th and
75th percentile values: 36 and 97 days for weight loss groups, 41 and 139 days for weight
loss/depression groups). Women receiving treatment for weight loss or psychotherapy for
depression were excluded from entering the study, but participants were not restricted from
seeking depression or weight loss treatment outside the study protocol once enrolled.
Women taking antidepressant medications were not excluded from the study.
Linde et al. Page 4
Ann Behav Med. Author manuscript; available in PMC 2012 February 1.
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript
Intervention groups met between June 2003 and April 2006. The two treatment groups were
as follows:
Weight Loss Only Intervention—The behavioral weight loss intervention was adapted
from programs developed at the University of Minnesota over the last 20 years. This
program has been used in numerous randomized trials [32–35] and has consistently
produced mean weight losses at least as good as other currently available non-medical
treatments. Session content was centered on behavioral goal setting and self-monitoring of
caloric intake, physical activity, and body weight. Each session contained a topically
relevant quiz and motivational messages to enhance session content. Participants were given
daily caloric intake goals of 1,200 or 1,500 kcal as appropriate to produce weight loss of 1 to
2 lb per week, based on initial body weight. Participants were asked to restrict fat intake to
20% of daily caloric intake. Physical activity goals were increased in biweekly increments
of 500 kcal/week until the goal of 2,500 kcal/week was reached. Participants received
structured meal plans and specific skills training in environmental stimulus control (e.g., not
having high calorie snacks in the home). Some physical activity (e.g., group walks, light
strength training) was directed during sessions with specific activity content; otherwise,
participants were instructed to engage in activities of their choosing on their own or with
partners outside of sessions, and motivational materials and relevant quizzes were
distributed as part of session content. Participants were asked to keep food intake and
physical activity diaries that were reviewed weekly by study interventionists. The program
did not specifically address symptoms of depression. Treatment was delivered by trained
weight loss counselors with prior experience in nutrition or exercise physiology.
Combined Weight Loss/Depression Intervention—The combined intervention
integrated the behavioral weight loss program described above with essential elements of
cognitive-behavioral treatment for depression, as drawn from the Coping with Depression
manual [36]. Approximately 55% of didactic content in the combined program addressed
treatment for depression, with the remaining 45% of the content addressing behavioral
weight loss treatment; session content was interlaced, with order and content determined by
experts on the research team who represented each discipline. In addition to completing
behavioral goal setting and self-monitoring tasks for weight loss as described above,
participants in the combined groups set additional behavioral and cognitive change goals to
improve depressive symptoms and completed regular mood self-monitoring assignments as
part of their weekly diaries. Combined groups were led by a licensed clinical psychologist
who was cross-trained to deliver weight loss content by expert members of the research
team; a second licensed clinical psychologist (also a member of the research team) was
available as needed.
Intervention was delivered in 21 groups (13 weight loss only, eight combined) with 3–16
participants assigned to each group (3–11 for weight loss only; 4–16 for combined);
variability of group size was driven by participant preferences for group times and locations.
Groups were held during daytime and evening hours at three locations in the Seattle
metropolitan area. Thirty-eight women (22 weight loss only, 16 combined) were not
assigned to an intervention group because they were unable or unwilling to attend any of the
21 available groups (see Fig. 1 for study participant flow information). An additional 125
obese women who did not meet criteria for major depressive disorder were assigned to
weight loss only groups; outcomes for depressed versus nondepressed women in weight loss
only groups are described elsewhere [37].
The intervention, which took place in 26 sessions over 1 year, began with weekly sessions
for 16 weeks, followed by four sessions every other week, then tapering to monthly
maintenance sessions for 6 months; a summary of intervention group content is presented in
Linde et al. Page 5
Ann Behav Med. Author manuscript; available in PMC 2012 February 1.
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript
the Appendix. Weight loss only groups lasted 90 min, and combined groups lasted 120 min.
Each session included collection of self-monitoring logs and review of homework
assignments, didactic instruction, small-group discussion, a quiz to reinforce session
content, and specific skills training and practice. Participants received written summaries of
material covered in each treatment session. To enhance treatment fidelity and to provide
therapist feedback, all treatment delivery staff participated in weekly live telephone
supervision sessions that included expert consultation with relevant members of the research
team. During these weekly calls, therapists briefed the research team on the previous week’s
sessions, discussed any challenges or questions that had been raised in sessions or other
participant contacts, and were given the opportunity to consult directly with weight loss or
depression experts on the study team. Experts also answered any questions about delivery of
content for upcoming sessions to monitor ongoing comprehension of study materials and to
recalibrate procedures as needed.
