Abstract
Objectives. We determined the efficacy and cost-effectiveness of adding an evidence-based Internet behavioral weight loss intervention alone or combined with optional group sessions to ShapeUp Rhode Island 2011 (SURI), a 3-month statewide wellness campaign.
Methods. We randomized participants (n = 230; body mass index = 34.3 ±6.8 kg/m2; 84% female) to the standard SURI program (S) or to 1 of 2 enhanced programs: SURI plus Internet behavioral program (SI) or SI plus optional group sessions (SIG). The primary outcome was weight loss at the end of the 3-month program.
Results. Weight losses differed among all 3 conditions (S: 1.1% ±0.9%; SI: 4.2% ±0.6%; SIG: 6.1% ±0.6%; Ps ≤ .04). Both SI and SIG increased the percentage of individuals who achieved a 5% weight loss (SI: 42%; SIG: 54%; S: 7%; Ps < .001). Cost per kilogram of weight loss was similar for S ($39) and SI ($35); both were lower than SIG ($114).
Conclusions. Although weight losses were greatest at the end of SURI with optional group sessions, the addition of an Internet behavioral program was the most cost-effective method to enhance weight losses.
Excess adiposity is associated with increased health risk.1,2 Behavioral weight loss programs are the treatment of choice for overweight and moderate obesity. These programs produce weight losses of 8 to 10 kilograms during the initial months of treatment,3 which are associated with significant health improvements.4,5 However, the programs are intensive, expensive, and only accessible to a small portion of individuals in need. Thus, there is critical need to develop cost-effective weight loss programs that are accessible to large numbers of overweight and obese individuals.
Community weight loss campaigns are typically offered via the Internet and reach large numbers of people. However, weight losses produced in these programs are modest. Thus, the challenge is to use these wide-reaching community programs to attract large numbers of individuals and enhance the weight losses without substantially increasing cost. We have begun to programmatically test ways to improve weight outcomes in ShapeUp Rhode Island (SURI), an annual 3-month Internet-based community initiative that attracts thousands of overweight and obese adults. In an earlier trial,6 we showed that providing SURI participants with behavioral weight loss strategies via e-mail (i.e., sending weekly PowerPoint slides and providing feedback) significantly improved 3-month weight losses (–3.1 kg vs –1.2 kg). If providing access to evidence-based weight loss strategies was found to be both effective and cost-effective, the potential public health impact would be substantial. In addition, it is important to evaluate other approaches that might further enhance SURI outcomes.
This trial extended our previous work. Based on the content of our previous e-mail intervention, we developed an Internet behavioral weight loss Web site and examined whether it improved weight loss outcomes in SURI 2011. In addition, we examined whether adding optional group sessions to the new Internet program further improves weight losses. Secondary aims explored the cost-effectiveness of these strategies and examined weight loss trajectories following treatment. We hypothesized that SURI weight losses might be increased if, in addition to adding an Internet behavioral program, participants were offered optional group sessions. However, because adding group meetings would likely increase cost, we examined whether the expected increased weight loss would justify the increased cost. The primary endpoint was 3 months.
SURI 2011 was promoted to the community via employers, media, and mass mailings. During registration, participants chose whether to join the weight loss or physical activity division, or both. Those who joined the weight loss division (n = 3806) were asked whether they could be contacted for a weight loss research study. A total of 1139 agreed to be contacted. As this was a research study and all participants had to be processed through the protocol before the start of SURI (consent, assessment, randomization), individuals were enrolled on a first-come, first-served basis; specifically, the first 431 respondents who expressed interest were screened for study eligibility. Exclusion criteria were minimal and focused on safe participation in an unsupervised weight loss trial and practical issues. Specifically, exclusion criteria were age younger than 18 years or older than 70 years; body mass index (defined as weight in kilograms divided by the square of height in meters [kg/m2]) less than 25 kg/m2; pregnant, nursing, or plans to become pregnant; serious medical condition (e.g., cancer); unreliable Internet access; non-English speaking; current or previous participation in our weight loss studies; or planned relocation. Those who reported a medical condition that could interfere with safe participation (e.g., diabetes) obtained doctor’s consent to participate. Of the 431 individuals screened, 230 met inclusion criteria, completed orientation procedures, and were randomized (Figure 1).
