A systematic review of the differences between trial-eligible and trial-ineligible patients in real-world populations

Abstract Background Many people with a physical health condition are ineligible for randomised clinical trials (RCTs) of treatments for that condition. However, less is known about the differences in characteristics and outcomes between those that are eligible and ineligible for RCTs. Methods Systematic review without meta-analysis. Screened studies were those included in a previous review of trial eligibility (full-text rescreened for this study) and new studies identified in a Medline and Embase search from 12/02/18 to 13/04/23. Included studies were those which evaluated the proportion of patients in a real-world population eligible for one or more RCTs, which additionally compared eligible and ineligible patient characteristics. Titles and abstracts were screened by two reviewers independently and risk of bias assessed. Comparisons were demographic/clinical characteristics and outcome rates in real-world data (RWD). Median differences were compared using Wilcoxon signed rank. Results Forty-nine studies examining 122 trials in 52 RWD sources were included. 34/49 studies examined cardiovascular trials (57/122 trials), seven examined cancer (30 trials), two musculoskeletal diseases (24 trials) and two respiratory diseases (11 trials). Risk of bias was high for 33 (67.3%) studies. The median proportion of women was higher in the trial-ineligible (39.0%) than trial-eligible (34.6%, p = 0.008, 63 trials). Comorbidities examined by the included studies were less common in the trial-ineligible (e.g. hypertension median prevalence 61.0% vs 66.0%; p = 0.03, 35 trials), but baseline condition severity was higher (median proportion with markers of more severe index disease 37.9% vs 33.5%, p = 0.02, 43 trials). No difference was found for age. All-cause mortality did not differ, but serious adverse events occurred more frequently in the trial-ineligible (median 12.9% vs 6.3%, p = 0.017, 14 trials). Conclusions The data suggest a systemic exclusion of women, individuals without relevant risk factors, and those at increased risk of adverse events, but data are limited and heavily weighted towards cardiovascular trials which commonly recruit (and report on) cardiovascular comorbidities/risk factors. There is a need for more systematic research examining the impact of trial-eligibility and recruitment on who is included in trials and evidence. Trial registration PROSPERO registration CRD42023415960. Registered on April 26, 2023. ( https://www.crd.york.ac.uk/PROSPERO/view/CRD42023415960 ).