Obesity and sedentary lifestyles have reached epidemic proportions in the United States, resulting in increased rates of a variety of chronic diseases, increased risk of mortality, and substantial health care costs (
1). Adults with schizophrenia and other serious mental illnesses are more likely to be overweight or obese, which contributes to the higher rates of comorbid general medical disease and early mortality in this population (
2,
3). The association between serious mental illness and obesity may result from poor dietary habits and sedentary lifestyle (
4). Medication side effects, such as metabolic alterations associated with the use of second-generation antipsychotic agents, also contribute to the problem (
5,
6).
A recent systematic review of the literature identified 23 studies designed to help individuals with serious mental illness lose weight and become more physically active (
7). Interventions included both individual and group sessions. All provided general information about diet and exercise, and many included instruction in how to read food labels, count calories, practice portion control, and develop healthier diets. Some, but not all, interventions also incorporated exercise or fitness training elements. Most of the interventions incorporated behavioral strategies, including goal setting and behavioral self-management. Finally, all interventions had modifications or enhancements to address the range of motivational and cognitive deficits common among individuals with serious mental illness.
Of the 18 studies that conducted statistical analyses, ten reported significant weight loss findings. Among the seven randomized controlled trials (
8–
14), three of which did not find statistically significant weight loss outcomes (
8–
10), the mean±SD weight loss was 3.7±2.3 pounds (
7). By comparison, the mean weight loss from lifestyle interventions tested in the general population, according to recent meta-analyses, was between eight and 11 pounds (
15,
16). Collectively this research indicated that lifestyle interventions for individuals with serious mental illness can achieve modest weight loss and clinically meaningful reductions in cardiovascular risk factors. However, the challenges are considerable, and some studies failed to show benefit (
7). Notably, it is unknown which elements of the programs are critical to helping various participant groups.
A recent nonrandomized trial evaluated a psychosocial weight management program for veterans with schizophrenia in the context of a larger longitudinal Veterans Affairs (VA) evaluation of a chronic care model for schizophrenia (
17). The program, a 16-session group intervention, was based on the Lilly Solutions for Wellness program (
18). Weight loss did not differ among participants and nonparticipants (2.4±10.6 and 2.2±11.9 pounds, respectively) (
19). In 2006, the Veterans Health Administration (VHA) introduced the MOVE! Weight Management Program for Veterans to provide comprehensive, multidisciplinary weight management services for the nearly six million veterans served throughout its system of care (
20). The core components of MOVE! include delivery of behaviorally based dietary and physical activity self-management through a variety of modalities, including individual and group face-to-face counseling and telephone care. Weight loss medication, intensive outpatient treatment, residential treatment, and bariatric surgery may also be available as adjuncts to the core behavioral program.
Recent data indicated that 18.6% of veterans who participated in at least two MOVE! visits experienced a 5% loss of body weight by six months (
20,
21). The cohort of the studies was 31,854 veterans across the VHA with available weight measurements, and the data were compiled from a variety of sources, including administrative and clinical databases, electronic medical record reviews, and annual structured VHA facility self-reports. Of note, 12.5% of veterans in a comparison group matched for age, gender, body mass index (BMI), and comorbidity status who were not treated with MOVE! also had at least a 5% loss of body weight. Among those who participated in more intense and sustained treatment (eight or more MOVE! visits and continuous engagement in MOVE! for at least four months), 31.6% achieved clinically relevant weight loss; only 13.4% of the cohort, however, received intense and sustained treatment.
MOVE! has not yet been tested in specialty mental health settings to facilitate weight management among veterans with serious mental illness. Using content developed for MOVE!, we created a manualized six-month weight management intervention for individuals with serious mental illness that included both individual and group sessions. This article describes the modifications that we used to manualize MOVE! for veterans with serious mental illness and reports findings from a randomized controlled trial comparing MOVE! with a control condition that offered basic information about diet and exercise at monthly weigh-ins. We hypothesized that over a six-month period, the number of veterans who achieved weight loss of 5% or more would be greater among veterans randomly assigned to MOVE! versus the control condition. We also hypothesized that those assigned to MOVE! would experience greater reduction in individual metabolic syndrome criteria (hypertension, increased waist circumference, dyslipidemia, and hyperglycemia) as well as diagnosis of metabolic syndrome (three or more individual criteria). Secondary hypotheses focused on differential outcomes between study conditions in physical activity and dietary management as well as a range of attitudinal and functional variables.