Data Analysis
All analyses were conducted using SAS Version 9.1. Initial comparisons of baseline
measures between the two treatment groups were made using t tests for continuous variables
and chi-square tests for categorical variables. No significant differences were observed
between groups with regard to baseline demographic variables (age, race, marital status,
education), body weight, or depressive symptoms (p=0.18–0.90). Primary analyses
compared participants in terms of mean weight change and mean SCL-20 score change by
group from baseline to 6- and 12-month assessments using general linear regression models.
The study was powered to detect a mean 1.7 kg difference (SD± 5 kg) in weight and a mean
0.28-point difference (SD± 0.64 points) on the SCL-20 for depressed women in weight loss
only versus the combined intervention, assuming 80% power and a group size of 90
participants in each condition at 12 months. Cohen’s d statistic [38] was calculated to
determine effect sizes for weight and depression outcomes at 6 and 12 months. Secondary
analyses compared the treatment groups in terms of participation in the initial weekly
treatment sessions and attrition from treatment as well as changes in dietary intake or
physical activity during treatment.
To reduce bias, missing weight data were replaced prior to analysis using information from
two sources. The first source was the electronic medical record available for all Group
Health members; the closest weight within 3 months of the scheduled interview date was
extracted from the record. If electronic medical record data were unavailable, missing
weights were substituted with the last study observation carried forward. To assess the
accuracy of weights extracted from medical records as compared to measured study weights,
medical record weights within 3 months of the baseline assessment were compared to
measured weights at baseline; the resulting correlation was 0.98 [39]. Percentages of
measured weight data, electronic medical record data, and last observation carried forward
data at 6 and 12 months by treatment group are presented in Table 1. Across both treatment
groups at 6 months, 79% of weight data were measured for the study, 13% were taken from
electronic medical records, and 8% were taken from last observation carried forward; at 12
months, 74% of weight data were measured for the study, 14% were taken from electronic
medical records, and 12% were taken from last observation carried forward. Rates of
available data were greater for combined versus weight loss only participants at 6 months
[86% combined versus 71% weight loss only, χ2(1)=6.45, p<0.05]; however, differences
were absent at 12 months [76% combined versus 71% weight loss only, χ2(1)=0.57, p=0.44].
For depressive symptom data, missing SCL-20 scores at 6 and 12 months were replaced by
last observation carried forward.
Linde et al. Page 6
Ann Behav Med. Author manuscript; available in PMC 2012 February 1.
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript
Results
Participant Characteristics
Baseline participant characteristics are shown in Table 2. Women were 52 years of age on
average, and approximately 50% were married. Over 75% of women were of white
ethnicity, and nearly 90% had attained at least a high school education. Mean BMI was in
the obese range and did not differ significantly by treatment group (39.5 kg/m2 for weight
loss only, 38.6 kg/m2 for combined). Mean depression scores on the SCL-20, the primary
means of assessing depressive symptoms at follow-up interviews, indicated that study
participants were experiencing moderate depression at baseline; mean PHQ scores at the
point of study randomization were comparable, with participants scoring in the low end of
the moderate to severe range of depressive symptoms. Overall rates of antidepressant use
were higher than past month prevalence rates of approximately 10% reported by comparable
adults in NHANES survey data [40], with 77% of women having taken any antidepressant
medication and 52% having used antidepressant medication for 180 days or more. Among
the antidepressant medications used, participants were most often prescribed those
associated with weight loss or no weight effect (e.g., 34% fluoxetine, 19% buproprion)
rather than weight gain (e.g., 7% doxepin, 3% amitriptyline, 1% mirtazapine, 1%
imipramine). Rates of antidepressant use did not differ significantly between treatment
groups [75.4% weight loss only versus 78.2% combined, χ2(1)= 0.21, p=0.65] and were not
considered as a confounder of weight or mood group effects over time in the study. No
participants received additional formal weight loss treatment outside of the study protocol,
and rates of psychotherapy outside of the study protocol were small and comparable across
groups [9% weight loss only versus 7% combined, χ2(1)=0.25, p=0.62].
Treatment Session Participation
Session attendance rates by treatment group are presented in Fig. 1. Out of 26 possible
sessions, approximately one half of participants attended at least 12 sessions (43% weight
loss only, 50% combined), and an additional one quarter of participants (24% weight loss
only, 30% combined) attended at least one session. The remaining participants (33% weight
loss only, 20% combined) were assigned to treatment groups but did not attend any sessions.
Attendance rates did not differ significantly by treatment group [χ2(2)=4.12, p=0.13] nor
were there any differences in mean number of sessions attended by group [mean (standard
deviation)=11.3 (8.4) versus 10.5 (9.0) sessions attended in weight loss only versus
combined, t(201)=0.63, p=0.53]. No study-related adverse events were reported by
participants in either treatment group.
