The Impact of an Adaptive mHealth Intervention on Improving Patient-Provider Health Care Communication: Secondary Analysis of the DIAMANTE Trial.

Background: Depression and diabetes are highly comorbid conditions and are significant causes of global disability, particularly among individuals with low income or those from racial or ethnic minority backgrounds. While digital interventions offer promise for managing these chronic conditions (such as via lifestyle modification), there is also emerging evidence suggesting that digital support may strengthen or complement existing health care relationships, particularly by improving patient perceptions of communication and connection with their health care providers. Objective: This study aimed to examine the impact of an adaptive mobile health (mHealth) texting-based intervention on patient ratings of communication with their health care providers among individuals with diabetes and depressive symptoms. Methods: This study used data from the DIAMANTE (Diabetes and Mental Health Adaptive Notification Tracking and Evaluation) trial, a prospective, randomized controlled trial that varied SMS text messaging strategies to encourage physical activity to support both mental and physical health for patients with diabetes and depression. Patients were recruited from safety-net clinics in San Francisco and through social media during the COVID-19 pandemic, and were randomized into three trial arms: (1) personalized SMS text messaging about physical activity via an adaptive learning algorithm, (2) randomly selected SMS text messaging about physical activity, and (3) a control group that received no SMS text messages. As a secondary outcome, we examined pre-post changes in patient-reported health care communication, assessed via surveys with the validated Consumer Assessment of Healthcare Providers and Systems (CAHPS) communication subscale. Bivariate comparisons examined changes in CAHPS scores, including age, gender, preferred language, race or ethnicity, nativity, marital status, and education. Our primary analysis used mixed-effects modeling within an intent-to-treat analysis to determine differences in CAHPS scores by trial arm. Results: A total of 195 patients participated in the DIAMANTE trial from 2020 to 2022. After excluding patients who had incomplete or missing baseline or exit surveys, 156 patients were included in this secondary analysis. Overall, there was a substantive but nonsignificant decrease in the average CAHPS score over the 6-month trial period (-2.6; P=.11), with similar trends across patient demographic subgroups. Upon evaluating health care communication across the three randomized controlled trial (RCT) arms, there were no significant differences in patient-provider communication. Conclusions: Digital health interventions are rapidly increasing in clinical practice and have the ability to reach historically underserved communities by overcoming barriers such as language, geography, and time constraints. However, RCTs have primarily focused on the clinical impact of these tools, rather than how the tools may impact patients' perceptions of their relationships with providers. While our study found no significant changes in patient-provider communication by RCT arm, the temporal implications of the COVID-19 pandemic during the DIAMANTE study period remain unclear and should be further studied. This study is the first of its kind to examine the influence of an adaptive mHealth intervention on patient-reported health care communication, with insights that may be important as digitally-enabled chronic care management rapidly expands.