A tool to evaluate rapidly the quality of clinical trials on topical acne treatment

Practicing dermatologists rarely evaluate therapeutic trials. This study was designed to build a simple evaluation tool, use it to evaluate the quality of published trials and thereby consider their applicability. Reports on 146 therapeutic trials, published from 1969 to 2004, were selected. A 50-item list was used for an initial assessment. After comparison with existing grids, a shortened nine-item list was established and used for this analysis. The results were analyzed item by item (logistic regression) and globally (principal component analysis). An index was then established and validated using Cronbach's coefficient alpha. The overall quality of trials was poor: 9% to 73% positive response range. This rate increased after 1990 for six items. Two factors explained almost 60% of the total variance: factor 1 summing six items concerning the formal content of the articles and factor 2 its underlying basis. Using the sum of these six variables, Cronbach's coefficient alpha reached 0.716; their mean sum improved from 1969 to 1970-2001-2005. The significantly improved quality of trials, based on the mean sum over time, supports well-founded basis of this tool. Regarding the application of evidence-based medicine in clinical practice, this tool is user friendly and should facilitate the updating of our knowledge.