Recruitment and retention in a preclinical AD trial: comparisons between academic and non-academic sites.

Abstract

Background: Alzheimer's disease (AD) clinical trials enroll participants at various site types including research-focused academic institutions and independent non-academic sites. Limited research has examined the impact of site type on recruitment and retention outcomes.

Methods: To evaluate potential differences between site types, we used data from the Anti-Amyloid Treatment for Asymptomatic AD (A4) trial, the largest completed preclinical AD trial to date. We first compared the frequency of varying recruitment sources between site types. We then examined potential differences in participant- and site-level characteristics. To assess potential site type differences in retention, we fit a multivariable logistic regression model adjusting for variables associated with site type. For participants who prematurely discontinued, we examined potential differences by site type in reasons for dropout.

Results: One thousand and fifty-eight participants were randomized at 50 academic (N = 835) and 15 non-academic (N = 223) sites in North America. Academic sites had higher proportions of participants recruited through earned media and organizational referrals and lower proportions recruited through internal referrals and advertisements, compared to non-academic sites. Participant-level characteristics differed between site types. Compared to non-academic sites, academic sites had higher proportions of participants with a family history of dementia and a professional degree (highest education category), but lower proportions of individuals with a history of diabetes, a CDR-SB score above 0, and belonging to a racial and ethnic underrepresented group. Though the results were not statistically significant, non-academic sites had a higher screening rate (number of participants screened/site/month), but a lower randomization rate (randomized/screened) compared to academic sites. No site type differences in completion rates were observed. When examining reasons for discontinuation, we found that among the 72 participants who discontinued the trial at non-academic sites, 56 (77.8%) withdrew consent or were lost to follow up. In contrast, 140 out of 243 (57.6%) participants who discontinued the trial in academic sites withdrew consent or were lost to follow up.

Conclusion: Our findings shed light on important site type differences that investigators should consider when making choices around site, design, and conduct in multisite preclinical AD trials.