TL;DR
This study investigated the completion and compliance rates of an intensive 6-month mHealth intervention designed for caregivers of individuals with traumatic brain injury (TBI). The findings reveal surprisingly high compliance across various data collection methods, suggesting that well-designed mHealth interventions can be effective even with demanding protocols. This is crucial for developing sustainable digital health solutions for often-overlooked caregiver populations.
What Problem Does This Paper Address?
Caregivers of individuals with traumatic brain injury (TBI) face significant emotional, physical, and logistical burdens. While mobile health (mHealth) interventions show promise for supporting these caregivers, their effectiveness hinges on participant engagement and compliance. Previous research indicates that compliance rates in mHealth studies can be highly variable, influenced by factors like study duration, participant demographics, and individual characteristics. This paper addresses the specific challenge of maintaining high compliance in an intensive mHealth study design over a prolonged period (6 months) for this vulnerable population, aiming to understand what drives engagement in such a demanding context.
Methods
This study was a secondary analysis of a randomized controlled trial (RCT) testing the "CareQOL" app, an mHealth intervention designed to promote self-awareness and self-care among TBI caregivers. The intervention involved a multi-faceted data collection protocol over 6 months, including:
- Daily Ecological Momentary Assessments (EMAs): Three brief questions delivered daily.
- Monthly Patient-Reported Outcome (PRO) Surveys: Comprehensive surveys assessing various health-related quality of life domains.
- Continuous Activity and Sleep Monitoring: Using a wearable Fitbit device.
- Follow-up PRO Surveys: Administered at 3 and 6 months post-monitoring.
Participants were randomized into two arms: one receiving self-monitoring features only, and the other receiving self-monitoring plus personalized self-care push notifications. A self-monitoring dashboard (CareQOL app) provided participants with graphical displays of their daily data. The analysis focused on overall and component-specific completion and compliance rates, exploring the impact of demographic and person-specific factors.
Datasets
The study utilized data collected from 254 TBI caregivers who participated in the randomized controlled trial. This dataset includes:
- Daily EMA responses.
- Monthly and follow-up PRO survey responses.
- Fitbit data on daily steps and sleep duration.
- Demographic information of the participants.
No new datasets were introduced or benchmarks established; the paper focuses on analyzing the compliance data from the existing trial.
Key Insights and Ideas
- Intensive mHealth is Feasible for Caregivers: The most striking insight is that even a demanding 6-month mHealth protocol, involving daily assessments and continuous wearable monitoring, can achieve high compliance rates in TBI caregivers. This challenges the assumption that such interventions are inherently too burdensome for this population.
- Design Matters: The paper strongly suggests that specific study design elements are critical for fostering engagement. These likely include survey brevity, customizable administration windows, regular reminders, and monetary compensation.
- Demographics Play a Role, But Not Exclusively: While race (specifically being Black) and relationship to the care recipient (being a friend or other family member vs. a partner/child) were associated with lower compliance in certain aspects (Fitbit data), they were not the sole predictors of engagement. This highlights the interplay between individual characteristics and study design.
- Positive Perceptions Drive Compliance: Caregivers' positive perceptions of the study's feasibility, acceptability, and ease of use were consistently linked to better completion and compliance rates across all data types. This underscores the importance of user experience in mHealth research.
- Subgroup Analysis Reveals Patterns: The cluster analysis identified distinct groups of participants based on their compliance patterns (e.g., "high compliance-all data," "high compliance-PROs and steps only," "moderate PRO and low Fitbit compliance"). These subgroups were primarily predicted by race and relationship to the caregiver, suggesting targeted strategies might be needed for specific segments of the caregiver population.
Key Results
- High Overall Compliance:
- Daily EMA PRO completion rate: 84% (SD 19%)
- Fitbit-based step count compliance: 90% (SD 21%)
- Fitbit-based sleep duration compliance: 75% (SD 32%)
- Even Higher Survey Compliance: Monthly PRO survey completion rates ranged from 97% to 100%, and follow-up completion rates were 95%.
- No Significant Differences Between Study Arms: Compliance rates did not differ significantly between the self-monitoring only arm and the self-monitoring plus self-care push notifications arm.
- Predictors of Compliance: Positive perceptions of feasibility and acceptability were associated with better compliance. Being Black was linked to lower Fitbit compliance (steps and sleep). Being a friend or other family member was associated with lower Fitbit step count compliance. Being single or caring for someone with more functional deficits was associated with lower Fitbit sleep compliance.
- Cluster Analysis: A 3-cluster model identified distinct compliance profiles, with cluster membership predicted by race and relationship to the caregiver.
Limitations and Future Work
The authors acknowledge several limitations:
- Post-hoc Analysis: This was a secondary analysis, and the impact of specific design elements (like compensation or survey length) was not systematically assessed. Future work should directly investigate the influence of these factors.
- Sample Characteristics: The sample did not universally report high levels of caregiver strain, suggesting they might have been higher functioning or had more resources than the general caregiver population, potentially contributing to higher compliance.
- Clustering Model Selection: The choice of a 3-cluster model was based on a combination of data-driven and clinical interpretability, but other models could have been explored.
Future work should focus on:
- Directly assessing the impact of specific study design elements on compliance.
- Understanding the reasons behind lower compliance in disadvantaged groups (racial/ethnic minorities, single, or nontraditional caregivers) and developing targeted interventions.
- Exploring how these compliance patterns relate to actual improvements in caregiver well-being and self-care.