Different ways to estimate treatment effects in randomised controlled trials.