FIGURE 1— Flow of participants through the ShapeUp Rhode Island 2011 Research Study.
Note. BMI = body mass index; S = ShapeUp Rhode Island 2011 (SURI) program; SI = SURI + Internet behavioral program; SIG = SURI + Internet behavioral program + optional group sessions.
With a random number generator, we assigned participants by using a 1:2:2 randomization scheme to SURI alone (S; n = 46); SURI plus Internet behavioral weight loss program (SI; n = 90); or SURI plus Internet behavioral weight loss program plus optional group sessions (SIG; n = 94). To avoid contamination, we randomized individuals on the same team (see next section) to the same intervention. The study statistician conducted the randomization.
SURI 2011 was a 3-month, statewide program. Participants joined in teams, entered the weight loss or physical activity division, or both, and competed with other teams on these domains. Throughout the 3-month program participants had access to a reporting Web site where they submitted their weekly weight and activity data and viewed their personal and team progress. They also received paper logs to record weight and activity, a pedometer, access to newsletters and community workshops, and recognition for meeting goals. Participants in the S arm (n = 46) of this trial had access to the SURI program only and did not receive any behavioral weight loss treatment.
Participants in the SI arm (n = 90) received the 3-month SURI program plus a 3-month Internet behavioral weight loss intervention. Before SURI began, participants attended a 1-hour group meeting during which they received their weight loss goal (lose 1 to 2 pounds per week), calorie and fat gram goal (starting weight < 250 lbs: 1200–1500 kcal/day, 40–50 g of fat; starting weight ≥ 250 lbs: 1500–1800 kcal/day, 50–60 g of fat), and activity goal (gradually increase to 200 minutes of aerobic activity per week). During this session, participants were also taught self-monitoring skills and oriented to the behavioral weight loss program Web site. The Web site included 12 weekly, 10- to 15-minute multimedia lessons based on the Diabetes Prevention Program7 and a self-monitoring platform where participants tracked their daily weight, calorie, and activity information. On the basis of the information reported into the self-monitoring platform, participants received weekly automated feedback on their progress. The Web site also included information on meal plans, prepackaged meals, and meal replacements. Participants only had access to the Web site during the 3-month SURI program.
Participants in the SIG arm (n = 94) received everything described in SI (previous section) and were given the option to attend weekly group meetings at the research center during the 3-month SURI program. As SURI was Internet-based and very low intensity, we wanted this latter arm to appeal to the SURI audience. Thus, unlike typical in-person behavioral weight loss programs where attendance at each group session is expected,3 we made group sessions optional. The 12 weekly, optional group sessions were led by Masters-level staff with extensive training in behavioral weight loss. Sessions involved private weigh-ins and covered topics that supplemented the Internet intervention (e.g., recipe modification, portion control).
Participants completed measures in person at baseline and 3 months (posttreatment) and at 6-and 12-month follow-ups, unless noted otherwise. Questionnaires were completed with paper-and-pencil scanforms and weight and height were objectively assessed. Assessment staff was blinded to treatment arm. Participants were compensated $25 for the 3- and 6-month assessments and $50 for the 12-month assessment. The primary outcome was weight loss at 3 months (i.e., posttreatment).
We obtained basic demographic information at baseline. Weight was measured to the nearest 0.1 kilogram with a digital scale. Height was measured to the nearest millimeter with a stadiometer (baseline only). During the 3-month program, the SURI Web site encoded submission of weight information each week, the behavioral weight loss study Web site tracked submission of self-monitoring information and number of lessons viewed, and attendance at group sessions was recorded by intervention staff.
We used the Weight Control Practices Questionnaire8 to assess whether participants engaged in 7 core weight loss strategies targeted in treatment (self-monitoring diet and physical activity, counting calories and fat grams, decreasing caloric and fat intake, and self-weighing; “Yes” = 1; “No” = 0). We summed item responses to create a composite behavioral strategy score. This questionnaire and a similar scoring approach has been used previously and is associated with weight outcomes in behavioral treatment.6,8 The Paffenbarger Questionnaire, a widely used and well-validated measure of physical activity,9,10 assessed weekly kilocalories expended in moderate to vigorous activity.
