Effect Of Personalized Feedback Messages On Adherence To The Dietary, Physical Activity And Weight Loss Goals- The Smarter RCT

Background: A core component of standard behavior treatment for weight loss is establishing goals for caloric restriction and weekly minutes of physical activity (PA). Adherence to these behavior change goals is essential for successful weight loss outcomes. The aim of this paper is to examine adherence to the behavior change goals in the two treatment conditions: Self-Monitoring plus personalized Feedback (SM+FB) or SM alone. Methods: We conducted a single-site, two-arm randomized controlled trial from August 2018 to March 2021. Adults (N=502) 18 years of age or over with body mass index (BMI) 27-43 kg/m 2 were randomized with equal allocation to SM+FB (n=251) or SM (n=251). Feedback messages were sent daily for diet and every other day for PA. Adherence to the dietary and PA goals were defined as: 1) for diet the number of days one was adherent to the calorie goal (i.e., between 80% and 120% with the wide range due to the lack of precision in self-reported food intake) and 2) for PA the percentage of minutes of fairly and very active minutes per week meeting the PA goal of at least 150 minutes for fairly and very active minutes based on the Fitbit Charge 2™. Associations between treatment assignment and adherence to the dietary and PA goals were examined using linear mixed modeling with intention-to-treat method. Results: The sample (N=502) was mostly white (84%, 423 of 502), female (79%, 399 of 502) and on average 45.0±14.4 years of age and BMI of 33.7±4.0 kg/m 2 . Study retention at 12 months was 79%. The trajectory of change in adherence to the dietary goal was non-linear over 12 months (b=-.18, 95% CI [-.20, -.17]; b 2 =.006, 95% CI [.005, .007]; b 3 =-.0001, 95% CI [-.00007, -.00006], all P<.0001). The rate of decline over time in adherence to the diet goal was nonlinear and different by treatment (b trt*time = .03, 95% CI [.01, .06]; b trt*time*time = -.002, 95% CI [-.003, -.001]; b trt*time*time*time = .00003, 95% CI [.00002, .00004], all P<.05) Notably, the effect of treatment (b= .41, 95% CI [.10, .72], P=.009) on adherence to the dietary goal was significant. Additionally, the effects of time with polynomial contrasts showed declining curvilinear trajectories of change in the PA goal over 12 months (b=-1.75, 95% CI [-2.57, -0.92], b 2 =.08, 95% CI [.04, .11], b 3 =-.001, 95% CI [-.002, -.001], all P<.0001). The rate of decline over time in adherence to the PA goal was non-linear and different by treatment (b trt*time = -.80, 95% CI [-1.40, -.20]; b trt*time*time = .02, 95% CI [.004, -.03], all P<.05); the effect of treatment on adherence to the PA goal was not significant (b= 19.07, 95% CI [-3.86, 42.01], P = .103). Conclusions: In this study, adherence to the dietary and PA goals was associated with daily FB, which may be related to the frequency of the feedback messages. Further studies are needed to determine the best dose of tailored feedback messages to achieve sustained adherence to the prescribed behavior change goals.