INTRODUCTION

Adding financial incentives to a structured behavioral weight-control program has been shown to increase weight losses relative to the same weight-control approach without financial incentives, although benefits are better established for weight-loss induction than for weight maintenance ((1-3)). Given the challenges faced by the obesity treatment field in promoting prolonged weight control ((4)), as well as the significant health benefits that accrue with sustained weight loss ((5, 6)), identifying strategies to enhance long-term weight outcomes is critical.

Few studies, to our knowledge, have examined financial incentives to promote weight-loss maintenance, and most that have explored this approach to enhancing weight maintenance have enrolled individuals who had previously lost weight (in programs that did not offer incentives for weight-loss induction). In these studies, the addition of incentives did not promote better weight-loss maintenance than the program without incentives ((7-9)). However, better weight-loss maintenance was observed with the addition of financial incentives and coaching compared with a low-contact, health education behavioral approach ((10)).

Contemporary studies integrating financial incentives for both initial weight-loss induction and extended weight management within the same obesity treatment program have shown promise. In one study, those receiving financial incentives for achieving target weight-loss outcomes and performing weight-control behaviors both during and after a structured in-person program achieved significantly, albeit modestly, better 12-month weight losses (−2.3 kg for the incentivized program vs. −0.8 kg for the structured program alone) ((11)). In a text-based weight-loss intervention, Dombrowski and colleagues similarly found better 12-month weight outcomes with the addition of financial incentives relative to the text intervention alone (−2.5 kg for text plus incentives vs. −1.29 kg for the text alone) ((12)). However, others have found that offering financial incentives for target weight outcomes and weight-management behaviors as part of an in-person program did not produce better weight losses than achieved with the program alone ((13, 14)). Therefore, it remains unclear whether the addition of incentives to behavioral weight control meaningfully improves long-term weight outcomes and whether they should be considered as part of comprehensive behavioral obesity treatment.

Another issue surrounding the use of financial incentives in weight control is how weight-control behaviors and weight outcomes are impacted when incentives are discontinued. From the few studies that have followed individuals after financial incentives were terminated, it appears that any advantages accrued when incentives were provided dissipated once incentives cease ((7, 8, 15, 16)). Consequently, the sustainability of weight control and healthy lifestyle behaviors produced with incentives may be problematic, but the literature examining this important factor is sparse.

Therefore, the objectives of the current study were twofold: 1) to examine the impact of an integrated incentive scheme that traverses weight-loss induction and weight-loss maintenance with explicit weight-loss and weight-maintenance incentive targets; and 2) to explore what happens to weight outcomes and critical weight-management behaviors once financial incentives end.

METHODS Study design

This parallel randomized controlled trial allocated participants from two clinical sites (South Carolina and Vermont) in a 1:1 ratio to either of the following groups: 1) an 18-month, online, group-based behavioral lifestyle program with synchronous group text-based chat sessions (Internet-Only); or 2) the same 18-month, online, group-based program with financial incentives available contingent on key weight-control behaviors and achieving target weight-loss outcomes (Months 1-6) and weight stability goals (Months 7-12; Internet+Incentives). The initial 6-month outcomes of the weight-loss induction period have been reported previously ((17)); the current report extends this earlier work, focusing on weight-loss outcomes across the full incentive scheme, spanning the weight-loss induction phase (Months 1-6) and weight-maintenance phase (Months 7-18), and including a 6-month period without financial incentives (Months 13-18).

The study was approved by the Institutional Review Board at the University of South Carolina and Committee on Human Research in the Behavioral Sciences at the University of Vermont. This trial is registered at www.ClinicalTrials.gov (NCT02688621), and written informed consent was obtained from all participants.

Participants

Participants were recruited over a period of 24 months from 2016 to 2018 through community flyers, targeted emails using available distribution lists (e.g., worksites, professional organizations, sororities), and word of mouth. In order to be eligible, participants had to be 18 years or older, have a BMI between 25 and 50 kg/m2, have access to a computer (at home or work), the Internet, and a smartphone, and complete 3 consecutive days of dietary self-monitoring. Individuals were ineligible if they reported substantial recent weight loss, a history of bariatric surgery, enrollment in another weight-management program, taking medications that could affect weight, or a condition for which weight loss and/or exercise was contraindicated. Participants indicated availability for predetermined group times; these intact groups were randomized using a biased-coin approach by the University of Vermont Biostatistical Coordinating Center. Sample size was calculated to detect a between-conditions difference of 2.2 kg at 6 months.