We assessed costs from the payer, participant, and societal perspective (i.e., sum of payer and participant costs). Payer costs were the sum of labor, rent, and intervention materials. We estimated labor costs associated with intervention preparation and delivery by using market values for staff time.11 We assessed rental space for intervention sessions ($50 per hour) and intervention material costs by using local costs in Providence. Participant direct costs included the SURI registration fee ($20) and transportation costs associated with attending intervention sessions. We estimated indirect costs for participant time spent on the intervention (e.g., using Web sites, attending sessions) and related travel time by using average wage rates of US adults.12 Additional details on the cost-effectiveness analyses are provided in the Appendix (available as a supplement to the online version of this article at http://www.ajph.org).
We performed statistical analyses by using SAS version 9.3 (SAS Institute, Cary, NC). We examined baseline group differences with analyses of variance and the χ2 test for continuous and categorical variables, respectively. We examined adherence metrics and correlates by using simple descriptives and correlations. We used PROC MIXED to conduct a longitudinal linear mixed model analysis, fit with an autoregressive covariance structure to examine group differences in weight loss. PROC MIXED accommodates missing values under the assumption of missing at random13; thus, we included all randomized participants. To evaluate the primary aim, within the larger longitudinal model, contrast tests examined group differences at month 3. To examine group differences in number of participants achieving a 5% or greater weight loss, we used generalized estimating equations with robust standard errors with PROC GENMOD.14 We analyzed effects of treatment on weight control practices and physical activity longitudinally with PROC MIXED, using an autoregressive covariance structure.
Cost analyses included participants with both cost and weight data. We allocated program costs to individual participants. We allocated attendance-related costs to participants who attended sessions; other costs were allocated evenly to all participants in the treatment group. We determined per capita costs from the payer, participant, and societal perspectives. With these estimates, we drew 10 000 bootstrap samples to construct 95% bias-corrected confidence intervals (CIs) and P values.15
Participants were predominantly female, non-Hispanic White and had a mean body mass index of 34.3 kg/m2 (SE = 0.5; Table 1). Groups differed on education (P = .02). Thus, we examined effects of education on weight loss; weight loss did not differ by educational attainment (Ps > .57). Ninety-three percent of participants attended the posttreatment assessment and 91% and 86% of participants attended the 6- and 12-month follow-up assessments, respectively.
TABLE 1— Baseline Characteristics of Participants: 2011 ShapeUp Rhode Island Research Study
| Characteristic | S (n = 46), Mean ±SE or % (95% CI) | SI (n = 90), Mean ±SE or % (95% CI) | SIG (n = 94), Mean ±SE or % (95% CI) | P |
| Gender | .74 | |||
| Female | 82.6 (71.6, 93.6) | 82.2 (74.3, 90.2) | 86.2 (79.1, 93.2) | |
| Male | 17.4 (6.4, 28.4) | 17.8 (9.8, 25.7) | 13.8 (6.8, 20.9) | |
| Age, y | 46.5 ±1.7 | 46.2 ±1.2 | 47.7 ±1.1 | .63 |
| Race | .17 | |||
| White | 91.3 (83.1, 99.5) | 90.0 (83.8, 96.2) | 81.9 (74.1, 89.8) | |
| Non-White | 8.7 (0.5, 16.9) | 10.0 (3.8, 16.2) | 18.1 (10.2, 25.9) | |
| Ethnicity | .08 | |||
| Not Hispanic/Latino | 97.8 (93.6, 100.0) | 97.8 (94.7, 100.0) | 91.3 (85.5, 97.1) | |
| Hispanic/Latino | 2.2 (0.0, 6.4) | 2.2 (0.0, 5.3) | 8.7 (2.9, 14,5) | |
| Education | .02 | |||
| Vocational, high school, or less | 15.2 (4.8, 25.7) | 3.3 (0.0, 7.1) | 11.8 (5.2, 18.4) | |
| Some college or college graduate | 69.6 (56.2, 83.0) | 61.1 (51.0, 71.3) | 54.8 (44.6, 65.0) | |
| Postgraduate | 15.2 (4.8, 25.7) | 35.6 (25.6, 45.5) | 33.3 (23.7, 43.0) | |
| Weight, kg | 95.3 ±3.4 | 94.3 ±1.9 | 91.2 ±2.2 | .43 |
| BMI, kg/m2 | 35.1 ±1.3 | 34.7 ±0.7 | 33.4 ±0.7 | .27 |
Note. BMI = body mass index; CI = confidence interval; S = ShapeUp Rhode Island 2011 (SURI) program; SI = SURI + Internet behavioral program; SIG = SURI + Internet behavioral program + optional group sessions. The sample size was n = 230.
