Inadequacy of sample sizes in clinical trials of laboratory parameters attributable to invalid statistical assumptions

Clinical trials often determine the sample size based on the use of statistical methods such as analysis of variance, t tests, and rank sum tests, which compare mean or median values. The resulting studies rarely are big enough to show that the method is based on mistaken assumptions. Data from a recent clinical trial of nephrotoxicity associated with the use of contrast agents during angiography found a significant difference on the order of the difference that previous studies had intended to detect. It also showed that a central assumption does not apply to changes in serum creatinine. As a result, the previous studies had considerably lower power than believed. Their lack of significance reflected only the mistaken assumption. Analysis of variance, t tests, and rank sum tests may be just as invalid for other clinical parameters. Lack of significance cannot be automatically taken to imply a small treatment effect. Clinical Pharmacology and Therapeutics (1994) 56, 437–444; doi: 10.1038/clpt.1994.158