In Reply: The points made by Dr Chan and colleagues are important for both well-informed treatment decisions in the clinic and the progress of medical research generally. The recent systematic review by Bassler et al1 of inflated treatment effects in trials stopped early for benefit demonstrates the importance of prespecified stopping rules that require a significant number of events before stopping. It shows the need for full access to the protocol, including prespecified stopping rules, to independently assess the accuracy and reliability of the reported treatment effect.
Miller JD. Registering Results From Clinical Trials—Reply. JAMA. 2010;303(21):2138-2139. doi:10.1001/jama.2010.704