Methods
Study setting and sample
Veterans were recruited from outpatient mental health clinics within the VA Maryland Health Care System and the District of Columbia VA. Inclusion criteria were DSM-IV chart diagnosis of schizophrenia, other psychotic spectrum disorder, bipolar disorder, major depression, or severe anxiety disorder; age 18 to 75; community dwelling; BMI ≥25; and English speaking. Exclusion criteria were documented debilitating general medical diagnoses, including serious renal, cardiac, neurological, and respiratory diseases; prior or planned bariatric surgery; use in past three months of prescription pharmacological agents for weight loss; and abnormal results of blood tests—including hemoglobin A1c >9 mg, fasting blood sugar >160, and values outside a normal range for hematocrit, creatinine, and liver enzymes within the past six months—that indicate a serious medical condition that would contraindicate participation in a weight loss program. All study participants gave written informed consent, and the study was approved by all relevant institutional review boards.
Procedures
A total of 1,223 veterans were screened for eligibility. Of the 329 who met initial eligibility criteria, 227 were approached and asked to participate and 147 (65%) consented. The main reasons for ineligibility included diagnosis not listed as an inclusion criterion, disengagement from mental health treatment services, and presence of exclusionary general medical conditions. All consents occurred between January 2007 and June 2009. Subsequently, 38 veterans were excluded after a review of blood work collected as part of screening procedures and responses to the Physical Activity Readiness Questionnaire (
22) indicated a self-reported medical condition or prompted the primary medical provider to withdraw preapproval for the veteran to participate.
A total of 109 veterans completed the baseline assessment and were randomly assigned to the MOVE! (N=53) or control condition (N=56) by using a permuted-block randomization scheme with varying block sizes stratified by antipsychotic medication categories. Antipsychotic medication categories were defined as being prescribed at baseline either olanzapine or clozapine or any other medication regimen. All study participants in both conditions were asked to return monthly for a weigh-in and again at the end of six months to complete a battery of follow-up assessments.
Interventions
MOVE!
In adapting MOVE! for individuals with serious mental illness, we strove to include strategies associated with successful weight loss programs targeting individuals with serious mental illness, including psychoeducation focusing on nutritional counseling, caloric expenditure, and portion control (
23). We also emphasized behavioral and motivational self-management strategies, including weekly goal setting and regular weigh-ins.
Table 1 presents a summary of the key components of the intervention and reviews the enhancements used to optimize MOVE! for use with veterans with serious mental illness.
Research staff who had previous experience leading psychosocial and behavioral interventions with seriously mentally ill adults delivered the intervention. Participants completed individual sessions once a week for the first month, followed by weekly group sessions of about 60 minutes during months 2–4. Months 5 and 6 included four biweekly sessions and two individual sessions to further reinforce weight loss efforts and review important information covered in the curriculum. We did not incorporate supervised exercise as part of the intervention because veterans did not have access to exercise facilities at the study sites and they were recruited from across multiple programs.
Control condition.
Participants assigned to the control condition received standard services in the context of their existing treatment plans. In addition, once a month they met with research staff and were weighed and provided with brochures and handouts about diet and exercise. Notably, at the time of the study trial, MOVE! was not widely available at the facilities where the study was conducted, nor were programmatic efforts or procedures in place to screen and refer veterans with serious mental illness for a service related to MOVE!
Assessments and measures
Participants completed a 1.5-hour assessment administered at baseline and again six months later. Weight measurements were recorded for all participants at both time points and monthly during the intervening months. The subset of participants recruited from the VA Maryland Health Care System (N=71) also had their blood pressure measured at baseline and at the six-month follow-up and provided a blood sample after a ten-hour, overnight fast to measure total cholesterol, triglycerides, low-density and high-density lipoprotein cholesterol, and glucose.
All participants completed a demographic questionnaire and the MOVE!23, a 23-item questionnaire developed by the VHA National Center for Health Promotion and Disease Prevention about general medical health, eating habits, exercise patterns, and barriers that may get in the way of weight management efforts (
24). The study assessment battery also included instruments to measure dietary habits, including the Block Fruit, Vegetable, and Dietary Fat Screeners (
25,
26); attitudinal and motivational measures, including the Diet and Exercise Confidence Survey (
27); and the Impact of Weight on Quality of Life Survey (
28). The 12-Item Short Form (SF-12), a self-reported rating of general medical health, mental health, and physical functioning, was also included (
29).
Statistical analyses
Baseline characteristics were compared across conditions by using independent t and chi square tests. Baseline values for participants who completed the six-month assessment and those who were lost to follow-up were also compared.
Monthly weight data were analyzed both as a binary outcome, indicating whether the participant had lost 5% of his or her body weight, and as a continuous weight measure. All participants were included in the analyses regardless of the number of assessment points or intervention sessions they completed (intent to treat). A generalized estimating equation model (SAS, version 9.2, Proc Genmod) was used to compare binary outcomes between groups over time, with month, treatment group, and month × treatment group as independent variables. For continuous outcomes, a general linear mixed model (Proc Mixed) was used. An unstructured covariance matrix was specified for all repeated-measures analyses.
Changes in measures of metabolic or laboratory, dietary management, physical activity, and attitudinal variables obtained at baseline and six-month follow-up were analyzed by using the same models described above.
Results
Seventy-three percent (N=41) of the control group and 57% (N=30) of participants in the MOVE! condition completed the six-month assessment. Those who did not complete the follow-up composite assessment (N=38, 35%) were younger (t=2.56, df=56, p=.013) and reported poorer functioning on the mental health subscale of the SF-12 (t=2.89, df=107, p=.005) than the completers (N=71, 65%).
Sample characteristics
Baseline characteristics for the 109 veterans randomly assigned to the two conditions are presented in
Table 2. The sample was predominantly African American and had a high burden of medical comorbidity, current tobacco use, and alcohol use in the past 30 days. Twenty-five percent of the sample resided in a supervised community residence; approximately half of the sample also reported being dependent on others for grocery shopping (47%) and preparation of meals (52%). There were no significant differences between participants in study conditions across any of the variables listed except that a greater proportion of persons in the control condition (36%) than in MOVE! (19%) reported a diagnosis of diabetes (p=.049).
Intervention fidelity
A total of 38 group sessions were video recorded, and the intervention was rated by using a fidelity checklist that focused on both content—for example, coverage of essential nutritional and physical activity information—and process-oriented aspects—such as facilitator attention to barriers or obstacles voiced by participants. Items were rated on a scale from 1, not at all evident, to 4, fully evident. The mean score on all items across the 38 videos was above 3.7.
Intervention completion rates
Among the 53 veterans assigned to MOVE!, 70% (N=37) completed all four individual sessions, 15% (N=8) completed two or three sessions, and 15% (N=8) completed fewer than two sessions (mean=3.3 sessions). During months 2–6, a total of 34% (N=18) attended 75% or more of the 16 group sessions; 15% (N=8) completed between 50% and 75% of the sessions, 15% (N=8) completed between 25% and 50%, and 36% (N=19) completed fewer than 25%. The mean number of group sessions completed was 7.4, and the median was five.
Responses to questions on the MOVE!23 (
24) indicated a nonsignificant change from baseline to follow-up in the number of participants in the control group who reported lifetime receipt of weight loss services (z=.84, p=.399), suggesting that they did not receive additional weight loss programming during the course of the trial.
Monthly weight and metabolic outcomes
A total of seven participants (6% of the full sample) lost 5% of their baseline weight over the six-month study period. Neither a main effect of month or month × treatment effect was significant. When continuous weights were analyzed, no significant effects were found for month or month × treatment effects. We also compared all weight outcomes between participants in the MOVE! condition who completed eight or more group sessions (N=26) and those who completed seven or fewer (N=27); there were no significant differences in weight outcomes in relation to dose of the intervention.
Further, there were no significant differences over time in any of the variables used to classify metabolic syndrome between the 30 veterans in MOVE! and the 41 veterans in the control condition for whom we had complete laboratory data (
Table 3).
