Monitoring in clinical trials: benefit or bias?