Changes in Weight and Depression Outcomes by Treatment Group
Results for changes in weight and depression outcomes by treatment group are presented in
Table 3. There was no statistically significant interaction between treatment group and
session attendance in weight loss or depression models (data not shown). During the study,
participants lost weight on average over time, with no statistically significant differences by
treatment group at 6 or 12 months (mean changes at 6 months, -2.8 kg weight loss only
versus -1.8 kg combined, p=0.26; 12 months, -3.1 kg weight loss only versus -2.3 kg
combined, p= 0.55). Weight change results for those who were interviewed at 12-month
follow-up (n=73 weight loss only, n= 77 combined; data not shown) did not differ from
weight change outcomes for which missing data were imputed using a combination of
medical chart weights and last observation carried forward, as described in the “Data
Analysis” section above.
Depressed mood, as measured by the SCL-20, decreased over time as well, with scores
falling from moderate levels (1.82 weight loss only, 1.87 combined at baseline) to mild
Linde et al. Page 7
Ann Behav Med. Author manuscript; available in PMC 2012 February 1.
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript
levels by 6 months that were maintained at 12 months (6 months, 1.28 weight loss only, 1.36
combined; 12 months, 1.30 weight loss only, 1.22 combined). At baseline, 26% of
participants in each treatment group rated their depressive symptoms in the nonclinical to
mild range on the SCL-20 [weight loss only versus combined χ2(1)= 0.00, p=0.97]; by 12
months, 61% of those in weight loss only and 67% of those in combined groups rated their
depressive symptoms as nonclinical to mild [weight loss only versus combined χ2(1)=0.94,
p=0.33]. The change in proportion of those who reported nonclinical to mild symptoms at
baseline versus 12 months was statistically significant in each group [weight loss only
2χ
(1)=35.28, combined χ2(1)=24.92, p values<0.0001]. There were no statistically
significant differences in depressive mood changes by treatment group at 6 or 12 months in
general linear models (mean changes at 6 months, -0.54 weight loss only versus -0.50
combined, p=0.70; 12 months, -0.53 weight loss only versus -0.65 combined, p=0.25).
Depressive symptom change results for those who were interviewed at 12-month follow-up
(n=73 weight loss only, n=77 combined; data not shown) did not differ from outcomes for
which missing data were imputed by last observation carried forward as described in the
“Data Analysis” section above.
Effect sizes, calculated as Cohen’s d statistic, were derived from the statistical models of
weight changes and depressive symptom changes at 6 and 12 months. Effect sizes were
quite small for both weight changes (d=0.16 at 6 months, 0.09 at 12 months) and depressive
symptom changes (d=0.06 at 6 months, 0.16 at 12 months) at both assessment points.
Effects of Attendance on Weight Loss and Depression Outcomes
As described above and in Fig. 1, 43% of weight loss only participants and 50% of
combined participants attended 12 or more sessions (i.e., were exposed to at least 75% of
core content sessions). Therefore, weight change and depressive symptom change results
were also examined by attendance status (i.e., attending 12 or more sessions versus fewer
than 12 sessions) to determine the impact of attendance on outcomes. Results are presented
in Fig. 2a, b.
Attendance had a significant effect on weight change outcomes for both treatment groups
(Fig. 2a), with those attending 12 or more sessions faring better in terms of weight loss than
those attending fewer than 12 sessions (p<0.01, Cohen’s d=0.62 for weight loss only and
p<0.001, Cohen’s d=0.79 for combined at 6 months; p=0.17, Cohen’s d=0.29 for weight loss
only and p<0.05, Cohen’s d=0.50 for combined at 12 months). Regardless of treatment
group, the weight change trajectory during the study for those who attended 12 or more
sessions was statistically significant [time trend F(2, 91)=11.18, p<0.0001], whereas the
weight change trajectory for those attending fewer than 12 sessions was marginally
significant [time trend F(2,106)=2.40, p=0.095].
As presented in Fig. 2b, attendance did not have a significant effect on depressive symptom
outcomes, as assessed by the SCL-20, for either treatment group (p=0.72 for weight loss
only and p=0.47 for combined at 6 months, p=0.98 for weight loss only and p=0.37 for
combined at 12 months, all Cohen’s d<0.21). That is, regardless of attendance status, all
participants showed comparable and statistically significant changes in depressive symptoms
over time [time trend F(2, 82)=36.64 for 12+ sessions, time trend F(2, 57)=20.20 for <12
sessions, p values<0.0001].