Participants reported their weight and physical activity data into the SURI Web site an average of 8.4 of the 12 weeks (SE = 0.3). S participants used this Web site significantly more (9.8; SE = 0.5) than did SI participants (7.4; SE = 0.5; P = .002); SIG (8.6; SE = 0.4) did not differ significantly from the other 2 conditions. More reporting was associated with greater percentage weight loss (r = 0.25; P < .001). There were no significant differences in adherence to the Internet behavioral weight loss program between SI and SIG (Ps ≥ .25). Participants in SI and SIG logged into the study Web site an average of 8.9 weeks (SE = 0.3) out of 12 and viewed 6.3 of the 12 multimedia lessons (SE = 0.3); logins and lessons viewed were positively associated with percentage weight loss (r = 0.45; P ≤ .001; and r = 0.34; P ≤ .001, respectively). Similarly, during the 84-day program, participants in SI and SIG reported their weight on 57.9 days (SE = 2.1) and their calories on 56.4 days (SE = 2.1); again, more frequent reporting was associated with greater percentage weight loss (r = 0.47; P < .001; and r = 0.47; P < .001, respectively). Participants in the SIG arm attended an average of 6.2 (SE = 0.5) of 12 group sessions, with greater attendance associated with better weight outcomes (r = 0.61; P ≤ .001).
We examined whether participant characteristics were associated with intervention adherence in each condition. Younger age was consistently associated with poorer adherence. Younger participants reported their information less frequently into the SURI Web site in SIG (r = 0.23; P = .024), viewed fewer multimedia lessons in SI (r = 0.35; P = .001), and, in both SIG and SI, logged in less frequently to the study Web site (r = 0.22; P = .034; and r = 0.27; P = .012, respectively) and reported self-monitoring data less often (r’s = 0.23–0.30; Ps ≤ .029). Racial/ethnic minority status was associated with reporting fewer weeks of data into the SURI Web site in SI (White: 7.7; SE = 0.5 vs non-White: 4.3; SE = 1.6; P = .048), and men viewed fewer multimedia lessons than women in SIG (4.0; SE = 1.2 vs 6.6; SE = 0.4; P = .021). Remaining effects of participant characteristics on intervention adherence were not significant.
Weight losses for the 3 treatment conditions are shown in Figure 2. Overall, there was a significant group-by-time interaction (P < .001). Our primary hypothesis, which focused on weight loss at the end of the 3-month SURI program, was confirmed with significant differences in weight loss among each of the 3 conditions. Percentage weight loss was greatest in SIG (6.1%; SE = 0.6) followed by SI (4.2%; SE = 0.6) and lowest in S (1.1%; SE = 0.9; all comparison Ps ≤ .04). During the no-treatment follow-up phase, initial weight losses were largely maintained through month 6 (SIG: 6.2 [SE = 0.6]; SI: 3.6 [SE = 0.6]; S: 1.0 [SE = 0.9]; all comparison Ps ≤ .02); however, at month 12, the 3 groups no longer differed from one another (SIG: 3.3 [SE = 0.6]; SI: 2.2 [SE = 0.6]; S: 1.2 [SE = 0.9]; Ps ≥ .05).
FIGURE 2— Percentage weight loss in the 2011 ShapeUp Rhode Island Research Study.
Note. S = ShapeUp Rhode Island 2011 (SURI) program; SI = SURI + Internet behavioral program; SIG = SURI + Internet behavioral program + optional group sessions. Different letters indicate that weight losses differ significantly from one another at that time point. Error bars represent standard error of the mean.
