Weight Loss Treatment for Veterans With Binge Eating

RESEARCH DESIGN AND METHODS

Design Overview To address the specific aims of this proposal, the investigators will conduct
an RCT comparing MOVE! (treatment-as-usual; MOVE!) to MOVE! plus gshCBT (MOVE!+gshCBT).
Overweight/obese Veterans (108 men and women) with recurrent binge eating as reported on the
binge eating item from the MOVE! survey will be enrolled. Potential participants will be
recruited from MOVE! group orientation meetings at two VA CT sites which take place several
times per month. Major assessments will be performed by an independent evaluator at
pre-treatment (baseline), at the end of gshCBT treatment (3 months), and at 6- and 12-months
following treatments (9 and 15 months after randomization). Analyses are powered for the
primary outcome of binge remission (categorical). Analyses for binge frequency (continuous),
and eating disorder and mental health outcomes will be performed. Exploratory analyses will
be conducted on treatment engagement (MOVE! session attendance) and weight outcomes (weight
loss and metabolic outcomes), as well as for potential predictors and moderators (age,
gender, clinical severity, and comorbidity), and correlates (change in weight control
strategies and physical activity) of treatment outcome.

Participants Criteria have been kept to a minimum to increase the generalizability of
findings. Inclusion Criteria: aged 21 and older, BMI>=25, and presence of recurrent binge
eating on the MOVE! Survey. Exclusion Criteria: more than 4 MOVE! sessions in the prior year;
active psychosis or suicidal ideation; medical or psychiatric illness, or cognitive deficits
that interfere with providing consent or completing assessments, and; pregnancy or lactation.
The investigators will aim for a sample that is 34% racial/ethnic minority (66% Caucasian) to
be consistent with the demographics of a nationally representative sample of Veterans seeking
weight management treatment through VHA3. The investigators will also aim for 25% women (75%
men), a figure that represents an oversampling of women compared to the rate of 6% for all
VHA users. This will allow us to examine gender as a potential predictor or moderator of
treatment outcome and address the VHA's strategic plan to better serve women Veterans.

Recruitment Potential participants will be recruited from orientation sessions for MOVE!
groups conducted at two sites, the West Haven and Newington campuses, of VA CT. While MOVE!
is delivered in different formats (i.e., group, individual and telehealth), the group format
accounts for 75% of weight management services delivered across the VHA. The Research
Assistant will be given a few minutes at the beginning of each orientation session to
describe the study. Orientation sessions are available to Veterans three times per month
between the two sites. Names and numbers will be collected of those who are interested so
that they may be contacted for individual sessions to assess interest and eligibility using
the criteria above. Dr. Ruser (Co-I, Chief of Primary Care and MOVE! Physician Champion for
VA CT) oversees the program at both campuses, and Dr. Dorflinger (Co-I, Health Behavioral
Coordinator who is directly responsible for the delivery of MOVE!) splits her time between
the two sites. According to MOVE! performance measures, a greater number of patients are seen
for weight management at the Newington campus compared to the West Haven campus. The
investigators plan to test, and control, for potential site differences in the analyses. In
addition, Dr. Driscoll (Co-I), women Veteran's health researcher for the VHA's Women's
Practice Based Research Network, and clinical supervisor for the local women's clinic, will
be responsible for recruiting women Veterans into the MOVE! orientation sessions and
treatment study. She has experience successfully oversampling for women Veterans in a
previous trial that resulted in 25% women (compared to 6% of all VHA users)38.

Screening for "recurrent binge eating" The presence of binge eating will be determined by the
MOVE! Survey, an 11-item self-report survey that must be completed prior to attending MOVE!
programs. This survey was created to aid in the tailoring of MOVE! treatment for patients'
specific needs. The item that assesses the presence of binge eating reads as follows: "On
average, how often have you eaten extremely large amounts of food at one time and felt that
your eating was out of control at that time?" The response set for this item is never, less
than one time per week, one time per week, two to four times per week, or five or more times
per week. Participants reporting binge eating at least one time per week will be included
following published methods. These methods come from two effectiveness studies for recurrent
binge eating that have been conducted with participants from a large integrated healthcare
system18,19. Further assessment with the Eating Disorder Examination-Interview 12th Ed.
(EDE)46, described below, will enable us to further categorize participants who meet full
DSM-5 diagnostic criteria for BED (clinical BED) and those who do not (subclinical BED).

Informed consent and pre-treatment assessment If the potential participant is eligible and
interested, a pre-treatment assessment of one to two sessions will be scheduled. This
assessment will help to determine whether participants are able to comply with the
requirements of the program (such as regular appointments and record-keeping) and also to
minimize the attrition rate of the trial, since only those who complete these requirements
will be enrolled. The Research Assistant will describe the study, and the consent will be
reviewed verbally and in writing, before IRB written informed consent is obtained. During the
pre-treatment assessment session(s), the following will be obtained/completed: anthropometric
measures (height, weight, waist circumference, blood pressure; methods described below),
demographic information (age, gender, race/ethnicity, etc.), self-report measures (described
below), blood draw, psychiatric evaluation with the Mini-Neuropsychiatric Interview (MINI)49,
and the eating disorder evaluation with the Eating Disorder Examination-Interview 12th Ed.
(EDE)46.

