Outcome reporting bias in clinical trials: why monitoring matters