There was an overall group effect for percentage of individuals meeting a 5% weight loss (P < .001) and a marginal group-by-time interaction (P = .05). At 3 months, a greater percentage of participants in SIG (54.3%) and SI (42.2%) achieved 5% or greater weight loss compared with S (6.5%; Ps < .001); SIG and SI did not differ from one another (P = .1). This initial effect was maintained at the 6-month no-treatment follow-up assessment (SIG: 44.7%; SI: 34.4%; S: 15.2%); however, at the 12-month follow-up, there were no significant differences among the 3 groups (SIG: 25.5%; SI: 18.9%; S: 13.0%).
There was a significant group-by-time interaction for the use of behavioral weight loss strategies (P < .001). The SIG and SI groups had greater increases in strategies from pre- to posttreatment than did the S group (SIG: mean = +4.1 [SE = 0.2]; SI: mean = +3.8 [SE = 0.2]; S: mean = +1.3 [SE = 0.3]; Ps < .001), with no significant difference between SIG and SI (P = .23; Table 2). Increased use of behavioral strategies was associated with greater percentage weight loss in all 3 conditions at posttreatment (S: r = 0.39; P = .015; SI: r = 0.60; P < .001; SIG: r = 0.57; P < .001). After treatment, SIG was engaging in more weight control practices than S at month 6 (P = .003), with SI not significantly different from the other 2 conditions (Ps ≥ .07). At the 12-month follow-up, there were no differences between groups in the use of behavioral strategies (Ps ≥ .07).
TABLE 2— Pre- to Posttreatment Changes in the Use of Behavioral Weight Loss Strategies: 2011 ShapeUp Rhode Island Research Study
| S (n = 46) | SI (n = 90) | SIG (n = 94) | ||||
| Strategy | Baseline, % or Mean ±SE | Month 3, % or Mean ±SE | Baseline, % or Mean ±SE | Month 3, % or Mean ±SE | Baseline, % or Mean ±SE | Month 3, % or Mean ±SE |
| Total no. of strategies | 1.9 ±0.29 | 3.3 ±0.29a | 1.9 ±0.21 | 5.7 ±0.21b | 2.0 ±0.20 | 6.1 ±0.20b |
| Use of specific strategies | ||||||
| Recording dietary intake | 32.6 | 43.5 | 26.7 | 84.4 | 34.0 | 92.6 |
| Recording physical activity | 15.2 | 41.3 | 15.6 | 72.2 | 16.0 | 73.4 |
| Counting calories | 30.4 | 41.3 | 17.8 | 85.6 | 24.5 | 92.6 |
| Counting fat grams | 15.2 | 26.1 | 15.6 | 78.9 | 18.1 | 88.3 |
| Decreasing calories | 43.5 | 76.1 | 50.0 | 91.1 | 50.0 | 94.7 |
| Decreasing fat grams | 37.0 | 60.9 | 45.6 | 85.6 | 46.8 | 89.4 |
| Daily self-weighing | 19.6 | 39.1 | 17.8 | 72.2 | 10.6 | 74.5 |
| Activity kcal/wk | 950.8 ±172.8 | 1749.6 ±172.8a | 954.1 ±23.5 | 1214.6 ±23.5b | 962.0 ±20.8 | 1486.1 ±20.8a,b |
Note. S = ShapeUp Rhode Island 2011 (SURI) program; SI = SURI + Internet behavioral program; SIG = SURI + Internet behavioral program + optional group sessions. Month-3 immediate posttreatment values with different superscripts indicate a significant difference in change scores between conditions at P < .05.
There was no group-by-time effect for physical activity; however, all groups reported significant increases in physical activity over time (P < .001), particularly during the initial 3-month program. Over the first 3 months, increase in physical activity was significantly greater in S than SI (mean = +798.7 kcal/week [SE = 161.7]; mean = +260.5 kcal/week [SE = 115.6]; P = .012); The SIG group was not significantly different from the other 2 conditions (+524.1 kcal/week [SE = 113.1]; Ps > .21). Improvements in physical activity were associated with greater weight loss at posttreatment in SI and SIG (SI: r = 0.21; P = .048; SIG: r = 0.33; P = .003); however, they were not associated with weight loss in S (r = 0.16; P = .3). Improvements in physical activity that occurred during the 3-month intervention declined once treatment ended; at the 12-month follow-up overall improvements in activity did not differ among groups (Ps > .05).
