The authors respond

From BMJ USA 2002;September:523

Geis characterizes CLASS as a single study, as was done in JAMA, which reported patients to be “randomly assigned on a 2:1:1 basis.”1 This description is misleading; there were two separate trials, with two separate patient recruitment and randomization procedures, and therefore requiring separate analyses to preserve randomization. Nonetheless, the two trials were combined by simply adding up numbers.1

The assumption underlying this approach is that allocation of patients to the two trials was ruled by chance alone. We tested this assumption by comparing the trials in terms of the 18 patient characteristics available at baseline.2 If allocation to the two trials had been governed by chance we would expect approximately one of the 18 comparisons to be statistically significant at P<0.05. However, patients in the two …