Machine Learning Analysis to Identify Digital Behavioral Phenotypes for Engagement and Health Outcome Efficacy of an mHealth Intervention for Obesity: Randomized Controlled Trial

Abstract

Background: The digital health care community has been urged to enhance engagement and clinical outcomes by analyzing multidimensional digital phenotypes.

Objective: This study aims to use a machine learning approach to investigate the performance of multivariate phenotypes in predicting the engagement rate and health outcomes of digital cognitive behavioral therapy.

Methods: We leveraged both conventional phenotypes assessed by validated psychological questionnaires and multidimensional digital phenotypes within time-series data from a mobile app of 45 participants undergoing digital cognitive behavioral therapy for 8 weeks. We conducted a machine learning analysis to discriminate the important characteristics.

Results: A higher engagement rate was associated with higher weight loss at 8 weeks (r=-0.59; P<.001) and 24 weeks (r=-0.52; P=.001). Applying the machine learning approach, lower self-esteem on the conventional phenotype and higher in-app motivational measures on digital phenotypes commonly accounted for both engagement and health outcomes. In addition, 16 types of digital phenotypes (ie, lower intake of high-calorie food and evening snacks and higher interaction frequency with mentors) predicted engagement rates (mean R² 0.416, SD 0.006). The prediction of short-term weight change (mean R² 0.382, SD 0.015) was associated with 13 different digital phenotypes (ie, lower intake of high-calorie food and carbohydrate and higher intake of low-calorie food). Finally, 8 measures of digital phenotypes (ie, lower intake of carbohydrate and evening snacks and higher motivation) were associated with a long-term weight change (mean R² 0.590, SD 0.011).

Conclusions: Our findings successfully demonstrated how multiple psychological constructs, such as emotional, cognitive, behavioral, and motivational phenotypes, elucidate the mechanisms and clinical efficacy of a digital intervention using the machine learning method. Accordingly, our study designed an interpretable digital phenotype model, including multiple aspects of motivation before and during the intervention, predicting both engagement and clinical efficacy. This line of research may shed light on the development of advanced prevention and personalized digital therapeutics.

Trial registration: ClinicalTrials.gov NCT03465306; https://clinicaltrials.gov/ct2/show/NCT03465306.

Keywords: clinical efficacy; digital phenotype; in-app engagement; machine learning analysis; mobile phone.

Figure 1

A conceptual framework of mobile health components and examples of digital phenotypes.

Figure 2

Relationships between engagement and health outcomes. The health outcome larger than zero indicates weight loss compared with baseline.

Figure 3

Multivariate patterns of conventional and digital phenotypes for predicting engagement (red) as well as short-term (green) and long-term (blue) health outcomes. Points indicate the averaged β coefficients across 100 repetitions of net elastic analysis (see the Methods section for details). A positive β estimate of a phenotype indicates an association between the phenotype and higher in-app activities (engagement) or more weight loss (health outcomes). The points, which contain zero in the simulated 95% ranges, are omitted.

Figure 4

Common predictors between engagement and health outcomes for (A) health outcome (short term) versus engagement, (B) health outcome (long term) versus engagement, and (C) health outcome (long term) versus health outcome (short term). Each axis indicates the β estimate for predicting engagement and health outcomes. A positive β coefficient indicates a positive association with engagement but negative associations with health outcomes (weight changes).

Figure 5

Two examples of common predictors between short-term and long-term health outcomes: (A) carbohydrate intake and (B) confidence in digital phenotypes. VAS: Visual Analogue Scale.