Abstract
Obesity is a common health problem for veterans. This study explored background and program characteristics associated with a 5% weight reduction for veterans enrolled in MOVE!®, a weight management program. For data analysis, 404 veteran records were examined using logistic regression. Background characteristics included socio-demographic variables, comorbidity, body mass index, rurality, and Veterans Administration (VA) priority group. Program characteristics included the program type (group attendee or self-managed) as well as the number and type of provider contacts. Thirteen percent of participants achieved a 5% weight reduction. Age in years (odds ratio [OR] = 1.04) and the number of group visits (OR = 1.05) were significant predictors for achieving a 5% weight reduction. Given the importance of weight reduction, health professionals should consider these significant predictors when planning weight-reduction programs for veterans.
Congruent with the global problem of obesity, the prevalence of obesity among United States veterans was similar to or higher than the general population, and veterans receiving care at the Veterans Administration Medical Centers (VAMCs) had a higher prevalence of obesity than their counterparts who did not receive care at VAMCs (Almond, Kahwati, Kinsinger, & Porterfield, 2008; Das et al., 2005; Koepsell, Forsberg, & Littman, 2009). Since 2006, the United States Veterans Administration (VA) disseminated nationwide a weight management program designed specifically for veterans, MOVE!® (Kinsinger et al., 2009). On July 1, 2008, the VA stopped requiring payment from veterans to attend MOVE!®, so the program became free of charge to any veteran (Elimination of co-payment for weight management counseling, 2008). MOVE!® was a multi-tiered program with five levels of weight-reduction strategies featuring participant decisional control and tailored strategies according to participant responses to a questionnaire (Veterans Health Administration, 2006).
The MOVE!® Program
VAMCs implemented the MOVE!® program differently depending on local needs. The Charlie Norwood Veterans Administration Medical Center (CNVAMC) program offered two levels to individuals with a body mass index (BMI) of 25 or greater. Participants referred themselves to the program or were referred by a health care provider. Level 1 (self-managed) was a tailored educational program with self-management support and optional individual consultation and/or telephone contacts. The self-management approach included (1) educational materials tailored to each participant’s response to questions at program entry and (2) distribution of a pedometer following orientation. Level 2 (group attendee) included (1) educational materials tailored to each participant’s response to questions at program entry, (2) distribution of a pedometer following orientation, and (3) group visits consisting of 1-hr classes providing education on goal-setting, nutrition, behavior change, physical activity, and use of MyHealtheVet, an online resource for health care management. Initially, the program had eight classes; over time, more classes were added to expand nutrition and behavior change content. The classes were taught by a variety of professionals including a nurse, dietitian, physical activity specialist, and psychologist. Participation in Level 1 (self-managed) or Level 2 (group attendee) was based on participant preference. If a participant began with Level 1 and changed to Level 2, then the participant was considered a Level 2 participant. Participants determined their own participation schedule; therefore, missing a scheduled class was never considered a missed appointment. Participants returned for missed classes and repeated classes as frequently as desired. Participants in Level 1 or Level 2 had the option to request individual meetings in person or by telephone with the nurse, dietitian, physical activity specialist, and/or the psychologist. Prior to beginning the program, BMI was calculated from baseline height and weight and all participants completed the MOVE!® 23 Questionnaire. Their responses were used to generate the initial tailored plan; for example, the recommended educational materials were adjusted based on the needs identified from responses to items on the MOVE!® 23 Questionnaire. Ongoing tailoring occurred with individual visits and telephone contacts based on participant needs and desire for these resources.
Little research has focused on the MOVE!® program. One study focused on the trajectory of weight for participants in the Miami MOVE!® program (Dahn et al., 2011; Jay, 2011). White participants experienced a better weight-reduction trajectory than their Black counterparts, and participants attending group sessions experienced a better weight-reduction trajectory than their self-managed counterparts (Dahn et al., 2011; Jay, 2011). In a study of the trajectory of weight for participants in the Los Angeles MOVE!® program, an average weight reduction of 4.8 pounds a year after enrollment was reported for participants attending at least three group sessions (Romanova, Liang, Deng, Li, & Heber, 2013). In a study of MOVE!® programs in four western states, a weight reduction of 3.7 pounds was associated with six or more intervention encounters (Littman, Boyko, McDonell, & Fihn, 2012).