Determination of binge status: Subclinical vs. Clinical BED The Eating Disorder
Examination-Interview 12th Edition (EDE)46 is the structured diagnostic interview that is
considered the gold-standard to diagnose, and assess the symptoms of, eating disorders
including binge eating and eating disorder pathology. The EDE will be used to categorize
participants as clinical BED (meeting full DSM-5 criteria including a binge frequency of at
least one time per week on average for the past three months) or subclinical BED (reporting
at least one time per week binge eating on the MOVE! survey but not meeting DSM-5 criteria
for BED). The Research Assistant will be trained and supervised by Dr. Masheb (PI) who has
extensive experience with the EDE. Binge status will be used as a potential covariate in
analyses.

Randomization procedures Eligible patients who agree to participate and sign consent forms
will be randomly assigned to MOVE! or MOVE!+gshCBT in the order in which they complete the
evaluation process. Equal numbers of participants will be randomly assigned to each of the
two conditions. The randomization schedule will be created with a random number generator,
independent of the investigators.

Control condition: MOVE! (treatment-as-usual) Based on scientific review and practice
guidelines of the NIH, and in collaboration with VHA Central Office, the MOVE! program was
developed as an evidence-based weight management and physical activity initiative that
included assessment and treatment procedures, clinical algorithms, patient and provider
information, and instructional materials. The program was widely disseminated and implemented
beginning in January 200653. Every VHA facility across the country is required to have some
form of the MOVE! program, if not several programs differing in intensity, which utilize
these materials. At a minimum, on-site weekly, multidisciplinary groups must be held. MOVE!
group at VA CT is the most widely used weight management service and was thus chosen as the
treatment-as-usual for the present study given the potential for disseminability. Research
has demonstrated minimal weight losses with MOVE! and these findings have largely been
attributed to poor treatment engagement (i.e., attendance) and weight gain in the year prior
to treatment initiation54,55. One initiative to increase the reach and impact of weight
management in the Veteran population has been to identify vulnerable subgroups of
overweight/obese patients, such as those with binge eating8, pain56, and serious mental
illness57. The hope is that the present study will extend this initiative.

All participants will receive MOVE! group treatment as it is delivered at VA CT. The
60-minute groups cycle every 16 weeks and are led by a physical therapist, dietitian, and/or
clinical health psychology predoctoral intern. Clinicians utilize materials available
on-line, and encourage the use of, but do not require or provide, food diaries. Participants
in both MOVE! and MOVE!+gshCBT will be encouraged to attend these sessions throughout their
involvement in the entire study (i.e., during treatment and follow-up phases). MOVE! is
similarly delivered at both the VA CT West Haven and Newington facilities as both programs
are overseen by Drs. Ruser and Dorflinger (Co-Is).

Experimental condition: MOVE!+gshCBT One-half of participants will be randomly assigned to
MOVE!+gshCBT and will receive the MOVE! treatment described above as well as a gshCBT
treatment for binge eating. gshCBT will be administered via 8 individual sessions over a
3-month treatment period. The first two sessions will be 60 minutes each and subsequent
sessions will be 30 minutes. CBT has emerged as the treatment of choice for BED for its
robust effects on reducing binge eating and improving the behavioral and psychological
aspects of the disorder9. The gshCBT treatment will adhere to a manual developed by Dr.
Masheb (PI) and generally follows the book Overcoming Binge Eating 58 that has been used by
Drs. Masheb and Grilo (Co-I) in previous published studies of CBT11,21,32-35, as well as by
others 12,40,59. A copy of the gshCBT manual is included in the Appendix.

Participants will be given the gshCBT manual with a schedule to read at home. During
sessions, the Psychologist will use the following session guideline: 1) setting an agenda, 2)
reviewing food diaries and self-monitoring, 3) introduction of new material from the manual,
4) reviewing old material and problem-solving difficulties, and 5) establishing new homework
and goals. The treatment follows three stages. Stage 1 consists of presentation of the CBT
model; introduction and discussion of the structure, goals, interventions, and expected
outcome of treatment; education regarding binge eating and dieting; and the introduction of
self-monitoring techniques for identifying problematic eating practices and triggers. Stage 2
consists of maintaining the normalized eating and self-monitoring procedures but becomes
increasingly cognitively oriented. Stage 3 focuses on practicing skills, consolidation of
progress and relapse prevention.

