Identifying behavioural phenotypes among 1,084 healthy living app users.
Abstract
BACKGROUND: Low mobile health (mHealth) app engagement leading to little or no effect is typical. In partnership with WayBetter Inc., makers of the deposit contract-based WayBetter (WB) mHealth app, we aim to begin to address the low app engagement issue. PURPOSE: To identify homogenous subgroups (“behavioural phenotypes”) of new WB app users. METHODS: A cross-sectional study was conducted between November 2, 2023 and January 3, 2024. New WB app users creating an account during the study period and engaging in a seven-day free trial were recruited via email to complete a short 27-item study survey. Previously validated survey items purported to predict mHealth app engagement were included. Latent class analyses were conducted using survey responses to identify behavioural phenotypes. RESULTS: The total sample included 1084 users (age [mean, SD]=42.00, 9.65) years; % women=90.59). Four behavioural phenotypes were identified based on key model fit statistics (e.g., AIC=39681.76), the results of likelihood ratio tests (Vuong-Lo-Mendell-Rubin LRT p