Measuring atopic dermatitis severity in randomized controlled clinical trials: what exactly are we measuring?

Well-designed clinical trials are a fundamental aspect of evidence-based medicine. Such trials are dependent on the use of valid, reliable, and relevant outcome measures. Wide variation in outcome methodology can have important detrimental effects on the correct interpretation and comparison of results. The objective of this study was to describe the variation in outcome methodology in randomized controlled trials of therapeutic interventions for atopic dermatitis published between January 1994 and December 2001. Of the 93 eligible randomized controlled trials identified using a systematic electronic database search strategy, 85 (91%) incorporated an objective measurement of clinical signs. Only 23 (27%) of these trials used a published severity scale, however. The remainder used either modified versions of published scales (14%) or unnamed scales with no data on validity or reliability (59%), although unpublished scales were used significantly less frequently over the last 2 y compared to previously (21%vs 74%, p<0.01). There was lack of consensus on which clinical features best reflect disease severity, with 31 different descriptions of clinical signs being used across all scoring systems. Fifty-six different "objective" clinical scales were identified. Patient symptoms were recorded in 80 trials (86%) and disease extent in 62 trials (67%). Quality of life was measured in only three trials (3%). This wide variation in outcome methodology is hindering evidence-based practice, and the widespread use of unvalidated outcome measures is a potential source of bias and inaccuracy. More emphasis should be placed on measuring things that are important to patients such as symptoms and quality of life.