Changes in Dietary Intake and Physical Activity over Time
Patterns of dietary and physical activity changes in repeated measures analyses (depicted in
Fig. 3a, b) were statistically comparable across treatment groups between baseline and 12
months. Women decreased their dietary intake and increased their physical activity
Linde et al. Page 8
Ann Behav Med. Author manuscript; available in PMC 2012 February 1.
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript
significantly over time in both groups [dietary intake time trend F(2, 103)=29.77, physical
activity time trend F(2, 128)=10.88, p values<0.0001]. With the exception of lower dietary
intake for weight loss only versus combined treatment participants at 6 months among those
who attended at least 12 treatment sessions (p=0.03, data not shown), attendance did not
affect dietary or physical activity patterns during the weight loss trial.
Discussion
This study examined the effects of group behavioral weight loss treatment versus a program
of weight loss treatment combined with cognitive-behavioral therapy for depression for
obese women with major depressive disorder. Results for both outcomes were comparable
between the two intervention conditions. Study findings are indicative of modest reductions
in weight and depressive symptoms for depressed women in either treatment. Although the
changes were not of expected magnitude, these findings are encouraging in light of a prior
population study that suggested poor health and treatment responses among those with
significant obesity-depression comorbidity: the authors reported increased risk of suicidal
ideation among those with comorbid obesity and major depression and suggested that this
finding might have implications for obesity treatment, as those with depression tend to
respond poorly to treatment and with reduced compliance [1]. Here, we found that obese,
depressed women were able to lose weight and improve mood, whether or not they received
specialized depression treatment as part of weight loss treatment. Weight losses among the
depressed women in the present study were equal to or greater than those observed on
average at 12 months (-2.28 kg) in a sample of nondepressed managed care organization
members of similar age and BMI (mean age=51, mean BMI=34 kg/m2) who were
randomized to either a telephone- or mail-based weight loss curriculum developed by
members of the current study team [41].
Women in both treatment groups responded comparably in terms of study engagement (i.e.,
attendance at groups and follow-up interview visits over 12 months) and in terms of
experiencing reductions in depressive symptoms over the course of the study, as well as
comparable changes in weight over time. We do not exclude the possibility that depressive
symptoms remitted spontaneously over time either during the waiting period before groups
started or during the course of weight loss only treatment [42], but we are optimistic that this
study, as with other promising trials such as the ongoing study of whether a pre-weight loss
course of behavioral therapy for depression will improve weight loss outcomes in depressed
obese women [15], will advance this field so that women with these comorbid conditions are
better able to receive appropriate care.
Despite concerted efforts by study team experts to create a unique depression and weight
loss management intervention (versus weight loss only) by augmenting basic nutrition and
physical activity advice with additional cognitive-behavioral depression management
strategies, commonalities between the two intervention programs may have led to
comparable treatment outcomes for the women enrolled in this study. For example, though
foci and behavioral targets (strictly weight loss versus a combination of weight loss and
depression management) differed between intervention conditions, both programs involved
behavioral activation, problem solving skills, cognitive restructuring, and social support.
The extensive overlap between programs may have contributed to our inability to detect
differences between group outcomes at 12 months as estimated by a priori power
calculations for this study; however, the confidence intervals around the results presented in
Table 3 for both outcomes include the expected differences from our initial power
calculations. As we did not make any predictions regarding which program might provide
greater benefit to participants, future studies may consider examining preferences for
treatment focus (e.g., weight loss alone versus combined treatment or treating depression
Linde et al. Page 9
Ann Behav Med. Author manuscript; available in PMC 2012 February 1.
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript
alone), as well as examination of results with other demographic groups (e.g., men or
younger adults).
Study limitations include restrictions on demographic variability of the sample, particularly
with respect to age, race/ethnicity, and gender. Although we were able to account for 13–
14% of missing weight data at 6 and 12 months by using proximally measured chart weights
from medical records, we relied on last observation carried forward as an imputation method
for 8% of cases at 6 months and 12% of cases at 12 months; the last observation carried
forward method of imputation is limited by the assumption that weight remained unchanged
during the time period of the study and may represent an overly generous interpretation of
the data.
As depressive symptom data (assessed via SCL-20) were collected rather than using
repeated SCID administrations at study follow-up points, we are limited to commenting on
symptom changes over the course of the study rather than on changes in DSM-IV diagnostic
status. The mean lag time of approximately 90 days between baseline measurement of
depression status (by SCL-20 and SCID) and the start of intervention groups, without
ascertainment of depression status at the outset of treatment, does not allow us to account for
potential remission of symptoms during the period of time prior to treatment start date. Due
to higher rates of female recruitment in weight loss programs in general, much more is
known about women and weight loss than about men and weight loss, and we are not able to
generalize the present findings to men. Another limitation is that approximately three
quarters of the participants were receiving antidepressant medication, which may have
hindered our ability to detect differential changes in depression levels. In addition, because
combined sessions were 30 min longer than weight loss only sessions (to accommodate
added depression content and to extend homework review of additional self-monitoring
tasks in the combined treatment), participants in the combined intervention were potentially
exposed to an average of 5.25 more treatment hours than those in weight loss only (based on
mean session attendance of 10.5 sessions for the combined group); the added contact may
have affected outcomes for combined participants relative to those in weight loss only.
