Key result
Cardiovascular RCTs in clinical guidelines enrol only ~47% of screened patients, with widespread underreporting.
Meta-Analysis
Answers are grounded in this paper and cite their sources.
TO THE EDITOR: Recent years have witnessed an exponential growth in studies that utilize ‘real-world’ data.1 However, those studies have often attracted criticism due to their inherent biases.1 While randomized controlled trials (RCTs) are considered gold standard in evidence hierarchy,1 narrow representation of the general population due to stringent enrolment criteria generate concerns.2 Indeed, studies have documented underrepresentation of women, elderly, racial minorities, and multimorbid patients in RCTs.3 Moreover, an upstream enrolment to screening gap in RCTs may further limit the generalizability of their findings.4,5 Nonetheless, the magnitude of this gap remains under-assessed in cardiovascular RCTs. We examined the American College of Cardiology/American Heart Association and European Society of Cardiology Cardiovascular guidelines to systematically assess the enrolment to screening ratio in RCTs that formed the basis of guidelines. Two authors (A.N.L. and S.U.K.) reviewed the guidelines on the management of coronary artery disease, atrial fibrillation, heart failure (HF), blood cholesterol, stroke, ventricular arrhythmia, and valvular heart disease for RCTs through April 2020. The enrolment-to-screening ratio was calculated as the number of patients enrolled divided by the number of patients screened.5 Study-level estimates were pooled using DerSimonian and Laird random-effects model. Subgroup analyses explored potential sources of heterogeneity. Of 565 articles screened, 160 trials provided relevant information. The median number of participants screened and enrolled were 1176.5 (interquartile range, IQR 407.2–5106) and 422 (IQR 168.7–1028.7), respectively, corresponding to an enrolment to screening ratio of 46.8%. The enrolment to screening ratio increased over time and varied according to disease state, sponsorship, intervention, and region (Table 1). The most frequent clinical exclusion criteria were infection, participants having life expectancy <1 year, uncontrolled hypertension, end-stage kidney disease, cancer, HF Class III–IV, recent acute coronary syndrome or stroke, and women of child-bearing potential. Enrolment to screen ratio in cardiovascular randomized controlled trials Enrolment to screen ratio in cardiovascular randomized controlled trials In this review of over 30 years of RCTs cited in professional cardiovascular guidelines, less than half of the RCTs reported quantitative recruitment information. Overall, approximately one in two participants screened were enrolled in cardiovascular RCTs, which varied markedly according to study characteristics and type of intervention. Reporting of study selection criteria was sporadic, and excluded the participants ranging from healthy women to chronic comorbidities of variable severity. This degree of selection bias raises the concerns whether findings of cardiovascular RCTs can be comfortably generalized to the global population. Observational studies are criticized for inherent confounding and residual biases. However, real-world evidence can complement RCTs data, especially when RCT data is sparse to inform practices for the general population.2 In future, the CONSORT guidelines should push the investigators to document the number of participants identified for enrolment and fraction of those who were recruited to preserve the internal and external validity of the RCTs.4 Target population base should be expanded to closely approximate the general population. A multi-incentive approach regarding the potential benefits of therapy might limit the number of eligible candidates lost due to consent withdrawal or protocol deviation. Finally, innovative strategies like the application of randomization to real-world data might be instrumental in promoting their role in the evidence framework.1 Data set: Available from Dr. Khan (e-mail, safinmc@gmail.com). Conflict of interest: none declared.
No takes yet. Share an insight, caveat, or question.
Khan et al. (2020) conducted a meta-analysis in Cardiovascular diseases (coronary artery disease, atrial fibrillation, heart failure, blood cholesterol, stroke, ventricular arrhythmia, and valvular heart disease). Cardiovascular guideline-cited RCT recruitment screening was evaluated on Enrolment-to-screening ratio (number of patients enrolled divided by number of patients screened). Cardiovascular RCTs cited in clinical guidelines had a pooled enrolment-to-screening ratio of 46.8%, with only 160 of 565 trials reporting recruitment data.
Synapse has enriched 10 related papers. Consider them for comparative context: