A statistical framework for quantifying clinical equipoise for individual cases during randomized controlled surgical trials

Abstract

Abstract Background Randomised controlled trials are being increasingly used to evaluate new surgical interventions. There are a number of problematic methodological issues specific to surgical trials, the most important being identifying whether patients are eligible for recruitment into the trial. This is in part due to the diversity in practice patterns across institutions and the enormous range of available interventions that often leads to a low level of agreement between clinicians about both the value and the appropriate choice of intervention. We argue that a clinician should offer patients the option of recruitment into a trial, even if the clinician is not individually in a position of equipoise, if there is collective (clinical) equipoise amongst the wider clinical community about the effectiveness of a proposed intervention (the clinical equipoise principle). We show how this process can work using data collected from an ongoing trial of a surgical intervention. Results We describe a statistical framework for the assessment of uncertainty prior to patient recruitment to a clinical trial using a panel of expert clinical assessors and techniques for eliciting, pooling and modelling of expert opinions. The methodology is illustrated using example data from the UK Heel Fracture Trial. The statistical modelling provided results that were clear and simple to present to clinicians and showed how decisions regarding recruitment were influenced by both the collective opinion of the expert panel and the type of decision rule selected. Conclusions The statistical framework presented has potential to identify eligible patients and assist in the simplification of eligibility criteria which might encourage greater participation in clinical trials evaluating surgical interventions.

Keywords

• Equipoise
• Randomised controlled trial
• Surgery
• Statistical model