Purpose
To date, there are no published studies of weight reduction among veterans in the MOVE!® Program examining the influence of background characteristics and program characteristics with the level of detail undertaken in this study. A better understanding of these characteristics could provide evidence for developing future tailored interventions to increase the likelihood of weight reduction among veterans. Therefore, the purpose of this study was to determine which background characteristics and program characteristics were associated with a 5% or greater weight reduction for veterans enrolled in MOVE!® at CNVAMC.
Methods
Design, Sample, and Data Collection
A secondary data analysis was conducted using records of veterans enrolled in the weight management program, MOVE!®, at CNVAMC. Records of all veterans entering the MOVE!® program at CNVAMC from July 1, 2008, to May 31, 2010, were included. Data collection included the period between July 1, 2008, and June 30, 2010. Therefore, the data collection period varied for each participant and we controlled for the variation in observed time in the analyses (see Control Variable). Participants were excluded if records indicated baseline age of 90 years or older; a history of surgical intervention for weight reduction; pregnancy during the study period; or death during the study period. Georgia Regents University Internal Institutional Review Board and CNVAMC Research and Development Committee approved this study.
Measures
Weight reduction
The outcome measure was 5% or greater weight reduction. A 5% weight reduction is the first goal of the MOVE!® program at CNVAMC. This percentage of weight reduction has been frequently associated with clinical benefit in the literature (Mertens & Van Gaal, 2000; Neter, Stam, Kok, Grobbee, & Geleijnse, 2003; Pontiroli et al., 2005; Stevens et al., 2001). This percentage was the target outcome of two recent randomized controlled trials of structured commercial programs (Jebb et al., 2011; Jolly et al., 2011). Most recently, The Obesity Society released a guideline identifying 5% as the initial goal for weight-reduction interventions (Jensen, 2013). Weights were obtained from the electronic health record. Weight reduction was defined as the baseline weight in pounds minus the final documented weight in pounds. A weight reduction of 5% or more was based on the percentage of weight reduction from the baseline weight.
Background characteristics
Background characteristics included age, gender, ethnicity, race, marital status, comorbidities, BMI, rurality, and VA priority group. Comorbidities were examined in three ways. First, active diagnoses on record at the time of enrollment were counted to determine the number of comorbidities. Second, specific diagnoses of diabetes, hypertension, and hyperlipidemia/dyslipidemia were identified. Third, a Charlson Comorbidity Index (CCI) was calculated for each participant. The widely used CCI contains 19 categories of comorbidity (Charlson, Pompei, Ales, & MacKenzie, 1987) defined primarily by ICD-9 (International Classification of Diseases, version 9) codes and a few procedure codes (Deyo, Cherkin, & Ciol, 1992). Each of the 19 categories has an associated weight based on the adjusted risk of 1-year mortality (Charlson et al., 1987). The total CCI score reflects the cumulative increased likelihood of 1-year mortality (Charlson et al., 1987). With CCI scores ranging from 0 to 37, the higher CCI scores reflect more severe comorbidity burden (Charlson et al., 1987). BMI was calculated from height and weight. Rurality was based on zip codes transformed into Rural Urban Commuting Area (RUCA) codes (Economic Research Service, 2013). RUCA Codes 1 to 3 were classified as urban while Codes 4 to 10 were classified as rural. VA priority groups are classifications, based on level of service and disability and defined by federal legislation, which determine health care coverage for inpatient and outpatient care, medication, and extended-care services for veterans (Department of Veterans Affairs, 2010b). Therefore, priority group classifications change over time based on federal legislation. They were not clearly linked to duration of the veteran’s service to the country, but the priority groups were used to make decisions regarding resource allocation. For example, a veteran who was a prisoner of war was in Priority Group 1. VA Priority Groups ranged from 1 to 8 with Priority Group 1 being the highest priority and entitling the veteran to the most coverage in all categories of care, including inpatient, outpatient, medication, and extended care, with no co-payment (Department of Veterans Affairs, 2014).