Additional follow-up outcomes
There were no significant group × time differences across the full range of dietary management, physical activity, attitudinal, and functional measures. All baseline and follow-up results are presented in
Table 3.
Discussion
This study did not find any significant differences in weight loss or related metabolic outcomes between persons receiving MOVE! and persons receiving monthly weight monitoring and brochures and handouts related to diet and exercise. Further, there were no significant differences between the study and the control groups across any of the dietary, physical activity, and attitudinal measures. These results underscored the significant challenges associated with helping seriously mentally ill adults adopt and sustain healthier lifestyle behaviors.
Consideration of the baseline characteristics of the study sample highlights the range and magnitude of these challenges. Participants were eating foods extremely high in fat and smoking at rates greater than the general population. At baseline, the participants’ mean weight was 236 pounds and their mean BMI was 34.6, both in the obese range. More than one-third met criteria for metabolic syndrome, and most were coping with one or more co-occurring chronic general medical conditions. Most were using an antipsychotic or an antidepressant or other mood-stabilizing medications, including 13% who were taking olanzapine, a second-generation antipsychotic associated with weight gain. Given the severity and extent of these co-occurring conditions and lifestyle challenges, future studies might consider programming and services that integrate a range of lifestyle issues and emphasize self-management of general medical health and wellness over a more narrow focus on weight. In light of the motivational deficits common among seriously mentally ill adults, work on how to best engage this population in lifestyle interventions is especially important. We also recommend that more work be done to link weight management programming with other recovery-oriented services designed to support a holistic approach to wellness.
Participants also experienced significant environmental challenges that may have made it harder for them to make meaningful changes in their diet. Specifically, one-quarter were living in supervised living situations, and close to half reported having no control over food shopping or food preparation. These findings suggested a need for added efforts to work directly with facilities to improve food services to support healthier dietary intake among consumers living in residential facilities. Others have suggested that interventions include more in vivo activities, such as shopping and cooking, to help support lifestyle changes. Recent pilot work suggested that modifying meals and snacks served in psychiatric day programs—as part of a larger intervention effort—is associated with positive weight loss results (
30).
The lack of significant findings also warrants critical consideration of the manualized MOVE! intervention developed for this study. Although delivered with fidelity and packaged with a range of evidence-based intervention features, including psychoeducation focused on nutritional counseling, caloric expenditure, and portion control (
23) along with behavioral goal setting and regular weigh-ins, our MOVE! intervention did not include in-session opportunities for supervised physical activity, which have been shown in other studies to be associated with weight loss (
7). The attrition rate among those in the MOVE! condition also raised the question about the optimal amount and intensity of programming needed to effectively engage participants in weight loss interventions. Of note, among the 23 studies summarized in the introduction, interventions ranged in duration from 30 minutes to 52 weeks.
Future studies would benefit from qualitative feedback from participants regarding their experiences with intervention burden and its effects on retention. Future studies should also compare the relative benefits of standardized group-based interventions, such as the version of MOVE! we evaluated, and more individualized approaches that afford greater opportunity to tailor content and focus on an individual’s specific needs, preferences, and challenges. Using mobile or Web-based technology to augment or replace some in-person services should also be considered.
Our intervention also did not include concerted efforts to engage other elements of the MOVE! program, such as coordinated service delivery across multiple disciplines, for example, dietetics, exercise physiology, behavioral health, nurses, and primary and specialty medical care. The VHA has been a national leader in creating service delivery models, such as the Patient Aligned Care Teams (
31), that are designed to provide comprehensive patient-centered care that coordinates primary care with a full range of other care needs. Although our version of MOVE! was not successful, future work is needed to consider how to best select and configure elements of structured weight management interventions to support and complement such integrated program models.
Conclusions
Despite the negative findings of this study, further research is crucial to identify effective lifestyle interventions and related supports and services to reduce overweight and obesity among mentally ill veterans.
Acknowledgments and disclosures
The study was supported by a Research, Rehabilitation, and Development merit grant from the U.S. Department of Veterans Affairs. The authors thank the National Program for Weight Management, National Center for Health Promotion and Disease Prevention, Office of Patient Care Services, Veterans Health Administration.
The authors report no competing interests.