Evaluating the process and ethics of randomizing African American and Latino adults with diabetes to a community health worker intervention

The purpose of this presentation is to describe and evaluate the process of a randomized control trial for a community health worker intervention with African American and Latino REACH Detroit participants.

We recruited 164 clients and randomly assigned them to two arms of our12-month intervention: an immediate and a six-month, delayed control group. Our overall retention rate is 71.7%, and was higher for Latinos (81%) compared to African Americans (62%). We had a higher drop-out rate in our delayed group (33%) compared to the immediate group (24%). Drop-out rates were similar across groups for African Americans (about 38%), but higher for the delayed group among Latinos (11% vs. 27%).

In-depth interviews with 20 clients across the two arms by race/ethnicity and drop-out status revealed that individuals did not report adverse effects due to randomization, but a few of the clients in the delayed group expressed considerable disappointment in not being selected for the immediate group. Although procedures were presented in the initial orientation and documented in the consent forms, many clients from both groups did not fully understand the process. A few clients questioned the fairness of the process and whether it was truly random. However, the majority expressed satisfaction, regardless of whether they were clear on the process. Drop out was not due to perceptions of fairness, but for other reasons, including work and time commitments.

The study offers lessons learned for designing randomized control trials with under-represented groups and policies related to ethical considerations in health disparities research.