At each session participants will receive new, blank food diaries. Participants are expected
to complete these diaries on a daily basis, and they will be reviewed and checked at each
session by the Psychologist. If participants would like to use these food diaries for their
MOVE! treatment as well, they may do so. [NOTE: Food diaries are encouraged in MOVE! but are
not provided or required.] gshCBT participants will also be instructed in how to record
(i.e., self-monitor) binge eating. Recording of this data will be mandatory and participants
who arrive to session with incomplete food diaries or self-monitoring will be asked to use
session time to complete these records for at least the previous day or two. This technique
emphasizes the importance of food diaries, raises patients' awareness of their eating, allows
clinicians to address any concerns about how to complete diaries, and demonstrates that
diaries can be completed without too much effort. The investigators will continue this
strategy through all CBT sessions until the patients arrive to sessions with diaries
completed. This typically takes two sessions, however sometimes food diary non-compliance can
reemerge later in treatment. In this scenario in-session diary completion is reintroduced
following the same procedure. Patients will also be encouraged to gradually increase daily
modest physical activity with the goal of establishing 30 minutes of activity (walking
sessions or other activity) five times per week. This prescription is consistent with
recommendations from several expert-based organizations (NIH and Centers for Disease
Control).

Measurements Participant assessments will be conducted by the Research Assistant and
independent of the treating Psychologist so that assessment will be unbiased. These will
include outcomes for: binge eating and mental health; engagement and weight; weight control
behaviors and physical activity, and; treatment satisfaction and adherence.

Statistical considerations

Analytic Plan. Statistical analyses will be conducted by a biostatistician and who has
extensive experience in statistical modeling, particularly in the analysis of longitudinal
data. Baseline characteristics will be compared between the treatment groups using ANOVA and
chi-squares as appropriate in order to determine the level of success of the randomization.
Distributions of outcomes will be examined in order to determine if assumptions of specific
model types are met. Change of model type as well as transformation of the data will be
considered if model assumptions are not met. Mixed-effects models, a method for statistical
modeling that uses both fixed effects and random effects, will be used to examine continuous
outcomes such as binge eating frequency, mental health symptoms, and percent weight loss for
completer, as well as intent-to-treat (all randomized persons), analyses. Logistic regression
models will be used for categorical outcomes such as binge remission and achieving clinically
meaningful weight loss (i.e., 5% weight loss or greater). Missing data and covariate
adjustment will be handled.

Aim 1 addresses binge eating (primary outcome) and mental health outcomes. Binge remission
(H1; defined as zero binges for the 28 days prior to the end of treatment) will be examined
as a binary outcome in logistic models including treatment group and relevant covariates.
Reduction in frequency of binge eating over time will be examined using mixed effects models.
Previous work indicates that the outcomes may need to be log-transformed to better
approximate normality11. Mental health outcomes (H2), using validated symptom scales for
depression and quality-of-life, will also be examined using mixed models. Relevant
demographic and clinical characteristics will be included in the logistic and mixed models,
and main as well as interaction effects will be investigated.

Aim 2 addresses level of engagement in weight loss treatment (i.e., number of MOVE! treatment
sessions) and weight outcome. Engagement (H1) will be explored using Poisson regression on
counts of MOVE! sessions completed using a binary treatment variable (MOVE! vs.
MOVE!+gshCBT). The distribution of engagement will be reviewed to determine if there is
over-dispersion (in which case quasi-likelihood estimation will be used to account for it).
Models will include relevant covariates such as demographic and clinical characteristics. The
treatment effect on weight (H2) will be examined with a mixed effects model, as described
above, using longitudinal change in percent weight loss during the intervention and follow-up
phases. Categorical analyses will also be performed comparing treatments on the proportion of
participants achieving 5% weight losses using logistic regression analysis.

Exploratory analyses will be performed on the metabolic outcomes (HbA1C, total cholesterol,
HDL, LDL, triglycerides, systolic and diastolic blood pressure, and waist circumference)
using mixed models. These outcomes were chosen for analysis to be consistent with interests
of health services research although the investigators recognize that changes in these
measures may not be sufficient to detect a difference. These analyses also could be important
downstream in adding credibility in the eyes of patients, and providers, and could be used to
enhance incentives for participation (or patient referral).

Aim 3 is an examination of predictors, moderators and correlates of treatment outcome.
Predictors are variables present at baseline that influence outcome regardless of treatment
condition, and moderators are variables present at baseline that influence outcome depending
on the treatment condition. Potential predictors and moderators (H1) will be explored with
demographic (age and gender)81, and clinical characteristics (clinical severity and
comorbidity) from the relevant literature on BED82,83. Logistic and multiple regression
analyses, that have been detailed conceptually and analytically by Kraemer et al.84, will be
used. Associations (H2) between changes in weight control strategies and physical activity,
and weight loss 71 will be examined with correlation analyses.

Inclusion Criteria:

- BMI>=25

- Presence of recurrent binge eating on the MOVE! Survey

Exclusion Criteria:

- More than 4 MOVE! sessions in the prior year

- Active psychosis or suicidal ideation

- Medical or psychiatric illness, or cognitive deficits that interfere with providing
consent or completing assessments

- Pregnancy or lactation