As in the companion study of the effects of major depression on behavioral weight loss
treatment among obese women [37], one of the most significant limitations of this study was
that rates of treatment participation were fairly low (one quarter of women assigned to
treatment groups failed to attend a single session, and only half attended 12 or more of 26
sessions). The study design, in which women were screened for treatment eligibility
following participation in an epidemiological survey, resulted in lag time between group
assignment and group start dates, to allow for treatment groups of optimal size to be
constructed at several study clinic locations. This procedure may have led to attrition prior to
initial group sessions, either due to wait times or to failure to place all participants in groups
at optimally convenient clinic locations.
Given the association between attendance and weight loss observed here, either program
might have yielded more significant weight loss results had participant attendance been
more robust. However, it is equally likely that greater weight loss success promoted better
attendance over the course of the study; the correlational nature of our data limits our ability
to draw causal inferences with regard to our findings. It is also possible that low attendance
rates across both conditions may have masked differences between the two programs, given
that relatively few participants attended the number of sessions considered to be an optimal
dose of treatment. However, it is encouraging that attendance did not differ by treatment
condition, suggesting that program content (addressing depressive symptoms and weight
loss versus weight loss only) did not contribute to differential attendance rates among these
depressed women.
Linde et al. Page 10
Ann Behav Med. Author manuscript; available in PMC 2012 February 1.
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript
Results for depression symptom change outcomes were not affected by attendance, which is
consistent with observations of symptom remission during cognitive-behavioral therapy for
depression in a dosage range of 8–12 sessions [43] nor was depression (assessed by PHQ-9
administered at pre-treatment study screening for randomization purposes and by SCL-20
administered at treatment study baseline and 6 and 12-month interviews to assess symptom
severity during treatment) correlated significantly with attendance (r=-0.00 to -0.11,
p=0.15 to 0.99). Depression, as assessed by the SCID prior to the start of treatment, was
modestly correlated with total number of sessions attended (r=-0.14, p<0.05); however, this
result appears to have been driven largely by lower attendance at the final maintenance
session, rather than at active treatment sessions, as the correlations between baseline SCID
diagnoses and attendance rates at 6 months (at which time 20 of 26 sessions had been held)
or 12 months (at which time 25 of 26 sessions had been held) were nonsignificant (r values=
-0.08 and -0.09, p values=0.29 and 0.23, respectively).
The current study has several important strengths, including examination of an understudied
but prevalent clinical population, a relatively large sample, rigorous assessment of
depression symptoms prior to study entry, evaluation of a novel intervention, and delivery of
an intensive behavioral program to all participants. As all participants were invited to the
intervention study after a larger-scale screening trial [16], the recruitment success of the
current project has positive implications for public health agency or other outreach strategies
to address obesity and depression in the community.
Overall, the present results suggest that obese women with depressive symptoms are able to
adjust their dietary intake and physical activity frequency to achieve weight loss, regardless
of whether depression is targeted specifically during intervention. Most importantly,
comparable adherence to treatment recommendations and significant improvements in
depressive mood may be achieved in programs that address weight loss only, as well as
those that address both depressed mood and weight control. Future weight loss trials are
encouraged to consider the option of enrolling women regardless of depression status to
provide a more valid evaluation of program effectiveness.
Acknowledgments
This study was funded by National Institutes of Health Grant R01MH068127 (G. E. Simon, PI); ClinicalTrials.gov
identifier: NCT00169273, “Epidemiology and Care of Comorbid Obesity and Depression.”
Appendix
Appendix
Session topics: weight loss only and combined weight loss/depression group visits
Session Weight loss only (90-min sessions) Combined weight loss/depression (120-min sessions)
1 Orientation: group norms, introduction to selfmonitoring, behavior change processes
Orientation: group norms, introduction to selfmonitoring, behavior change processes, relation
between depression and weight
2 Energy balance and healthy food choices Energy balance and healthy food choices
3 Diet quality: Fat and cholesterol Pleasant activities and depression I
4 High fiber, low fat eating Pleasant activities and depression II
5 Increasing physical activity Increasing physical activity
6 Lifestyle exercise Lifestyle exercise, barriers to exercise
7 Barriers to exercise Relaxation training
8 Eating patterns Problem solving I
Linde et al. Page 11
Ann Behav Med. Author manuscript; available in PMC 2012 February 1.
