Typology of the primary outcome construction in dermatology: a systematic review of published randomized controlled trials.

The primary outcome is a major component of a randomized clinical trial (RCT) and several types of outcome can be chosen in a given disease. We systematically reviewed RCTs on nonneoplastic dermatological diseases published in 2009 and referenced in Medline, and described how the main outcome was defined and constructed. We assessed whether those characteristics were associated with a clearly defined primary outcome and whether they were associated with a significant statistical test for the primary outcome. Outcome construction variables were the three "VIP" questions (V denoting Variable: binary vs. quantitative variable; I denoting Item: data collection based on multiple vs. single item(s); and P denoting time Points: outcome assessment based on a "final time point" vs. "final and initial time points"). Among 122 RCTs, 32% did not have a clearly defined primary outcome. In multivariable analyses, a clearly defined primary outcome was associated with a binary variable (odds ratio (OR)=6.7; 2.5-17.7) and the composite variable "both blinding/placebo-controlled arm" (OR=4.5; 1.8-11.2); a significant statistical test was associated with a "final time point"-based outcome construction (OR=2.6; 1.2-5.5). Our study points to areas of improvement related to the definition and construction of the primary outcome in RCTs in dermatology.