The Digital Tracking of Arthritis Longitudinally (DIGITAL) study, an ancillary study of the ArthritisPower registry, demonstrated that a carefully designed digital study can optimize patient engagement and data completeness in rheumatoid arthritis (RA) research. The study, which was entirely web-based with no in-person visits, used a combination of a run-in period, automated prompts, and personalized communication to achieve high levels of participant adherence.
The study enrolled participants with self-reported RA and smartphone access, monitoring their engagement and adherence to protocol-specified data collection for weekly and daily electronic patient-reported outcomes (ePROs), as well as smartwatch data. Of the 470 participants expressing initial interest, 278 (59.1%) completed the run-in period and qualified for the main study. Over the 12-week main study period, 87.4% (243/278) of participants adhered to weekly ePRO data collection, 57.2% (159/278) to daily ePRO, and 81.7% (227/278) to smartwatch data collection. Composite adherence, defined as meeting all three adherence definitions, was achieved by 52.9% (147/278) of participants.
Key Design Elements for Enhanced Engagement
Building on lessons learned from previous studies, the DIGITAL study incorporated several design elements to promote engagement and minimize data missingness. These included:
- Run-in Period: A 14-day run-in period was implemented to increase app familiarity and identify motivated individuals. Participants were required to complete at least one set of weekly ePRO assessments and 10 of 14 daily ePROs to qualify for the main study.
- Automated and Manual Prompts: Participants received automated messages via email and smartphone notifications when smartwatch or ePRO data were missing. Study coordinators also contacted participants via SMS text messaging and phone to resolve adherence issues.
- Real-time Monitoring: Study coordinators monitored participant data daily using an Access database, which highlighted any need to identify or correct missing data. A heat map overview of data completeness was circulated to study leads regularly.
Factors Influencing Protocol Adherence
Modeling identified several factors associated with protocol adherence during the main study. These included age (OR 1.18 per 5-year increments, 95% CI 1.06-1.33), high adherence to daily ePROs (completing 10 of the first 14 days; OR 1.73, 95% CI 0.97-3.17), and weekly ePRO adherence during the run-in period (OR 5.31, 95% CI 1.27-36.19). Additional features included the most recent PROMIS fatigue score and the Outcome Measures in Rheumatoid Arthritis Clinical Trials RA Flare score.
Individuals who met the composite adherence measure were more frequently White and older than those who met fewer adherence measures. Frequency analysis showed that a larger proportion of participants who adhered to data submission were currently employed (55.4% vs 39.6%; P=.001) and receiving treatment with biologic disease-modifying antirheumatic drugs (63.3% vs 49.5%; P<.001).
Implications for Future Digital Studies
The DIGITAL study provides valuable insights into designing effective remote, web-based studies with ePROs. The findings suggest that a run-in period, combined with active monitoring and personalized communication, can significantly improve participant engagement and data completeness. The study also highlights the importance of considering participant characteristics, such as employment status and treatment type, when planning digital studies.
"This study demonstrates that it is possible to conduct rigorous, fully remote clinical research and collect high-quality data from participants with rheumatoid arthritis," said a lead researcher on the project. "By incorporating a patient-centric design and utilizing a combination of automated and manual prompts, we were able to achieve adherence rates comparable to those seen in traditional clinical trials."
Limitations
The study had several limitations. The run-in period limited the ability to model for factors related to adherence in a more general population. There was a potential for selection bias because only RA patient members of the ArthritisPower registry with an email address could be invited to participate in the study. The study did not integrate the results into in-person visits for clinical care, which has also been shown to increase adherence.
Conclusion
The DIGITAL study provides a valuable framework for designing and conducting future digital studies in rheumatoid arthritis and other chronic conditions. By incorporating the lessons learned from this study, researchers can improve participant engagement, data completeness, and the overall quality of digital health research.
