Problems of stopping trials early

When interim analyses of randomised trials suggest large beneficial treatment effects, investigators sometimes terminate trials earlier than planned. Gordon H Guyatt and colleagues show how this practice can have far reaching and harmful consequences