A relationship between ethics of clinical trials and therapeutic advances in cancer

Abstract

6090 Background: We hypothesized that there is a predictable relationship between equipoise and the outcomes of trials, i.e. the equipoise principle bounds therapeutic advances in a such way that, over time, distribution of successes between experimental and standard treatments is expected to be about equal, reflecting the unpredictability of treatment success for a particular intervention. We tested this hypothesis in randomized controlled trials (RCT) performed by the National Cancer Institute (NCI) sponsored Cooperative Groups (CG), systematically taking into account all factors (publication bias, methodological quality, choice of control intervention) known to affect outcomes. Methods: All completed (published and unpublished) phase 3 RCTs from 6 CGs (563 trials, ∼ 149,000 patients) since 1955 were evaluated. Success of experimental vs. standard treatments was evaluated semi-quantitatively and quantitatively using a meta-analytic technique. Results: Investigator’s preference for experimental vs. standard treatments were 43% and 57%, respectively (p=0.002). Hazard ratio for overall survival was 0.96 [99% CI (0.95, 0.98), p=0.00001] and for event-free survival was 0.92, [99%CI (0.90, 0.93), p=0.0001] slightly favoring experimental treatments but at the expense of increased treatment-related mortality [HR 1.12, 99%CI (1.01, 1.25), p=0.04]. The results adhere to the equipoise hypothesis and were not affected by any factors known to bias outcomes. Conclusions: While many successful treatments are discovered in phase 3 RCTs, innovative treatments in cancer which have progressed to testing in RCTs are, on average, as likely to be inferior or superior to standard treatments. This result was predicted, which amounts to description of the law of therapeutic advances in clinical oncology. If the distribution of the results were significantly skewed to consistently and predictably favor one type of treatments over another, the clinical trial system would come to a halt (patients would naturally request only those treatments that are expected to be superior making randomization impossible). Progress in treating cancer is ultimately determined by the ethical principle of equipoise which should guide clinical research. No significant financial relationships to disclose.