Background: The use of remote data capture for monitoring blood glucose and supporting digital apps is becoming the norm in diabetes care. One common goal of such apps is to increase user awareness and engagement with their day-to-day health-related behaviors (digital engagement) in order to improve diabetes outcomes. However, we lack a deep understanding of the complicated association between digital engagement and diabetes outcomes. Objective: This study investigated the association between digital engagement (operationalized as tagging of behaviors alongside glucose measurements) and the monthly average blood glucose level in persons with type 2 diabetes during the first year of managing their diabetes with a digital chronic disease management platform. We hypothesize that during the first 6 months, blood glucose levels will drop faster and further in patients with increased digital engagement and that difference in outcomes will persist for the remainder of the year. Finally, we hypothesize that disaggregated between- and within-person variabilities in digital engagement will predict individual-level changes in blood glucose levels. Methods: This retrospective real-world analysis followed 998 people with type 2 diabetes who regularly tracked their blood glucose levels with the Dario digital therapeutics platform for chronic diseases. Subjects included “nontaggers” (users who rarely or never used app features to notice and track mealtime, food, exercise, mood, and location, n=585) and “taggers” (users who used these features, n=413) representing increased digital engagement. Within- and between-person variabilities in tagging behavior were disaggregated to reveal the association between tagging behavior and blood glucose levels. The associations between an individual’s tagging behavior in a given month and the monthly average blood glucose level in the following month were analyzed for quasicausal effects. A generalized mixed piecewise statistical framework was applied throughout. Results: Analysis revealed significant improvement in the monthly average blood glucose level during the first 6 months (t=−10.01, P<.001), which was maintained during the following 6 months (t=−1.54, P=.12). Moreover, taggers demonstrated a significantly steeper improvement in the initial period relative to nontaggers (t=2.15, P=.03). Additional findings included a within-user quasicausal nonlinear link between tagging behavior and glucose control improvement with a 1-month lag. More specifically, increased tagging behavior in any given month resulted in a 43% improvement in glucose levels in the next month up to a person-specific average in tagging intensity (t=−11.02, P<.001). Above that within-person mean level of digital engagement, glucose levels remained stable but did not show additional improvement with increased tagging (t=0.82, P=.41). When assessed alongside within-person effects, between-person changes in tagging behavior were not associated with changes in monthly average glucose levels (t=1.30, P=.20). Conclusions: This study sheds light on the source of the association between user engagement with a diabetes tracking app and the clinical condition, highlighting the importance of within-person changes versus between-person differences. Our findings underscore the need for and provide a basis for a personalized approach to digital health.

Background: Depression and anxiety are the main sources of work and social disabilities as well as health-related problems around the world. Digital therapeutic solutions using cognitive behavioral therapy have demonstrated efficacy in depression and anxiety. A common goal of digital health apps is to increase user digital engagement to improve outcomes. However, there is a limited understanding of the association between digital platform components and clinical outcomes. Objective: The aim of the study is to investigate the contribution of specific digital engagement tools to mental health conditions. We hypothesized that participation in coaching sessions and breathing exercises would be associated with a reduction in depression and anxiety. Methods: Depression and general anxiety symptoms were evaluated in real-world data cohorts using the digital health platform for digital intervention and monitoring change. This retrospective real-world analysis of users on a mobile platform–based treatment followed two cohorts of people: (1) users who started with moderate levels of depression and completed at least 2 depression assessments (n=519) and (2) users who started with moderate levels of anxiety and completed at least 2 anxiety assessments (n=474). Levels of depression (Patient Health Questionnaire-9) and anxiety (Generalized Anxiety Disorder-7) were tracked throughout the first 16 weeks. A piecewise mixed-effects model was applied to model the trajectories of the Patient Health Questionnaire-9 and the Generalized Anxiety Disorder-7 mean scores in 2 segments (1-6 weeks and 7-16 weeks). Finally, simple slope analysis was used for the interpretation of the interactions probing the moderators: coaching sessions and breathing exercises in both depression and anxiety cohorts. Results: Analysis revealed a significant decrease in depression symptoms (β=–.37, 95% CI –0.46 to 0.28; P≤.001) during the period of weeks 1-6 of app use, which was maintained during the period of 7-16 weeks. Coach interaction significantly moderated the reduction in depression symptoms during the period of weeks 1-6 (β=–.03, 95% CI –0.05 to –0.001; P=.02). A significant decrease in anxiety symptoms (β=–.41, 95% CI –0.50 to –0.33; P≤.001) was revealed during the period of 1-6 weeks, which was maintained during the period of 7-16 weeks. Breathing exercises significantly moderated the reduction in anxiety symptoms during the period of 1-6 weeks (β=–.07, 95% CI –0.14 to –0.01; P=.04). Conclusions: This study demonstrated general improvement followed by a period of stability of depression and anxiety symptoms associated with cognitive behavioral therapy–based digital intervention. Interestingly, engagement with a coaching session but not a breathing exercise was associated with a reduction in depression symptoms. Moreover, breathing exercise but not engagement with a coaching session was associated with a reduction of anxiety symptoms. These findings emphasize the importance of using a personalized approach to behavioral health during digital health interventions.

