Sample size calculations for randomized controlled trials and for prediction models

Two study protocols are published in this issue of Colorectal Disease: FALCON, a multicentre randomised controlled trial of strategies to reduce surgical site infection, and AFAR, a predictive model of atrial fibrillation after colonic resection. Both are exemplars of excellent research design that surgeon researchers should seek to emulate. Trial statisticians were involved at an early stage and the protocols have been through several rounds of peer review by trial methodologists, prior to being funded by the National Institute for Health Research (NIHR). In this article we address the important question of sample size calculations and how they should be approached for these very different forms of study.