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript
Session Weight loss only (90-min sessions) Combined weight loss/depression (120-min sessions)
9 Eating in social situations Problem solving II
10 Eating in restaurants Eating patterns, eating in social situations, eating in
restaurants
11 Re-evaluating diet and exercise goals I Re-evaluating diet and exercise goals I
12 Re-evaluating diet and exercise goals II Re-evaluating diet and exercise goals II
13 Stress and eating Cognitive goals: monitoring thinking
14 Fad diets, weight loss medications, surgery for
weight loss
Cognitive techniques: what’s the evidence?
15 Cues for eating and exercise Cognitive techniques: thought balancing
16 Advanced diet change Advanced diet change
17 Advanced exercise change Advanced exercise change
18 Assertion and eating Social skills and assertiveness I
19 High risk situations Social skills and assertiveness II
20 Managing slips and lapses Managing slips and lapses
21 Summing up Summing up
22 Long-term self-care plan Long-term self-care plan
23 Monthly check-in: review and reinforcement (diet
and exercise; goals and motivation; plans for next
month), ad hoc session content determined by
group members
Monthly check-in: review and reinforcement (diet,
exercise,pleasant activities, and thought balancing;
goals and motivation; plans for next month), ad hoc
session content determined by group members
24–26 Same as session 23 Same as session 23
For both interventions, sessions 1–16 were held weekly, sessions 17–22 were held biweekly, and sessions 23–26 were held
monthly. Topics that are unique to the combined weight loss/depression program are highlighted in bold text
References
1. Carpenter KM, Hasin DS, Allison DB, Faith MS. Relationships between obesity and DSM-IV major
depressive disorder, suicidal ideation, and suicide attempts: Results from a general population
survey. Am J Public Health 2000;90:251–257. [PubMed: 10667187]
2. Simon GE, Von Korff M, Saunders K, et al. Association between obesity and psychiatric disorders
in the US adult population. Arch Gen Psychiat 2006;63:824–830. [PubMed: 16818872]
3. Thande NK, Hurstak EE, Sciacca RE, Giardina EGV. Management of obesity: A challenge for
medical training and practice. Obes 2008;17:107–113.
4. Goldstein LT, Goldsmith SJ, Anger K, Leon AC. Psychiatric symptoms in clients presenting for
commercial weight reduction treatment. Int J Eat Disord 1996;20:191–197. [PubMed: 8863072]
5. Clark MM, Niaura R, King TK, Pera V. Depression, smoking, activity level, and health status:
Pretreatment predictors of attrition in obesity treatment. Addict Behav 1996;21:509–313. [PubMed:
8830908]
6. Linde JA, Jeffery RW, Levy RL, et al. Binge eating disorder, weight control self-efficacy, and
depression in overweight men and women. Int J Obes Relat Metab Disord 2004;28:418–425.
[PubMed: 14724662]
7. McGuire MT, Wing RR, Klem ML, Lang W, Hill JO. What predicts weight regain in a group of
successful weight losers? J Consult Clin Psychol 1999;67:177–185. [PubMed: 10224727]
8. Anderson RE, Wadden TA, Bartlett SJ, Zemel B, Verde TJ, Franckowiak SC. Effects of lifestyle
activity vs. structured aerobic exercise in obese women: A randomized trial. JAMA 1999;281:335–
340. [PubMed: 9929086]
9. Wirth A, Krause J. Long-term weight loss with sibutramine: A randomized controlled trial. JAMA
2001;286:1331–1339. [PubMed: 11560538]
Linde et al. Page 12
Ann Behav Med. Author manuscript; available in PMC 2012 February 1.
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript
10. Barlow DH, Allen LB, Choate ML. Toward a unified treatment for emotional disorders. Behav
Therapy 2004;35:205–230.
11. Cooper Z, Fairburn CG. A new cognitive behavioural approach to the treatment of obesity. Beh
Res Therapy 2001;39:499–511.
12. Craigle MA, Nathan P. A nonrandomized effectiveness comparison of broad-spectrum group CBT
to individual CBT for depressed outpatients in a community mental health setting. Behav Therapy
2009;40:302–314.
13. Wing, RR. Behavioral approaches to the treatment of obesity. In: Bray, GA.; Bouchard, C.; James,
WPT., editors. Handbook of Obesity. New York: Marcel Dekker; 1998. p. 855-878.