Background: Remote data capture for blood glucose (BG) or blood pressure (BP) monitoring and the use of a supportive digital app are becoming the model in diabetes and hypertension chronic care. One of the goals in chronic condition management is to increase awareness and generate behavioral change in order to improve outcomes in diabetes and related comorbidities, such as hypertension. In addition, there is a lack of understanding of the association between BG and BP levels when using digital health tools. Objective: By applying a rigorous study framework to digital health data, this study investigated the relationship between BP monitoring and BG and BP levels, as well as a lagged association between BP and BG. We hypothesized that during the first 6 months of BP monitoring, BG and BP levels would decrease. Finally, we suggested a positive association between BP levels and the following month’s BG levels. Methods: In this retrospective, real-world case-control study, we extracted the data of 269 people with type 2 diabetes (T2D) who tracked their BG levels using the Dario digital platform for a chronic condition. We analyzed the digital data of the users who, in addition to BG, monitored their BP using the same app (BP-monitoring [BPM] group, n=137) 6 months before and after starting their BP monitoring. Propensity score matching established a control group, no blood pressure monitoring (NBPM, n=132), matched on demographic and baseline clinical measures to the BPM group. A piecewise mixed model was used for analyzing the time trajectories of BG, BP, and their lagged association. Results: Analysis revealed a significant difference in BG time trajectories associated with BP monitoring in BPM and NBPM groups (t=–2.12, P=.03). The BPM group demonstrated BG reduction improvement in the monthly average BG levels during the first 6 months (t=–3.57, P<.001), while BG did not change for the NBPM group (t=0.39, P=.70). Both groups showed similarly stable BG time trajectories (B=0.98, t=1.16, P=.25) before starting the use of the BP-monitoring system. In addition, the BPM group showed a significant reduction in systolic (t=–6.42, P<.001) and diastolic (t=–4.80, P<.001) BP during the first 6 months of BP monitoring. Finally, BG levels were positively associated with systolic (B=0.24, t=2.77, P=.001) and diastolic (B=0.30, t=2.41, P=.02) BP. Conclusions: The results of this study shed light on the association between BG and BP levels and on the role of BP self-monitoring in diabetes management. Our findings also underscore the need and provide a basis for a comprehensive approach to understanding the mechanism of BP regulation associated with BG.

Introduction: Digital therapeutics (DT) emerged and has been expanding rapidly for pain management. However, the efficacy of such approaches demonstrates substantial heterogeneity. Machine learning (ML) approaches provide a great opportunity for personalizing the efficacy of DT. However, the ML model accuracy is mainly associated with reduced clinical interpretability. Moreover, classical ML models are not adapted for the longitudinal nature of the DT follow-up data, which may also include nonlinear fluctuations. Objectives: This study presents an analytical framework for personalized pain management using piecewise mixed-effects model trees, considering the data dependencies, nonlinear trajectories, and boosting model interpretability. Methods: We demonstrated the implementation of the model with posture biofeedback training data of 3610 users collected during 8 weeks. The users reported their pain levels and posture quality. We developed personalized models for nonlinear time-related fluctuations of pain levels, posture quality, and weekly training duration using age, gender, and body mass index as potential moderating factors. Results: Pain levels and posture quality demonstrated strong improvement during the first 3 weeks of the training, followed by a sustained pattern. The age of the users moderated the time fluctuations in pain levels, whereas age and gender interactively moderated the trajectories in the posture quality. Train duration increased during the first 3 weeks only for older users, whereas all the users decreased the training duration during the next 5 weeks. Conclusions: This analytical framework offers an opportunity for investigating the personalized efficacy of digital therapeutics for pain management, taking into account users' characteristics and boosting interpretability and can benefit from including more users' characteristics.

