An Integrative, Digital Health Approach to Veteran-Centered PTSD Care

Setting of the study: Data will be collected from participants recruited from bike riding events hosted by collaborator Project Hero. Their rides include participants from across the U.S., but specific riding challenges will take place in major areas such as Texas, California, the Great Lakes region, and Nevada in the year 2025 onward. The same study procedure will occur at each location. The participants will also participate in the study from their home locations. Study design: The investigators propose a single-blind three-group randomized controlled trial comparing: (1) a digital health self-management intervention combined with an intensive group-based physical activity intervention, (2) an intensive group-based physical activity intervention without the digital health component (active comparator), and (3) a no-intervention control group. Mental health symptomatology will be measured across three time intervals to assess within-subject effects over a two-month period: pre-intervention (one month), during the intervention (one week), and post- intervention (one month). Participants in the experimental groups will be recruited from veteran registrants of the Project Hero Ride 2 Recovery 7-day Texas challenge (and 7-day California challenge or 10-day Europe challenge, based on registration numbers), while the participants for the control group will be selected from the broader Project Hero veteran community not registered for the Ride 2 Recovery events. Participants in the two intervention groups will be randomly placed into their groups using an RStudio randomization script. Participants in the control group will be selected from members of the veteran community that are not participating in the Ride 2 Recovery events of their own volition. All participants will wear an Apple Watch and download the FWD app for the duration of the study to measure stress moments (via machine learning model with heart rate and accelerometer data as inputs) and take psychological assessment surveys (GAD-7, PHQ-9, and PCL-5). Participants in the control group will not have access to the app features aside from tracking heart rate, stress moments, and psychological assessments. Participants in the control group and physical intervention group without digital intervention will still be asked to perform weekly self-assessments of mental health conditions as well as self-report stress moments and respond to app-detected stress moments to validate the ML algorithm. Participants in the digital intervention group will have full access to the First Watch Device app (FWD), which includes a wide suite of self-management material such as relaxation techniques, links to outside resources, educational material, emergency SOS messages to contacts, and monitoring of symptoms. One on one interviews may be performed with the participants at the start of each phase of the study: pre-intervention, during intervention, and post-intervention as well as a follow up at the end of the post-intervention period where they will be asked about their experience with the interventions, managing their symptoms, and the usability of the FWD app. At the end of the study, both versions of the FWD app (control and digital intervention) will be made available to participants at their request until app usage expires. This study combines Project Hero's Ride 2 Recovery, an intensive group-based cycling intervention, with a digital health self-management tool to enhance PTSD symptom monitoring and support. Each Ride 2 Recovery event is performed in a different location. The events of interest for this study are the 1 week long rides that take place in Texas (in Spring) and California (in Fall) once a year. Participants in the ride generally ride for most of the day (6-8 hours) across a predetermined route within the state. Eligible participants are veterans and first responders. The digital intervention leverages Apple smartwatches to (1) Continuously monitor key physiological indicators (e.g., HRV, skin temperature, electrodermal activity) for early distress detection; and (2) Deliver real-time, personalized interventions via the Mental Health Evaluation and Lookout Program (mHELP) bundle, offering symptom tracking, self-management tools, and access to counseling. Access to counseling is provided through the "Link" navigation screen on the app, which provides several contacts for veteran crisis lines, hotlines, call centers, and the Veterans Health Administration. Using advanced ML algorithms, the system will identify and examine physiological data to trigger timely alerts and tailored recommendations. For example, when FWD detects a hyperarousal event, akin to a panic attack, the app will prompt the user to open the app and perform relaxation mechanisms like breathing and also allow the user to send an SOS message to their emergency contacts by tapping on their watch. Additionally, user notifications are programmed into the app to remind users to review educational material as well as complete psychological surveys. A suite of self-management tools is provided for on-demand and recommended access, including breathing and focusing exercises, relaxing media, multimodal educational material, journaling tools, trends and visualizations of mental health status, mental health self-assessment, and more. The tool will also enable self-reports of hyperarousal events (double tapping on the watch interface or using an accessible feature on the mobile app). The integration of smartwatch technology with physical activity aims to provide scalable, accessible, and personalized support for veterans, enhancing their mental health and overall resilience. Data Collection: To assess the impact of the interventions on mental health symptoms, both subjective and objective data will be collected via Apple Watch and smartphone. Subjective measures will be gathered at four time points-one month before the Texas Challenge event (or at any number of recurring Challenge events thereafter depending on award date and registrant numbers), pre-event, post-event, and one month after the event-to evaluate participants' perceived mental health status and track changes over time. Psychological well-being will be assessed using validated self-report instruments, including: (i) the PTSD Checklist for DSM-5 (PCL-5) to measure PTSD symptom severity (Aim 1); (ii) the Patient Health Questionnaire (PHQ-9) to assess depression symptom severity (Aim 2); and (iii) the Generalized Anxiety Disorder 7-item (GAD-7) scale to evaluate symptom severity of generalized anxiety (Aim 2). These surveys will be programmed within the FWD app, which will act as a central hub for these participants during the course of the study. Notifications from the app will be sent to the participants on their phones to complete these if not done on a timely manner. These measures will provide valuable insights into the differential effects of the interventions on participants' mental health. As objective measures, physiological data, which include heart rate measured in beats per minute and accelerometer data, will be collected one month before, during, and one month after the Texas Challenge event via Apple Watch (Aim 3). The event will provide a dynamic environment to observe fluctuations in physiological measures under real-world conditions. Participants will wear an Apple Watch throughout the study to continuously monitor heart rate (BPM), accelerometer data, and subjective stress ratings. This timeline will allow for a comprehensive analysis of physiological changes over time, with a focus on responses to potential PTSD triggers during high-intensity activities such as long-distance cycling. Data analytics: The data will be analyzed using Generalized Linear Mixed Modeling (GLMM) to evaluate the impact of interventions on mental health symptomatology over time (pre-during-and-post event phases). GLMM is appropriate for this study as it accommodates the hierarchical structure of repeated measurements nested within individuals and individuals nested within groups. This approach allows for the inclusion of both fixed effects (e.g., intervention type, time points) and random effects (e.g., individual variability), providing a flexible framework for modeling complex data. GLMM is particularly advantageous for a longitudinal design as it handles repeated measures data, accounts for correlations among observations, and allows for non-normal outcome distributions. GLMM can also incorporate random intercepts and slopes to capture individual variability, tolerate missing data, and model interaction effects. This analysis will focus on the fixed effects of intervention type and time, as well as their interaction, to determine the efficacy of the digital health intervention, with simple slope and pairwise contrast analyses to examine time and group differences.