Frequentist and Bayesian Sequential Experiments for Pandemic Disrupted Clinical Trials: Can Novel Methodologies Deliver Greater Value for Health Care Systems?
摘要
We apply innovations in frequentist and Bayesian sequential experimental design to clinical and cost-effectiveness data from the DISC trial, a UK National Institute for Health and Care Research-funded clinical trial that was disrupted during the COVID-19 pandemic. Analysis of the cost-effectiveness data suggests that stopping recruitment at the pandemic’s outbreak was indicated by some models, but not others. Analysis of the clinical effectiveness data suggests that stopping recruitment was never indicated. Reflecting recent contributions [5, 6], we illustrate how trial management decisions during a pandemic might benefit from insights provided by a range of statistical and economic analyses of the same experimental data.