Online, group-based behavioral weight-loss intervention

The 18-month, manualized online intervention reflected approaches used in other evidence-based interventions ((18-21)) (see online Supporting Information for a detailed description of the intervention). All participants were given access to a password-protected website that provided dynamic components, including behavioral lessons, a bulletin board for private group communications, personalized real-time progress graphs, and educational resources. One-hour, online, synchronous, text-based chat sessions focused on reinforcing behavioral skills and cultivating group support were facilitated by an experienced behavioral weight-control interventionist. Group sessions were offered weekly during weight-loss induction (Months 1-6) and monthly during weight maintenance (Months 7-18), providing facilitated, synchronous group chat sessions focused first on achieving weight loss and then on cultivating weight stability skills.

All participants were instructed to weigh themselves daily and record their weight on the study website each day; if they did not have a scale, the study provided one. Participants were prescribed a daily calorie goal ranging from 1,200 to 1,800 kcal based on baseline weight. Participants were also given graded moderate-to-vigorous physical activity (MVPA) goals that progressed to 200 min/wk and graded step goals that progressed to 10,000 steps/d ((22)). They were instructed to monitor dietary intake, minutes of physical activity, and number of steps daily using the MyFitnessPal smartphone app and to record calorie intake, minutes of MVPA, total steps, and body weight on the study website daily. The interventionist emailed tailored feedback based on participant self-monitoring.

Internet+Incentive condition

Individuals randomized to Internet+Incentives received the same online, group-based behavioral intervention as the Internet-Only condition, with the same behavioral goals, lesson content, Web-based platform, and contact schedule. Only the availability of financial incentives differed between conditions.

Overall incentive scheme

Both behavioral (process) incentives and weight-loss (outcome) incentives were offered in an incentive scheme that traversed weight-loss induction and weight-loss maintenance (Figure 1). Incentives for achieving clinically meaningful weight-loss milestones at Months 2 and 6 were offered. Process incentives were offered for performing key self-regulatory behaviors during the early stages of weight-loss induction; specifically, individuals were rewarded weekly for consistent self-weighing and dietary self-monitoring. Importantly, self-monitoring behaviors were reinforced rather than caloric intake. After Month 6, incentive targets shifted from weight loss to weight stability; participants were paid at Month 12 if their weight was the same (or lower) than their Month 6 weight, and process incentives were offered for self-weighing and physical activity engagement (i.e., step targets). Incentives were discontinued during the final 6 months of maintenance (Months 12-18), allowing for observation of the durability of self-management behaviors and weight trajectory in the absence of financial incentives. Thus, the incentive scheme combined both outcome and process incentive targets and spanned the full range of weight-loss induction and weight-loss maintenance, including a period after incentives terminated. Incentives were paid out using electronic gift cards, with weekly payouts for process incentives and immediate payment for outcome incentives.

Process and outcome incentive scheme [Color figure can be viewed at wileyonlinelibrary.com] Weight-outcome incentives

Incentives in escalating amounts were offered for achieving weight-loss milestone targets at Months 2 and 6. Achieving clinically meaningful weight loss as early as 2 months foreshadows long-term weight-loss success up to 8 years later ((23)); therefore, incentives for meeting target weight-loss thresholds were offered at month 2 ($35 for ≥5% weight loss; $25 for ≥3%) and month 6 ($75 for ≥10% weight loss; $50 for ≥5%). This tiered-incentive approach allowed for greater reward for the higher weight target while also offering incentives for meaningful but lower weight-loss thresholds. When the incentive scheme shifted to rewarding weight stability (Months 7-12), participants received $75 at the 12-month assessment if they were below their baseline weight and had maintained their 6-month weight (or lost more weight). No outcome incentive was available for the final 18-month weight measurement to test weight maintenance in the absence of incentives.

