Path Analysis of RCT Recruitment: Secondary Analysis of Results from a Systematic Review

We examine recruitment processes for 71 pragmatic and comparative effectiveness trials identified in a systematic review, using path analysis to examine rates of refusal to screen, test, and consent to trial participation. Our analysis suggests that refusal rates might be on net slightly higher if potential subjects are screened or asked to undergo physical eligibility tests, but this was not significant in our sample of trials (p = .11 by Mann-Whitney test). We find that rates of refusing to provide informed consent are much lower for trials in which subjects have agreed to screening or testing (odds ratio = 0.40, Wilcoxon rank-sum z = 2.67, p = .008). We also observe that the overwhelming majority of trials examined secured consent after determining eligibility, even in trials involving screening or testing activities. The ethical implications and areas for future research are discussed.