Bayesian analysis of real‐world data as evidence for drug approval: Remembering Sir Michael Rawlins
British Journal of Clinical Pharmacology July 17, 2023 Balázs Szigeti, Lawrence D. Phillips, David Nutt 13 citations
Randomized controlled trials (RCTs) are often considered the gold standard in medical research, but they have limitations including reliance on null hypothesis significance testing and poor generalizability. Bayesian analysis of real-world evidence (RWE) offers a complementary approach. In a case series of 20 children with epilepsy treated with medical cannabis, all experienced reduced seizures; Bayesian analysis with a flat prior gives a 95% probability that the next patient will improve (95% credible interval 87%–100%). For treatment-resistant depression treated with psilocybin, the probability of a favorable response ranges from 62% (QIDS-16) to 82% (MADRS). These analyses require fewer patients than traditional RCTs and provide directly actionable probabilities for clinicians and patients.