ABSTRACT
BACKGROUND: The growth in the capabilities of telehealth have made it possible to identify individuals with a higher risk of uncontrolled diabetes and provide them with targeted support and resources to help them manage their condition. Thus, predictive modeling has emerged as a valuable tool for the advancement of diabetes management. OBJECTIVE: This study aimed to conceptualize and develop a novel machine learning (ML) approach to proactively identify participants enrolled in a remote diabetes monitoring program (RDMP) who were at risk of uncontrolled diabetes at 12 months in the program. METHODS: Registry data from the Livongo for Diabetes RDMP were used to design separate dynamic predictive ML models to predict participant outcomes at each monthly checkpoint of the participants' program journey (month-n models) from the first day of onboarding (month-0 model) up to the 11th month (month-11 model). A participant's program journey began upon onboarding into the RDMP and monitoring their own blood glucose (BG) levels through the RDMP-provided BG meter. Each participant passed through 12 predicative models through their first year enrolled in the RDMP. Four categories of participant attributes (ie, survey data, BG data, medication fills, and health signals) were used for feature construction. The models were trained using the light gradient boosting machine and underwent hyperparameter tuning. The performance of the models was evaluated using standard metrics, including precision, recall, specificity, the area under the curve, the F1-score, and accuracy. RESULTS: The ML models exhibited strong performance, accurately identifying observable at-risk participants, with recall ranging from 70% to 94% and precision from 40% to 88% across the 12-month program journey. Unobservable at-risk participants also showed promising performance, with recall ranging from 61% to 82% and precision from 42% to 61%. Overall, model performance improved as participants progressed through their program journey, demonstrating the importance of engagement data in predicting long-term clinical outcomes. CONCLUSIONS: This study explored the Livongo for Diabetes RDMP participants' temporal and static attributes, identification of diabetes management patterns and characteristics, and their relationship to predict diabetes management outcomes. Proactive targeting ML models accurately identified participants at risk of uncontrolled diabetes with a high level of precision that was generalizable through future years within the RDMP. The ability to identify participants who are at risk at various time points throughout the program journey allows for personalized interventions to improve outcomes. This approach offers significant advancements in the feasibility of large-scale implementation in remote monitoring programs and can help prevent uncontrolled glycemic levels and diabetes-related complications. Future research should include the impact of significant changes that can affect a participant's diabetes management.
ABSTRACT
BACKGROUND: Physical inactivity is a global issue that affects people's health and productivity. With the advancement of mobile technologies, many apps have been developed to facilitate health self-management. However, few studies have examined the effectiveness of these mobile health (mHealth) apps in motivating exercise adherence. OBJECTIVE: This study aims to demonstrate the enhanced action design research (ADR) process and improve the design of mHealth apps for exercise self-management. Specifically, we investigate whether sending motivational messages improves adherence to exercise plans, whether the motivational effect is affected by personality, the impact of message type and repetition, and the process of involving a field experiment in the design process and learning new design principles from the results. METHODS: This formative research was conducted by proposing an enhanced ADR process. We incorporated a field experiment into the process to iteratively refine and evaluate the design until it converges into a final mHealth app. We used the Apple ResearchKit to develop the mHealth app and promoted it via trainers at their gyms. We targeted users who used the app for at least two months. Participants were randomly assigned to 1 of the 12 groups in a 2×3×2 factorial design and remained blinded to the assigned intervention. The groups were defined based on personality type (thinking or feeling), message type (emotional, logical, or none), and repetition (none or once). Participants with different personality types received tailored and repeated messages. Finally, we used the self-reported completion rate to measure participants' adherence level to exercise plans. By analyzing users' usage patterns, we could verify, correct, and enhance the mHealth app design principles. RESULTS: In total, 160 users downloaded the app, and 89 active participants remained during the 2-month period. The results suggest a significant main effect of personality type and repetition and a significant interaction effect between personality type and repetition. The adherence rate of people with feeling personality types was 18.15% higher than that of people with thinking types. Emotional messages were more effective than logical messages in motivating exercise adherence. Although people received repeated messages, they were more likely to adhere to exercise plans. With repeated reminders, the adherence rates of people with thinking personality types were significantly improved by 27.34% (P<.001). CONCLUSIONS: This study contributes to the literature on mHealth apps. By incorporating a field experiment into the ADR process, we demonstrate the benefit of combining design science and field experiments. This study also contributes to the research on mHealth apps. The principles learned from this study can be applied to improve the effectiveness of mHealth apps. The app design can be considered a foundation for the development of more advanced apps for specific diseases, such as diabetes and asthma, in future research.
