Abstract 173: Tailoring Follow-Up Methods for Hard-to-Reach Participants in Outcomes Research: An Example from TRACE-CORE

Aims: Telephone follow-up for longitudinal outcomes research is common, with widely variable response rates. Non-respondents are often different from respondents and come from vulnerable groups, including racial and ethnic minorities, elders, or those with multiple co-morbidities. We aimed to create a comprehensive follow-up strategy to target follow-up resources to the hardest-to-reach participants. These methods could help outcomes researchers retain hard-to-reach patients in longitudinal studies. Methods: Using interview completion data from the first 614 patients hospitalized with an acute coronary syndrome and enrolled in the Transitions, Risks, and Actions in Coronary Events Center for Outcomes Research and Education (TRACE-CORE) cohort at our largest enrollment site, we fit a logistic regression model that used socio-demographic characteristics and difficulties with hearing and vision to predict likelihood of non-completion of a 1-month post-hospitalization telephone interview. Patients whose predicted probability of non-completion exceeded a threshold were triaged to a new follow-up protocol with more intensive calling by site-specific staff with access to medical records and follow-up appointment schedules. We compared the post-implementation follow-up completion rates to the expected completion rates based on the model. Results: The model predicted a completion rate of 69.8% (56 of 80) among the hard-to-reach group; our intervention achieved a 85.0% (68 of 80) completion rate (p=0.002). Increases in response rates were observed across most subgroups, although not all improvements reached statistical significance (Table). Improvements in completion rates were less pronounced for vulnerable groups. Conclusions: We used a statistical model to prospectively identify a “hard-to-reach” group for more intensive follow-up efforts. The statistical model provided an expected completion rate for patients receiving the new protocol against which the observed rate could be tested, allowing us to conclude that the enhanced protocol worked to improve follow-up in these typically hard-to-reach populations.