Design issues in clinical trials
Overall, J.E.
Progress in Clinical and Biological Research 317: 1291-1299
1989
ISSN/ISBN: 0361-7742 PMID: 2690096 Document Number: 333588
As noted, many issues in the design of clinical trials and the analysis of data therefrom deserve consideration with respect to Alzheimer's disease. This chapter has emphasized the different types of experimental designs that appear most appropriate in different phases of clinical drug development. Intensive within-subject designs that contrast different treatment conditions within the same group of subjects appear useful for early clinical trials. The standard parallel-groups design can provide substantial evidence of short-term efficacy in large-scale, multicenter trials; however, longer term treatment is required to evaluate the practical value of a new drug for the treatment of Alzheimer's disease. A survival analysis model is proposed as useful to ensure only minimal deterioration before termination of double-blind conditions for patients who might be assigned to placebo. Research protocols that plan in advance an interim analysis of data from large-scale clinical trials can be useful in reaching early decisions while not jeopardizing continuation of studies in which no interim decision is justified. A special issue regarding the reporting of two-tail p-values for tests of treatment effects that can only have practical consequences in one direction has been discussed as particularly relevant to FDA evaluation of placebo-controlled clinical trials. One-tail p-values represent the actual Type I error probabilities in cases where treatment effects in only one direction will ever be the basis for a claim of statistical significance.