A systematic review of 454 randomized controlled trials using the Dermatology Life Quality Index: experience in 69 diseases and 43 countries.

ABSTRACT

BACKGROUND: Over 29 years of clinical application, the Dermatology Life Quality Index (DLQI) has remained the most used patient-reported outcome (PRO) in dermatology due to its robustness, simplicity and ease of use. OBJECTIVES: To generate further evidence of the DLQI's utility in randomized controlled trials (RCTs) and to cover all diseases and interventions. METHODS: The methodology followed PRISMA guidelines and included seven bibliographical databases, searching articles published from 1 January 1994 until 16 November 2021. Articles were reviewed independently by two assessors, and an adjudicator resolved any opinion differences. RESULTS: Of 3220 screened publications, 454 articles meeting the eligibility criteria for inclusion, describing research on 198 190 patients, were analysed. DLQI scores were primary endpoints in 24 (5.3%) of studies. Most studies were of psoriasis (54.1%), although 69 different diseases were studied. Most study drugs were systemic (85.1%), with biologics comprising 55.9% of all pharmacological interventions. Topical treatments comprised 17.0% of total pharmacological interventions. Nonpharmacological interventions, mainly laser therapy and ultraviolet radiation treatment, comprised 12.2% of the total number of interventions. The majority of studies (63.7%) were multicentric, with trials conducted in at least 42 different countries; 40.2% were conducted in multiple countries. The minimal clinically importance difference (MCID) was reported in the analysis of 15.0% of studies, but only 1.3% considered full score meaning banding of the DLQI. Forty-seven (10.4%) of the studies investigated statistical correlation of the DLQI with clinical severity assessment or other PRO/quality of life tools; and 61-86% of studies had within-group scores differences greater than the MCID in 'active treatment arms'. The Jadad risk-of-bias scale showed that bias was generally low, as 91.8% of the studies had Jadad scores of ≥ 3; only 0.4% of studies showed a high risk of bias from randomization. Thirteen per cent had a high risk of bias from blinding and 10.1% had a high risk of bias from unknown outcomes of all participants in the studies. In 18.5% of the studies the authors declared that they followed an intention-to-treat protocol; imputation for missing DLQI data was used in 34.4% of studies. CONCLUSIONS: This systematic review provides a wealth of evidence of the use of the DLQI in clinical trials to inform researchers' and -clinicians' decisions for its further use. Recommendations are also made for improving the reporting of data from future RCTs using the DLQI.