Recruitment to publicly funded trials--are surgical trials really different?
Cook JA., Ramsay CR., Norrie J.
BACKGROUND: Good recruitment is integral to the conduct of a high-quality randomised controlled trial. It has been suggested that recruitment is particularly difficult for evaluations of surgical interventions, a field in which there is a dearth of evidence from randomised comparisons. While there is anecdotal speculation to support the inference that recruitment to surgical trials is more challenging than for medical trials we are unaware of any formal assessment of this. In this paper, we compare recruitment to surgical and medical trials using a cohort of publicly funded trials. DATA: Overall recruitment to trials was assessed using of a cohort of publicly funded trials (n=114). Comparisons were made by using the Recruitment Index, a simple measure of recruitment activity for multicentre randomised controlled trials. Recruitment at the centre level was also investigated through three example surgical trials. RESULTS: The Recruitment Index was found to be higher, though not statistically significantly, in the surgical group (n=18, median=38.0 IQR (10.7, 77.4)) versus (n=81, median=34.8 IQR (11.7, 98.0)) days per recruit for the medical group (median difference 1.7 (-19.2, 25.1); p=0.828). For the trials where the comparison was between a surgical and a medical intervention, the Recruitment Index was substantially higher (n=6, 68.3 (23.5, 294.8)) versus (n=93, 34.6 (11.7, 90.0); median difference 25.9 (-35.5, 221.8); p=0.291) for the other trials. CONCLUSIONS: There was no clear evidence that surgical trials differ from medical trials in terms of recruitment activity. There was, however, support for the inference that medical versus surgical trials are more difficult to recruit to. Formal exploration of the recruitment data through a modelling approach may go some way to tease out where important differences exist.