Program characteristics
Program characteristics included program type (group attendee vs. self-managed); number and type of provider contact including total number of group visits with each type of provider, total number of individual visits with each type of provider, and total number of telephone visits with each type of provider; and total number of group visits. Group attendee and self-managed were mutually exclusive categories. Group attendee was defined as anyone who attended a group session beyond orientation. If a participant began with self-managed and changed to group, then the participant was considered a group attendee. Participants had the option of visits with various types of providers during the program including nurses; dietitians, psychologists; or activity specialists such as physical, kinesthetic, or occupational therapists. The visits could have been group classes, individual visits, and/or telephone visits.
Control variable
To control for the variation in the length of time each participant was observed during the data collection period, we created a variable representing the time each participant was observed (total number of weeks from baseline weight to the final weight recorded during the observation period, July 1, 2008, and June 30, 2010). The time during the observation period ranged from 0 to 103 weeks. As this variable was not normally distributed, we created a categorical variable representing time observed in quartiles.
Statistical Analysis
Descriptive statistics were calculated for all variables. Logistic regression was used to examine variables associated with achieving a 5% or greater weight reduction. A dichotomous variable represented a weight reduction of at least 5% (0 = no, 1 = yes) from baseline to the final weight. Bivariate associations were calculated to describe the likelihood of achieving at least the 5% weight-reduction goal for each background and program characteristic.
Next, we developed models for achieving the goal of 5% or greater weight reduction between baseline and final weight. The control variable for time observed was entered first followed by background and program characteristics that were significant at p < .25 in the bivariate models. Significant variables were retained (p < .05). Several models were fit and models were compared by examining the coefficients of each variable with coefficients from the bivariate models, likelihood ratio tests, and model diagnostics. The decision on the best-fitting model was made on the basis of parsimony and clinical importance. We analyzed the data using SPSS 20.
Results
Sample Characteristics
The 404 MOVE!® program participants in the sample were primarily non-Hispanic (96%), urban (83%), and male (~80%; see Table 1). Age ranged from 21 to 81 years with a mean of 56.44 years (SD = 11.27) and a median of 59 years. Over half of the sample were Black (58.4%) and married (58.4%). The number of comorbidities ranged from 0 to 27 (M = 8.65, SD = 5.32); the CCI ranged from 0 to 7 (M = 0.86, SD = 1.18). The sample had common comorbidities associated with obesity, including Type 2 diabetes (30.2%), hypertension (60.9%), and hyperlipidemia (54.0%). The average BMI was 34.96 (SD = 5.94) and ranged from 25.33 to 62.69 with baseline weights ranging from 149 to 459 pounds, 236.86 pounds on average (SD = 44.5), and a median weight of 229.85 pounds. The sample included all eight VA Priority Groups, but Group 1 (35%) and Group 5 (24%) accounted for over half of the sample. For further analyses, the VA Priority Groups were examined as dichotomized resources in two ways. One way dichotomized those in Group 1, qualifying for the most resources, from those in all other groups. Another way dichotomized those in Groups 1 through 4 from those in Groups 5 through 8. On average, participants had a weight reduction of 2.8 pounds (SD = 11.4). Weight change ranged from a 28 pound weight gain to an 84 pound weight reduction. Thirteen percent (n = 53) or approximately one out of every eight participants achieved the goal of a 5% or more weight reduction.
Baseline Background Characteristics (N = 404 Participants).
Note. BMI = body mass index; VA = Veterans Administration.
Program Characteristics
Table 2 describes program characteristics (participation in group, individual, and telephone components of the weight management program). Over half of the participants were group attendees (n = 226, 55.9%). The majority of the visits were group visits. On average, participants attended 4.83 group visits (SD = 6.84) with a range of 1 to 71; the median was two group visits; 25% of participants attended more than six visits.
Program Characteristics.
Note. RD = registered dietitian; PA = physical activity.
Values are a percentage of the 404 participants.
Values are for those who participated in a given program component, not for the 404 total program participants. For example, of the 262 participants who had a telephone visit with the registered nurse, each participant had one visit and some participants had as many as three visits; on average, the number of telephone visits with the registered nurse was 1.94 for the 262 participants who had a telephone visit with the registered nurse.
Logistic Regression
Table 3 presents the bivariate associations for each of the background and program characteristics with the outcome of a 5% weight reduction. Controlling for variation in time observed from first to last recorded weight, the best-fitting regression model included age and total number of group visits (see Table 4). For every year of age, the odds of achieving the 5% threshold for weight reduction increased by 4%, holding other variables constant. For each group visit, the odds of achieving the 5% threshold for weight reduction increased by 5%, holding other variables constant.
Bivariate Associations for Background and Program Characteristics With 5% Weight Reduction.
Note. OR = odds ratio; CI = confidence interval; BMI = body mass index; CCI = Charlson Comorbidity Index; VA = Veterans Administration; RD = registered dietitian; PA = physical activity.
p < .05.
Final Logistic Regression Model for Successful Weight Reduction.
Note. Overall model Cox and Snell R2 = .04, χ2 = 18.07, df = 5, and p < .01; Hosmer and Lemeshow Test χ2 = 14.50, df = 8, and p = .07. OR = odds ratio; CI = confidence interval.
The fourth quartile representing 68 to 103 weeks was the reference.
p < .05.
Discussion
In this study, achieving a 5% or more weight reduction was associated with age and total number of group visits. Consistent with earlier reports of positive health behaviors, advancing age had a beneficial effect on weight reduction (Yarcheski, Mahon, Yarcheski, & Cannella, 2004). For each year of age in our study, an individual was 4% more likely to achieve a 5% weight reduction. For example, at 21 years of age, the odds of achieving a 5% weight reduction were .84; while at 71 years of age, the odds of achieving a 5% weight reduction were 2.84. Similarly, researchers reported that participants in group sessions at the Miami MOVE!® program were older and more successful with weight reduction than their counterparts in the self-managed group (Dahn et al., 2011; Jay, 2011). Large-scale studies such as the Diabetes Prevention Program and Look AHEAD studies also indicated that participants who lost more weight were older (Diabetes Prevention Program Research Group et al., 2002; Wing et al., 2011). The Diabetes Prevention Program Research Group proposed that older persons may have had fewer demands on their time thus allowing them to be more attentive to weight-reduction interventions (Diabetes Prevention Program Research Group et al., 2009). As the participants were previously unsuccessful at weight management, it is unlikely that a large enough number of progressively aged participants would experience a 5% weight reduction for reasons beyond effective program strategies during the relatively short observation in our study; however, other unmeasured factors, such as poor-fitting dentures, loss of appetite, lack of food, or other illnesses, could have contributed to weight reduction. In addition, in our study, there is no way to determine if the weight reduction was associated with a loss of fat or muscle. Furthermore, the design and delivery of the program may have been better suited to older adults.
The total number of group visits was associated with weight reduction. The odds of achieving a 5% weight reduction increased by 5% for each group visit attended. For example, a participant who attended three group visits had an odds of 0.15 of achieving a 5% weight reduction while a participant who attended 25 group visits had an odds of 1.25 of achieving a 5% weight reduction. Our findings are consistent with earlier studies supporting the benefit of group attendance and the benefit of more group visits for greater weight reduction (Dahn et al., 2011; Jay, 2011; Littman et al., 2012; Paul-Ebhohimhen & Avenell, 2009).
This study has implications for practice. First, it may be necessary to develop interventions that are more attractive to younger participants or more suited to those with limited time to focus on weight-reduction activities. Second, the effect of group attendance on a 5% weight reduction was dose-related indicating that more group visits were associated with greater weight reduction. The dose effect along with the chronic nature of obesity and the small percentage of participants meeting the 5% weight-reduction goal suggest that health care professionals should develop effective strategies for engaging participants in group weight-reduction activities over long periods of time. Third, as the sample on average was at a critical point for weight management in terms of BMI and comorbidities, participants should be carefully evaluated for other appropriate treatment options. The program may need to be expanded to accommodate participants who qualify for surgical intervention.
