The Usage Effects, Effect Modifiers, and Experiences of a Web-Based App for Healthy Habit Formation in Adults: Exploratory Analysis of a Quasi-Experimental Real-World Intervention.

BACKGROUND: Digital interventions provide a scalable approach to promote healthy lifestyles, but at-scale evidence of their effects and effect modifiers is limited-even scarcer for habit-based digital interventions, although habits are promising lifestyle intervention targets. OBJECTIVE: We conducted a 1-group pre-post effect and effect modification assessment of a web-based app designed to support the formation of health-promoting lifestyle habits by exploring the relationship between objective app use and subjective changes in behavioral outcomes. METHODS: Three-month app access was offered to a subsample of the population-based Healthy Finland survey participants aged 20-74 years via SMS text messaging or mail. The app provided personalized behavioral suggestions that translated evidence-based lifestyle guidelines into simple, repeatable actions ("habits"). Users could browse and select these habits, and report and monitor their performances. The assessment used app log data, questionnaires collected through the app at baseline and at 45 and 90 days, and background information obtained from the national population register and the Healthy Finland Survey. Linear mixed effects models explored the (1) associations between app use and changes in self-reported diet quality (Healthy Diet Index), physical activity (metabolic equivalent of task hours per week), and BMI, and (2) modification of these associations by baseline characteristics related to sociodemographics, health, lifestyle, and e-service use. App use measures comprised the percentage of app use days (ie, days with logins/days of follow-up×100) and the number of reported habit performances (total, diet-related, physical activity-related, and BMI-related). RESULTS: Of 6975 invitees, 1282 (18.4%) accepted the invitation and 382 (5.5%; mean age 51, SD 15 y; women: n=263, 69%) completed the data collection required for the assessment. Over 90 days, the median percentage of app use days was 5.9% (IQR 3.3%-11%), and the total number of reported habit performances was 22 (IQR 3-68.5). A higher percentage of use days (by 10 percentage points) was associated with a 3.09 (95% CI 0.79-5.39) metabolic equivalent of task hours per week greater increase in physical activity and a 1-unit higher logarithmic number of reported performances with a 0.51 (95% CI 0.07-0.95) Healthy Diet Index point greater improvement in diet quality. Other associations between app use and outcomes were nonsignificant. Greater physical activity and a more positive attitude to e-services at baseline appeared to enhance the effect of app use on physical activity (P values <.05). The app received a median acceptability rating of 2.9 (IQR 2.4-3.3, scale: 1-4) and a median overall score of 7 (IQR 5-8, scale: 0-10). CONCLUSIONS: An inexpensive, low-intensity digital support for healthy habits could foster small beneficial lifestyle changes, but effects require user engagement and may depend on the individual. The findings warrant confirmation in more robust study designs and call for further research for identifying those most likely to benefit from such digital support.