ADJUSTING FOR NONCOMPLIANCE AND CONTAMINATION IN RANDOMIZED CLINICAL-TRIALS

A method of analysis is presented for estimating the magnitude of a tr eatment effect among compliers in a clinical trial which is asymptotic ally unbiased and respects the randomization. The approach is valid ev en when compliers have a different baseline risk than non-compliers. A djustments for contamination (use of the treatment by individuals in t he control arm) are also developed. When the baseline failure rates in non-compliers and contaminators are the same as those who accept thei r allocated treatment, the method produces larger treatment effects th an an 'intent-to-treat' analysis, but the confidence limits are also w ider, and (even without this assumption) asymptotically the efficienci es are the same. In addition to providing a better estimate of the tru e effect of a treatment in compliers, the method also provides a more realistic confidence interval, which can be especially important for t rials aimed at showing the equivalence of two treatments. In this case the intent-to-treat analysis can give unrealistically narrow confiden ce intervals if substantial numbers of patients elect to have the trea tment they were not randomized to receive. (C) 1997 by John Wiley & So ns, Ltd.