Detecting Conversation Topics in Recruitment Calls of African American Participants to the All of Us Research Program...

Area of Science:

• Biomedical research
• Health disparities
• Clinical trial recruitment

Background:

• Clinical research diversity is crucial for equitable health outcomes.
• Underrepresented groups, especially African Americans, face historical trust barriers in research.
• The All of Us Research Program (AoURP) aims to increase diversity in biomedical research.

Purpose of the Study:

• Analyze recruitment conversations in the AoURP to identify enrollment influencers.
• Improve engagement and recruitment strategies for underrepresented populations in clinical research.

Main Methods:

• Observational, retrospective study using machine learning for content analysis.
• Structural topic modeling identified conversational themes in recruitment calls.
• Analysis focused on calls by Morehouse School of Medicine RAs (Feb 2021-Apr 2022).

Main Results:

• 12 coherent topics identified from 45 estimated topics.
• Successful enrollment linked to discussions on follow-up appointments, COVID-19 safety, precision medicine, and addressing concerns (cost, insurance, health fears).
• Non-enrollment conversations often involved technical issues and descriptions of physical measurements.

Conclusions:

• Machine learning effectively identifies themes in recruitment conversations with minimal subjectivity.
• This approach offers insights to improve clinical trial recruitment for underserved communities.