Enhancing pediatric clinical trial feasibility through the use of Bayesian statistics

Background Pediatric clinical trials commonly experience recruitment challenges including limited number of patients and investigators, inclusion/exclusion criteria that further reduce the patient pool, and a competitive research landscape created by pediatric regulatory commitments. To overcome these challenges, innovative approaches are needed. Methods This article explores the use of Bayesian statistics to improve pediatric trial feasibility, using pediatric Type-2 diabetes as an example. Data for six therapies approved for adults were used to perform simulations to determine the impact on pediatric trial size. Results When the number of adult patients contributing to the simulation was assumed to be the same as the number of patients to be enrolled in the pediatric trial, the pediatric trial size was reduced by 75–78% when compared with a frequentist statistical approach, but was associated with a 34–45% false-positive rate. In subsequent simulations, greater control was exerted over the false-positive rate by decreasing the contribution of the adult data. A 30–33% reduction in trial size was achieved when false-positives were held to less than 10%. Conclusion Reducing the trial size through the use of Bayesian statistics would facilitate completion of pediatric trials, enabling drugs to be labeled appropriately for children.