There were significant differences among the 3 groups in treatment costs from the payer, participant, and societal perspective. Costs per participant from the societal perspective were lowest in S ($36; 95% CI = $35, $38), followed by SI ($138; 95% CI = $131, $145) and highest in SIG ($594; 95% CI = $531, $658; Ps < .001 for all comparisons). Cost-effectiveness ratios (cost per kg of weight lost) were similar for S and SI ($39 per kg; 95% CI = $21, $155; $35 per kilogram; 95% CI = $28, $46, respectively) and both were lower than SIG ($114 per kg; 95% CI = $98, $134). Sensitivity analyses that reduced participant transportation costs and time traveling to and attending group sessions by 50% did not affect the overall results (see Appendix, available as a supplement to the online version of this article at http://www.ajph.org).
The primary aim of this study was to examine the 3-month efficacy of adding behavioral weight loss strategies to a community wellness campaign. Results showed that adding a novel evidence-based Internet behavioral weight loss program to the community campaign more than tripled the weight loss. Offering optional group sessions to supplement the Internet program further enhanced weight loss outcomes. Moreover, both enhanced conditions increased the odds of achieving a clinically significant weight loss 6-fold.
These results compare favorably to those of other studies that have combined empirically supported strategies with large-scale community campaigns. In our previous study6 we showed that providing SURI participants with behavioral strategies via e-mail increased the overall weight loss by 2.2%, or 1.9 kilograms. The current results are consistent with these initial findings (weight loss improvement: +3.2% or 2.2 kg). Brownell et al.16 conducted a series of studies to examine whether adding a weekly, face-to-face, behavioral weight loss program to a worksite wellness initiative enhances outcomes. Attrition was high in all studies (34%–58%) and, considering the intensity of the program (weekly in-person meetings), completer weight losses were modest (3.3 kg to 4.1 kg). Similarly, Graffagnino et al.17 more recently tested whether adding weekly individual counseling, dietary plans, and a free gym membership improves outcomes in a community campaign. Only 47% of participants completed and, despite the intensity, intent-to-treat weight losses were only 2.8 kilograms. In contrast to these previous studies, the more intensive, group approach tested herein yielded a substantially higher completion rate (95%) and better weight outcomes (–4.9 kg). The flexibility offered by optional group sessions and access to an Internet program for those who did not want to attend group sessions may be more appealing for this population of individuals who chose a low-intensity community intervention for weight loss. Future programs that attempt to enhance outcomes in lower-intensity campaigns may consider making intensive components optional and including lower-intensity alternatives (e.g., Internet intervention) to promote wider appeal, engagement, and, therefore, better weight loss outcomes.
Our process data suggest that weight losses achieved in the 2 enhanced conditions are likely attributable to engagement in our behavioral programs and the use of behavioral strategies. Web site use and session attendance were positively associated with percentage weight loss. From pre- to posttreatment, participants in the enhanced conditions had a greater increase in the use of behavioral weight loss strategies compared with those in the SURI-alone condition. Surprisingly, physical activity increases were higher in the SURI-alone condition compared with the enhanced Internet arm. Given that participants in SURI alone were not provided evidence-based weight loss skills (e.g., decreasing caloric intake, reducing calories from fat, and increasing behavioral skills), they may have focused solely on increasing physical activity, and thus achieved greater success in this domain, whereas the other groups were focusing on changing both energy intake and physical activity. It was interesting that the increase in physical activity in the SURI-alone condition was not associated with weight losses whereas it was in the other 2 conditions. This result is consistent with evidence that physical activity alone produces only modest weight losses.18 Taken together, the results from this study underscore the importance of incorporating all behavioral weight loss components (diet, activity, behavioral skills) into community interventions to improve outcomes.
