Comparative Effectiveness of Topical Drugs in Dermatologic Priority Diseases: Geometry of Randomized Trial Networks

Abstract

Among the 100 initial priority topics for comparative effectiveness research, three concern topical drugs in the following dermatologic diseases: psoriasis, chronic lower-extremity wounds (CLEWs), and acne vulgaris (AV). Our objective was to explore the geometry of the corresponding networks of randomized controlled trials (RCTs). We performed a review of RCTs on topical drugs in psoriasis, CLEWs, and AV. We searched MEDLINE, Embase, and CENTRAL for published trials from 2007 to 2012 and ClinicalTrials.gov for unpublished trials registered since 2011. RCTs comparing at least one topical treatment with any active or inactive comparator, regardless of RCT design and outcomes, were eligible. We produced network graphs (each node representing a treatment and links between nodes representing trials) and tested for co-occurrence (preference or avoidance of specific comparisons). We included 60 RCTs on psoriasis (14,255 patients) and 19 registered RCTs, 50 of CLEWs (5,916 patients) and 7 registered RCTs, and 90 of AV (22,984 patients) and 21 registered RCTs. Head-to-head comparisons were made in 78%, 32%, and 57% of published RCTs of these conditions, respectively. The co-occurrence test suggested that no specific head-to-head comparison was significantly preferred or avoided (P-value=0.53, 0.20, and 0.57, respectively). This study has limitations, the main being that the search period was restricted to 5 years. In conclusion, more comparative effectiveness trials are needed for CLEWs, for which head-to-head comparisons are fewer than those for psoriasis and AV.