April 16, 2009– Dr. Guyatt explained in a previous post why early stoppage of clinical trials results in falsely inflated efficacy of the therapy being studied. The April 16, 2009 edition of The New England Journal of Medicine has a great example how Kaplan curves tend to converge with time and erase early efficacy differences. The study looked at the drug valsartan for the treatment of atrial fibrillation and failed to show a benefit over the comparator cohort(see graph). If it were stopped early, it would have shown that the drug was effective. Note the curves touching, or converging, as time progresses.