In addition, we found that younger individuals were generally less adherent to the Internet-based interventions than were older participants. This finding is consistent with results from previous studies showing that young adults are difficult to engage in weight loss treatment, regardless of intervention modality (Internet or face-to-face).19,20 The findings from our study and these previous studies highlight the importance of developing novel interventions that engage younger individuals in weight loss treatment.
Cost-effectiveness analyses showed that combining SURI with an Internet behavioral program was a cost-effective method to enhance weight losses within a community campaign. The minimal staff time and participant time and travel required for the Internet program minimized the costs associated with this intervention. The addition of optional group meetings improved weight losses but significantly increased costs. The cost-effectiveness ratios observed in this trial were similar to estimates in other behavioral weight loss interventions. In a previous study of an Internet weight loss program21 the cost per kilogram lost averaged $26. Programs with group sessions are generally more expensive22–26 and, even though they yield greater weight losses, remain more expensive per kilogram weight loss ($409 per kilogram22). Cost of group meetings could be reduced in the future by using lay leaders and providing sessions throughout the community, thereby reducing time and travel expenses.
Our no-contact follow-up data are consistent with several studies suggesting a great need for effective maintenance interventions following initial weight loss. Weight loss was well maintained from 3 to 6 months without any contact; however, at month 12, the weight losses of SI and SIG averaged approximately 50% of the initial weight loss and were no longer significantly different from the standard SURI program. The differences between groups in the use of behavioral weight loss strategies also eroded with time. To help sustain initial weight losses, future programs should consider providing participants access to the evidence-based behavioral weight loss strategies after the 3-month SURI program or develop other Internet interventions that can be used during maintenance. Additional efforts are needed to develop and test effective weight maintenance programs that can be implemented in community initiatives.
This study is not without limitations. The sample was predominantly female and White; however, these sample characteristics are largely consistent with the broader ShapeUp population. Strengths of this study include its focus on the dissemination of evidence-based weight loss strategies, which, in light of the epidemic of obesity in the United States, has become increasingly imperative. This trial also involved a randomized design and included objective measures of the primary outcome and adherence metrics. Moreover, we conducted cost-effectiveness analyses. Such data are critical to public policy and to key stakeholders when they are deciding whether to implement large-scale intervention approaches. Finally, this trial included a longer-term follow-up, which allowed for examination of intervention effects on weight losses after treatment and underscored the importance of providing intervention beyond the 3-month weight loss period to better promote weight loss maintenance.
Our series of SURI trials have gradually moved from an efficacy focus to an effectiveness focus. Several components of the current trial were consistent with effectiveness research including the development of an evidence-based behavioral weight loss Web site with inherently wide reach, broad eligibility criteria, exclusion criteria primarily focused on safety, elimination of run-in procedures, minimal or optional treatment contact, and lack of targeted efforts to promote treatment adherence. Future efforts to further generalizability may include minimizing the involvement of the research team by offering orientations, assessments, and optional group sessions in the community. In addition, all participants in SURI who are interested in participating in the behavioral weight loss program could be offered access to this material, and the Web site for the behavioral weight loss intervention could be integrated into the SURI Web site. Such approaches would yield additional information on scalability and generalizability and further the dissemination efforts.
Adding a novel Internet behavioral weight loss program to a statewide community health initiative is a cost-effective approach to improving treatment outcomes and may, therefore, have substantial public health impact.
Acknowledgments
This study was supported by grant DK083248 from the National Institute of Diabetes and Digestive and Kidney Diseases.
We also wish to recognize the contributions of the following staff at the Weight Control and Diabetes Research Center at The Miriam Hospital and the Warren Alpert Medical School of Brown University: Sara Cournoyer, BA, Pamela Coward, Med, Linda Gay, RD, Kathryn McDermott, PhD, Deb Ranslow-Robles, Rachel Ogilvie, BA, Jessica Lawton, BA, Michelle Fisher, RN, Angelica McHugh, MA, and Kevin O’Leary, MS. We also recognize the contributions of the following staff at ShapeUp Rhode Island: Robert Vitek and Jenna Lafayette.
Human Participant Protection
This study was approved by The Miriam Hospital’s institutional review board.