Several issues should be considered in future research. While 13% of the participants were successful at reducing their weight by 5%, further investigation is needed to fully describe successful weight reducers and to develop methods for unsuccessful participants to become successful. Qualitative inquiry may provide insight into the processes of decision making, useful tools to aid in weight reduction, differences between successful weight reducers and non-reducers, and the role age plays in weight reduction. The intervention tools may have been more appealing to older veterans and more contemporary tools may be more beneficial to younger veterans. In addition, this study did not attempt to describe the change in trajectory of weight from before to after the program. Demonstrating even a minimal reduction in weight gain would be beneficial over an ongoing weight gain. A comparison of qualified, invited participating and non-participating veterans would contribute to determining the clinical value of the MOVE!® program. These inquiries should be followed or accompanied by an investigation of cost-effectiveness.
The primary strength of this study is that it involved data from a practice setting with veterans (Twisk, 2003). There is great potential value in practice-based evidence whereby researchers can examine outcomes of current treatments in the real world where veterans live (Horn & Gassaway, 2007). In addition, the translation of such research into practice may have greater applicability when research is based within practice.
Another strength of this sample is that Blacks and females were well represented. Actually, the sample had more Blacks and more females than the national average (Department of Veterans Affairs, 2010a). In addition, although weight reduction in men has been less frequently studied, this study included 80% males and thus contributes to the weight-reduction literature for males.
Using secondary data collected for clinical purposes has limitations. First, there could have been missing or inaccurate data. For example, some participants may have worn coats or shoes when weighed. In addition, there were differences in the instruments used to assess the outcomes, for example different scales were used to assess weight. These errors, omissions, and variations could threaten internal validity (Polit & Beck, 2012). Unidentified and unmeasured differences may have contributed to the participant’s selection of program- group attendee or self-managed, degree of participation, and the outcome of a 5% weight reduction. Additional measures of education, literacy, social support, previous health care experiences, motivation, cognitive appraisal, affective responses, medications, and intermediate outcomes of dietary intake and physical activity were desirable but unavailable for this study. These unidentified and unmeasured differences could represent a threat to internal validity (Polit & Beck, 2012). However, other actions to reduce weight (e.g., seeking health care provider assistance and/or medications with a side effect of weight reduction) were consistent with MOVE!® and would not be an alternative explanation.
An observational study such as this one has limited generalizability (Polit & Beck, 2012). This sample differed from the national veteran population. CNVAMC had more Blacks (58% vs. 11%) and more females (~21% vs. 8%) compared with the national average (Department of Veterans Affairs, 2010a). This sample had fewer married veterans (58% vs. 70%) than were reported nationally (Department of Veterans Affairs, 2010a). Furthermore, each VAMC implemented the program as the local providers saw best and as resources would allow. Thus, the VAMC’s MOVE!® program varies somewhat from site to site.
In conclusion, many studies have investigated weight reduction but none focused on MOVE!® and reported examining background characteristics and program characteristics with such detail. In this study, older participants benefited more than their younger counterparts. This study supported the benefit of group attendance with a dose effect for the number of group visits in the MOVE!® program. While weight reduction remains a complex issue, this study explained a segment of the phenomena and will provide valuable information for administrators, providers, and veterans.
Footnotes
Acknowledgements
The first author thanks M. Katherine Maeve, PhD, RN for her support as Program Director for the VA Predoctoral Nursing Fellowship. The authors thank Gayle Sprott, RN, CNVAMC MOVE!® Coordinator, for her help in understanding how the program worked at this particular VA, Caroline McKinnon, PhD, CNS/PMH-BC for assistance with data collection, and Richard T. Campbell, PhD, Emeritus Professor of Biostatistics and Sociology Research Scientist, University of Illinois at Chicago Institute for Health Research and Policy, for statistical support.
Authors’ Note
The contents of this publication do not represent the views of the Department of Veterans Affairs, the United States Government, or Georgia Regents University.
Declaration of Conflicting Interests
The author(s) declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: This material is based upon work supported by the Department of Veterans Affairs, Office of Academic Affairs, Predoctoral Nursing Fellowship to the first author, and by the Georgia Regents University College of Nursing.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was supported by the Department of Veterans Affairs, Office of Academic Affairs, Predoctoral Nursing Fellowship (first author) and Georgia Regents University College of Nursing Seed Grant (first author).