Process incentives

Treatment engagement in the first 2 months of a behavioral weight-control program is considered pivotal for continued weight-loss success ((24)); therefore, incentives were offered weekly during the first 8 weeks for daily self-weighing and dietary self-monitoring, both of which have a strong empirical relationship with better weight loss ((25-27)). Process incentives were awarded at the end of each week if self-report of targeted weight-control behavior thresholds were met in the previous week. In order to reflect the well-established dose-response relationship between self-regulatory behaviors, participants could receive partial incentive payouts for achieving the behavioral target on at least 5 days even if they did not meet criteria for the full incentive. Emails announcing the incentive earned were sent weekly; consistent with prospect theory, a “loss-framed” message ((28)) was sent to those who had not earned the full incentive and, therefore, had “left money on the table,” encouraging them to strive for the full incentive the following week. Each week was independent; therefore, individuals who failed to get an incentive one week had another chance the following week. No process incentives were offered in Months 3 to 6 to conserve funds; this program design decision was driven by evidence that earlier behavioral engagement appears most critical for future weight loss ((23, 24)) and an awareness that outcome incentives were in place to encourage continued engagement.

Weight monitoring and physical activity were selected as process incentive targets during the maintenance phase because both contribute significantly to successful weight outcomes ((29-32)). Incentives for daily self-weighing and achieving 10,000 steps/d started out on a fixed weekly schedule during the first 12 weeks of weight maintenance (Months 7-9), followed by a less predictable, variable payout schedule during Months 10 to 12 to accrue the benefits of a variable reinforcement schedule for promoting persistent behavior patterns ((33)). The density of payouts was structured to gradually fade over time (i.e., payouts given in a randomly selected 3 of the 4 weeks, then 2 of the next 4 weeks, and so on). Participants were not told which weeks were incentivized during the intermittent reinforcement period but were told at the start of weight maintenance they would not be getting incentives every week. As before, weekly email messages were sent either announcing the incentive earned that week (and money left behind by not engaging fully in the targeted self-management behaviors, if applicable) or announcing that although it was not an incentive week, the upcoming week might be.

Total financial incentive

The maximum total incentive possible over the full program was $665. During weight-loss induction, $230 was available ($110 for outcome incentives and $120 for process incentives). During weight maintenance, participants could earn up to $435 ($75 for outcome incentives and $360 for process incentives).

Outcome measures

Data were collected between February 2016 and July 2019 by assessors blinded to treatment assignment.

Body weight

Weight was measured in street clothes, without shoes, on a calibrated digital scale in the clinic at each time point. Height was assessed using a wall-mounted stadiometer at baseline. Percentage of weight loss from baseline was calculated at each clinic assessment visit. Weight stability from Month 6 to Months 12 and 18 was calculated using the criterion established by the incentive scheme (i.e., the proportion of those who were at or below their 6-month weight). The proportion of individuals achieving clinically meaningful weight loss of ≥5% and ≥10% ((34)) was also examined.

Sociodemographic characteristics

Self-reported sociodemographic characteristics were collected at baseline.

Treatment engagement

Interventionists documented attendance at online group chat sessions. Participants were asked to report daily on the study website whether they logged their foods/beverages, met their calorie goal, how many minutes of MVPA they completed, number of steps taken, and their weight. These self-monitoring entries were used to characterize treatment engagement in both study arms and to determine incentive payments for participants randomized to Internet+Incentive.

Statistical analysis

The primary focus of this analysis was to explore whether an integrated incentive scheme targeting both weight-loss induction and weight maintenance, with both outcome and process incentives, would improve weight stability and total weight loss at month 12 compared with the online program alone. Additionally, we examined weight-loss outcomes in the final 6 months of the maintenance period when no incentives were offered. Missing weight data were imputed by the fully conditional specification method (FCS option in SAS PROC MI, SAS Institute). Multiple imputation was used to create 100 complete data sets that were analyzed separately, and then the results were combined. Month 2 values were imputed with baseline weight, chat group, age, baseline BMI, sex, and race as covariates ((35)). Subsequent outcome points followed the same process, using the same covariates, as well as weight from previous assessments (both observed and imputed). Imputation was done separately for each arm ((36)). Weight-change analyses were conducted using a mixed-effects linear model, with repeated measures nested within participants and participants nested within randomized clusters (i.e., online chat groups). Sensitivity analyses were conducted in similar fashion using only those with weight data available (completers). Similar analyses were run to evaluate adherence measures (e.g., number of days reporting weight, number of steps per day). Missing adherence data were treated as indicating nonadherence. All analyses were conducted using SAS version 9.4 (SAS Institute). Statistical significance was defined as p < 0.05 (two-tailed).