Background: The management of type 2 diabetes (T2D) and obesity, particularly in the context of self-monitoring, remains a critical challenge in health care. As nearly 80% to 90% of patients with T2D have overweight or obesity, there is a compelling need for interventions that can effectively manage both conditions simultaneously. One of the goals in managing chronic conditions is to increase awareness and generate behavioral change to improve outcomes in diabetes and related comorbidities, such as overweight or obesity. There is a lack of real-life evidence to test the impact of self-monitoring of weight on glycemic outcomes and its underlying mechanisms. Objective: This study aims to assess the efficacy of digital self-monitoring of weight on blood glucose (BG) levels during diabetes management, investigating whether the weight changes may drive glucose fluctuations. Methods: In this retrospective, real-world quasi-randomized study, 50% of the individuals who regularly used the weight monitoring (WM) feature were propensity score matched with 50% of the users who did not use the weight monitoring feature (NWM) based on demographic and clinical characteristics. All the patients were diagnosed with T2D and tracked their BG levels. We analyzed monthly aggregated data 6 months before and after starting their weight monitoring. A piecewise mixed model was used for analyzing the time trajectories of BG and weight as well as exploring the disaggregation effect of between- and within-patient lagged effects of weight on BG. Results: The WM group exhibited a significant reduction in BG levels post intervention (P<.001), whereas the nonmonitoring group showed no significant changes (P=.59), and both groups showed no differences in BG pattern before the intervention (P=.59). Furthermore, the WM group achieved a meaningful decrease in BMI (P<.001). Finally, both within-patient (P<.001) and between-patient (P=.008) weight variability was positively associated with BG levels. However, 1-month lagged back BMI was not associated with BG levels (P=.36). Conclusions: This study highlights the substantial benefits of self-monitoring of weight in managing BG levels in patients with diabetes, facilitated by a digital health platform, and advocates for the integration of digital self-monitoring tools in chronic disease management. We also provide initial evidence of testing the underlying mechanisms associated with BG management, underscoring the potential role of patient empowerment.

Significant racial/ethnic disparities in healthcare and diabetes technology use have been observed in Type 2 diabetes mellitus (T2DM), which are associated with nonengagement in diabetes self-management and out-of-range glycemia. This study aimed to assess whether there were differences in the blood glucose levels achieved by several racial/ethnic groups using the same digital tool. Study objectives were to determine whether engagement with the digital tool and blood glucose levels differ among ethnic groups, and to determine whether any differences in the in-target-glycemia are related to engagement levels. The retrospective real-world analysis followed a group of 1000 people with Type 2 diabetes who used the Dario^TM digital therapeutic platform over 12 months. Participants included in the study had a blood glucose average > 180 mg/dL (hyperglycemia, high-risk) in their first month. The differences between/within the groups’ average blood glucose level (Avg.bg) and glycemic variability were evaluated. Furthermore, three general linear models were constructed to predict the Avg.bg by the number of blood glucose measurements (Bgm) in Model 1 (with the moderator White persons (WP)/people from racial and ethnic minority groups (REM)) and by the frequency of measurements by months (F.m) within REM and WP in Model 2 and Model 3, respectively. The Avg.bg was significantly reduced in each group over a year with no differences between REM/WP users. Blood glucose measurements in Model 1 and frequency of measurements by months in Model 2 and Model 3 predicted the Avg.bg (β₁ = −0.20, p = 0.045; β₂ = −4.38, p = 0.009; β₃= −3.77, p < 0.001, respectively). Findings indicate a positive association between digital engagement and glycemia, with no differences between REM and WP participants.