14. Wing RR, Jeffery RW. Benefits of recruiting participants with friends and increasing social
support for weight loss and maintenance. J Consult Clin Psychol 1999;67:132–138. [PubMed:
10028217]
15. Schneider KL, Bodenlos JS, Ma Y, et al. Design and methods for a randomized clinical trial
treating comorbid obesity and major depressive disorder. BMC Psychiatr 2008;8:77.
16. Simon GE, Ludman EJ, Linde JA, et al. Association between obesity and depression in middleaged women. Gen Hosp Psychiat 2008;30:32–39.
17. Spitzer R, Kroenke K, Williams J. Validation and utility of a self-report version of PRIME-MD:
The PHQ primary care study. JAMA 1999;282:1737–1744. [PubMed: 10568646]
18. American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders. 4th ed..
Washington, DC: American Psychiatric Association; 1994.
19. First, M.; Spitzer, R.; Gibbon, M.; Williams, J. Structured Clinical Interview for DSM-IV Axis I
Disorders (SCID-I), Clinician Version. Washington: American Psychiatric Press; 1997.
20. Shear MK, Greeno C, Kang J, et al. Diagnosis of nonpsychotic patients in community clinics. Am
J Psychiatr 2000;157:581–587. [PubMed: 10739417]
21. Derogatis, L.; Rickels, K.; Uhlenhuth, EH.; Covi, L. The Hopkins Symptom Checklist: A measure
of primary symptom dimensions. In: Pichot, P., editor. Psychological Measurements in
Psychopharmacology: Problems in Psychopharmacology. Basel, Switzerland: Kargerman; 1974. p.
79-110.
22. Mulrow CD, Williams JW Jr, Gerety MB, Ramirez G, Montiel OM, Kerber C. Case-finding
instruments for depression in primary care settings. Ann Intern Med 1995;122:913–921. [PubMed:
7755226]
23. Katon W, Robinson P, Von Korff M, et al. A multifaceted intervention to improve treatment of
depression in primary care. Arch Gen Psychiatr 1996;53:924–932. [PubMed: 8857869]
24. Simon GE, Von Korff M, Rutter C, Wagner E. A randomized trial of monitoring, feedback, and
management of care by telephone to improve depression treatment in primary care. BMJ
2000;320:550–554. [PubMed: 10688563]
25. Paffenbarger R, Wing A, Hyde R. Physical activity as an index of heart attack risk in college
alumni. Am J Epidemiol 1978;108:161–175. [PubMed: 707484]
26. Ainsworth BE, Haskell WL, Whitt MC, et al. Compendium of Physical Activities: An update of
activity codes and MET intensities. Med Sci Sport Exer 2000;32 Suppl:S498–S516.
27. Harris JK, French SA, Jeffery RW, McGovern PG, Wing RR. Dietary and physical activity
correlates of long-term weight loss. Obes Res 1994;2:307–313. [PubMed: 16358394]
28. Schakel SF. Procedures for estimating nutrient values for food composition databases. J Food
Comp Anal 1997;10:102–114.
29. Schakel SF. Maintaining a nutrient database in a changing marketplace: Keeping pace with
changing food products—A research perspective. J Food Comp Anal 2001;14:315–322.
30. Efron B. Forcing a sequential experiment to be balanced. Biometrika 1971;58:403–417.
31. Pocock SJ, Simon R. Sequential treatment assignment with balancing for prognostic factors in the
controlled clinical trial. Biometrics 1975;31:103–115. [PubMed: 1100130]
32. Jeffery RW, Wing RR, Thorson C, et al. Strengthening behavioral interventions for weight loss: A
randomized trial of food provision and monetary incentives. J Consult Clin Psychol
1993;61:1038–1045. [PubMed: 8113481]
Linde et al. Page 13
Ann Behav Med. Author manuscript; available in PMC 2012 February 1.
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript
33. Jeffery RW, Wing RR, Thorson C, Burton LR. Use of personal trainers and financial incentives to
increase exercise in a behavioral weight-loss program. J Consult Clin Psychol 1998;66:777–783.
[PubMed: 9803696]
34. Jeffery RW, Wing RR, Sherwood NE, Tate DF. Physical activity and weight loss: Does prescribing
higher physical activity goals improve outcome? Am J Clin Nutr 2003;78:684–689. [PubMed:
14522725]
35. Wing RR, Jeffery RW, Burton LR, Thorson C, Nissinoff KS, Baxter JE. Food provision vs.
structured meal plans in the behavioral treatment of obesity. Int J Obes 1996;20:56–62.
36. Brown R, Lewinsohn P. A psychoeducational approach to the treatment of depression: Comparison
of group, individual, and minimal contact procedures. J Consult Clin Psychol 1984;52:774–783.
