Background

Some guidelines state that in-person weight management interventions are more efficacious than those delivered digitally. However, digital programs are more scalable and accessible. We hypothesized that one-on-one health coaching via app-based video chat would simulate an in-person experience and help achieve outcomes comparable to those of in-person interventions.

Methods

A digital weight management program was provided to overweight or obese adults recruited from a large technology company. One-on-one health coaching sessions were offered weekly for 16 weeks, and less frequently thereafter, until one year after baseline. Focused on sustainable changes in activity and diet, the program uses SMART goals, in-app food and activity logs, Fitbit integration, sleep and stress modules, as well as digital scales that transmit weight data automatically. Self-Determination Theory, the Transtheoretical Model, and Behavioral Economics were incorporated to drive behavior change. Machine learning was used to tailor content. Multilevel mixed-effects models were utilized to accommodate nested, time-unstructured data.

Results

Six hundred eighty-three participants recorded 29,051 weights. At 12 months, mean percent changes in body weight were -7.2% (95% CI, -8.8% to -5.7%, p < .01) and -7.6% (95% CI, -8.9% to -6.4%, p < .01) in overweight and obese groups, respectively. Mean rates of change were -0.25 (95% CI, -0.30 to -0.19; p < .01) and -0.34 pounds per week (95% CI, -0.40 to -0.28; p < .01; time ✕ group coefficient = -0.09; 95% CI, -0.17 to -0.01; p < .05). Rates of change were heterogeneous across participants (χ2 (1)=20,586, p < .01). Limitations include an observational, uncontrolled design, as well as survivorship bias.

Conclusions

Further research is needed with randomization to in-person or digital programs. Preliminary results suggest that some weight management programs with app-based coaching may achieve results comparable to those of robust, in-person interventions.