Power JM, Hurley L, O’Shea NG, Nezami BT, Sciamanna C, Tate DF. Examining latent trajectories of participant engagement in a 12-month eHealth weight management intervention. Digital Health. 2026;12doi:10.1177/20552076261434062
This secondary data analysis examined data from a larger cluster randomized controlled trial that evaluated two internet-based weight loss programs in primary care compared to usual care over 12 months. The original trial included 27 providers and 550 adults with overweight or obesity. Patients were assigned to one of three groups: an online program, the same program with added feedback from their doctor, or usual care with a printed booklet. The two online programs were more effective than usual care. The current analysis focused on how participants engaged with self-monitoring tools for weight, diet, and physical activity over 12 months. The current analysis used latent class growth modeling (LCGM) to identify patterns of engagement and how these patterns related to weight loss. The study identified several engagement groups with different patterns of self-monitoring. “Sustained-engagers” made up 16% of the sample, where a group with little to no engagement also appeared, with “never-engagers” accounting for 22% of the sample. Nearly half of the participants were “low/declining-engagers” (48.5%), and they did not achieve 5% weight loss at 12 months. In contrast, “early-engagers” (13%) did reach clinically significant weight loss. While engagement decreased over time in both groups, “early-engagers” continued to engage at each timepoint, while “low/declining-engagers” dropped off, especially between months 6–12. This difference likely explains the gap in weight loss outcomes. Participants with a higher BMI at baseline were less likely to maintain strong engagement, suggesting that they may require more support. Older participants were more likely to stay engaged, while some racial groups were more likely to have low engagement. A key strength of the study was its data-driven approach, which captured real usage patterns across behaviors over 12 months. However, the sample size was modest (N = 363), with fewer participants having complete data for weight (n = 169, 46.6%), diet (n = 82, 22.6%), and activity (n = 98, 27.0%), resulting in small subgroup sizes. Future research could explore engagement using different tracking tools.