RESULTS Participants

A total of 418 participants were randomized (Figure 2). Participants were predominantly female, most had obesity, and 28% self-identified as African American or another racial minority group. There were no significant differences between conditions with respect to baseline characteristics (Table 1). Participants in 26 intervention groups were randomized (13 groups per condition). Retention at the final visit was 74%, with no significant differences in attrition between conditions. Participants who failed to provide 18-month outcome data were significantly younger (46 [11] years vs. 50 [11] years, t = 3.64, p < 0.001) and less likely to be college educated (69% vs. 83%, χ2 = 9.30, p = 0.002). No study-related serious adverse events were reported.

CONSORT (Consolidated Standards of Reporting Trials) diagram

TABLE 1. Baseline sample characteristics Internet-Only,n = 212 Internet+Incentives,n = 206 Age (y), mean ± SD 48 ± 11 50 ± 11 Female, n (%) 194 (92%) 187 (91%) Race, n (%) White 158 (77%) 145 (68%) African American 63 (30%) 44 (21%) Other 4 (2%) 4 (2%) Weight (kg), mean ± SD 97.8 ± 19.5 95.6 ± 15.9 BMI, mean ± SD 35.8 ± 5.9 35.5 ± 5.5 Obesity (BMI ≥30 BMI), n (%) 174 (82%) 169 (82%) College degree or higher, n (%) 169 (80%) 161 (78%) Employed full-time, (n (%) 177 (83%) 169 (82%) Retention, n (%) Month 2 194 (92%) 191 (95%) Month 6 172 (81%) 188 (91%)* Month 12 148 (70%) 162 (79%) Month 18 146 (69%) 152 (74%) Weight change over the 18-month program

As previously reported ((17)), the provision of financial incentives during weight-loss induction (Months 1-6) was associated with greater average weight loss than was produced with the online program alone in both the intent-to-treat and completers analyses (Table 2). A significantly greater proportion of those offered incentives achieved clinically meaningful weight-loss thresholds during weight-loss induction when examining the overall sample, but this advantage was not statistically significant among completers. Significantly greater total weight loss was no longer apparent for Internet+Incentives at Month 12 despite the continued provision of financial incentives for both behavioral targets and weight-maintenance benchmarks. However, significantly more Internet+Incentives participants achieved weight loss of ≥5% and ≥10% at Month 12 than did Internet-Only participants (in intent-to-treat but not completers analyses), and a significantly greater proportion of Internet+Incentives participants were weight stable from Months 6 to 12 in intent-to-treat analyses (Table 2). Nevertheless, despite these early indications of better overall weight maintenance among Internet+Incentives participants, by the end of treatment (Month 18) and following 6 months of no incentives, Internet+Incentives did not differ significantly from Internet-Only in any weight-loss outcomes (Table 2). Indeed, those in Internet+Incentives who were still participating (i.e., completers) gained significantly more weight from Months 12 to 18 than did those in Internet-Only.

