A multiple testing procedure for clinical trials.

摘要: A multiple testing procedure is proposed for comparing two treatments when response to treatment both dichotomous (i.e., success or failure) and immediate. The test statistic each the usual (Pearson) chi-square based on all data collected that point. maximum number (N) of tests (m1 + m2) observations between successive fixed in advance. overall size shown be controlled with virtually same accuracy as single sample N(m1 observations. power also found same. However, by affording opportunity terminate early one performs markedly better than other, may eliminate ethical dilemmas often accompany clinical trials.