[PubMed: 6501663]
37. Ludman EJ, Simon GE, Ichikawa L, et al. Does depression reduce the effectiveness of behavioral
weight loss treatment? Behav Med 2009;35:126–134. [PubMed: 19933059]
38. Cohen, J. Statistical Power Analysis for the Behavioral Sciences. 2nd ed.. Hillsdale NJ: Lawrence
Erlbaum Associates; 1988.
39. Arterburn D, Ichikawa L, Ludman EJ, et al. Validity of clinical body weight measures as
substitutes for missing data in a randomized trial. Obes Res Clin Pract 2008;2:277–281. [PubMed:
19956347]
40. Paulose-Ram R, Safran MA, Jonas BS, Gu Q, Orwig D. Trends in psychotropic medication use
among U.S. adults. Pharmacoepidemiol Drug Saf 2007;16:560–570. [PubMed: 17286304]
41. Jeffery RW, Sherwood NE, Brelje K, et al. Mail and phone interventions for weight loss in a
managed-care setting: Weigh-to-be one-year outcomes. Int J Obes 2003;27:1584–1592.
42. Andrews G. Placebo response in depression: Bane of research, boon to therapy. Br J Psychiatr
2001;178:192–194.
43. Butler AC, Beck AT. Cognitive therapy for depression. Clin Psychol 1995;48:3–5.
Linde et al. Page 14
Ann Behav Med. Author manuscript; available in PMC 2012 February 1.
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript
Fig. 1.
Study recruitment flow diagram
Linde et al. Page 15
Ann Behav Med. Author manuscript; available in PMC 2012 February 1.
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript
Fig. 2.
a Means and standard errors of weight change (in kilograms) over time by treatment session
attendance. b Means and standard errors of depressive symptom change over time by
treatment session attendance
Linde et al. Page 16
Ann Behav Med. Author manuscript; available in PMC 2012 February 1.
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript
Fig. 3.
a Means and standard errors of dietary intake changes by treatment group. b Means and
standard errors of physical activity changes by treatment group
Linde et al. Page 17
Ann Behav Med. Author manuscript; available in PMC 2012 February 1.
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript
Linde et al. Page 18
Table 1
Sources of weight data at 6 and 12 months
Weight loss only (n=102) Combined (n=101)
6 months
n (%)
12 months 6 months 12 months
Measured weight 73 (71) 73 (71) 87 (86) 77 (76)
Chart weight 20 (20) 13 (13) 6 (6) 15 (15)
Last study observation carried forward 9 (9) 16 (16) 8 (8) 9 (9)
Ann Behav Med. Author manuscript; available in PMC 2012 February 1.
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript
Linde et al. Page 19
Table 2
Baseline characteristics by random assignment to treatment group
Weight loss only (n=102) Combined (n=101)
Mean or percent SD Mean or percent SD
Age 52.1 6.5 52.3 6.2
Weight (kilograms) 106.6 22.3 103.4 17.7
Body mass index (kilograms/m2) 39.5 7.7 38.6 6.8
Any antidepressant medication 75.4% 78.2%
PHQ depression score from epidemiological survey (prior to treatment
randomization)
15.8 3.3 15.7 3.5
SCL-20 depression score at intervention study baseline (prior to start of treatment) 1.82 0.59 1.87 0.62
Percent married 45% 54%
Percent white 78% 79%
Percent with high school education or greater 86% 89%
Dietary intake (kilocalories/day)a 2,216 969 2,270 879
Physical activity (kilocalories/week) 506 772 549 789
There were no statistically significant differences in baseline characteristics between treatment groups
PHQ Patient Health Questionnaire, SCL-20 Symptom Checklist-20
a
Due to missing data, n=88 in weight loss only, 88 in combined
Ann Behav Med. Author manuscript; available in PMC 2012 February 1.
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript
Linde et al. Page 20
Table 3
Weight and depressive symptom changes at 6 and 12 months by treatment group
Weight loss only (n=102) Combined (n=101) p d
Mean SD 95% CI Mean SD 95% CI
Mean weight change from baseline (kilograms)
6 months -2.8 6.6 -4.1, -1.6 -1.8 6.6 -3.1, -0.5 0.26 0.16
12 months -3.1 8.9 -4.8, -1.3 -2.3 8.9 -4.1, -0.6 0.55 0.09
Mean depressive symptom change from baseline (SCL-20)
6 months -0.54 0.71 -0.68, -0.40 -0.50 0.70 -0.64, -0.36 0.70 0.06
12 months -0.53 0.81 -0.68, -0.37 -0.65 0.80 -0.81, -0.50 0.25 0.16
SCL-20 Symptom Checklist-20
Ann Behav Med. Author manuscript; available in PMC 2012 February 1.