TABLE 2. Weight losses All randomized participantsa Completers Internet-Only Internet+ Incentives p value Internet-Only Internet+ Incentives p value Total weight loss from baseline (kg), mean ± SD Month 2b 3.1 ± 3.4 4.3 ± 3.0 0.001 3.2 ± 3.4 4.3 ± 3.0 0.001 Month 6b 4.7 ± 6.6 6.4 ± 5.5 0.01 5.2 ± 6.5 6.7 ± 5.4 0.05 Month 12c 4.2 ± 7.4 5.8 ± 7.5 0.10 5.1 ± 7.5 6.8 ± 7.4 0.07 Month 18 3.5 ± 8.0 4.4 ± 7.7 0.38 4.5 ± 7.8 5.1 ± 7.2 0.41 Total weight loss from baseline (%), mean ± SD Month 2b 3.2 ± 3.1 4.5 ± 3.1 0.001 3.3 ± 3.1 4.5 ± 3.0 0.001 Month 6b 4.9 ± 6.4 6.8 ± 5.7 0.01 5.4 ± 6.1 7.1 ± 5.5 0.04 Month 12c 4.3 ± 7.5 6.2 ± 7.9 0.10 5.3 ± 7.3 7.3 ± 7.7 0.07 Month 18 3.5 ± 7.9 4.8 ± 8.1 0.26 4.6 ± 7.4 5.6 ± 7.6 0.23 Achieved clinically meaningful weight-loss thresholds ≥5% weight loss, n (%) Month 2b 46 (22%) 86 (42%) 0.01 46 (24%) 86 (44%) 0.001 Month 6b 85 (40%) 114 (55%) 0.01 85 (49%) 114 (61%) 0.06 Month 12 67 (32%) 92 (45%) 0.02 67 (45%) 92 (57%) 0.07 Month 18 54 (25%) 68 (33%) 0.10 54 (37%) 68 (45%) 0.16 ≥10% weight loss, n (%) Month 6b 34 (16%) 58 (28%) 0.02 34 (20%) 58 (31%) 0.06 Month 12 32 (15%) 52 (25%) 0.03 32 (22%) 52 (32%) 0.07 Month 18 27 (13%) 35 (17%) 0.28 27 (18%) 35 (23%) 0.33 Weight maintenance No regain from Month 6 (observed weight ≤ Month 6 weight), n (%) Month 12c 57 (27%) 81 (39%) 0.01 57 (40%) 81 (50%) 0.07 Month 18 50 (24%) 51 (25%) 0.79 50 (35%) 51 (34%) 0.83 Total weight regain from Month 6 (kg), mean ± SD Month 12 +0.5 ± 4.6 +0.7 ± 4.5 0.81 +0.5 ± 4.6 +0.4 ± 4.4 0.90 Month 18 +0.7 ± 3.1 +1.4 ± 3.1 0.15 +0.7 ± 2.9 +1.5 ± 3.0 0.04 Treatment engagement over the 18-month program

Targeted weight-control behaviors were reported more frequently by Internet+Incentives participants while process incentives were being provided, with some spillover into periods when there was no incentive for a particular behavior but when weight-outcome incentives were still offered (Table 3). For example, dietary self-monitoring was higher in the initial 6 months (when incentives for recording intake were available) among those in Internet+Incentives compared with Internet-Only, and greater tracking continued into Months 7 to 12 even though dietary monitoring was no longer an incentivized behavior. Similarly, those in Internet+Incentives met the goal of ≥50,000 steps/wk early (Months 1-6) before incentives were offered for that behavior. In the final 6 months of the program, when financial incentives were no longer available, the conditions did not differ in frequency of engaging in core self-regulatory weight-control behaviors. However, significantly greater declines were observed from Month 12 to Month 18 among those in Internet+Incentives compared with Internet-Only; the steeper decline in adherence was evident for self-weighing (p < 0.0001), dietary self-monitoring (p < 0.0001), and weeks with ≥50,000 steps (p < 0.0001).

TABLE 3. Treatment engagement Internet-Only Internet+Incentives p value Chat attendance Number of monthly group chat sessions attended Months 1-6 (out of 24 sessions) 16 ± 6.6 (67%) 18 ± 5.1 (75%) <0.0001 Months 7-12 (out of 6 sessions) 2.4 ± 2.2 (40%) 3.3 ± 1.9 (55%) 0.008 Months 13-18 (out of 6 sessions) 1.7 ± 2.1 (28%) 2.1 ± 2.0 (35%) 0.32 Incentivized weight-control behaviors Number of days reported weight Months 1-6 (out of 168 days)a 72 ± 51 (43%) 115 ± 43 (68%) <0.0001 Months 7-12 (out of 183 days)b 29 ± 50 (16%)