Design and methodological issues of within-person (split-body) randomized controlled trials evaluating a topical treatment: A systematic review.

Background Topical drugs are often used as first-line treatment for dermatological conditions. A within-person design may then be well adapted: it consists of randomizing lesions/body sites rather than patients, which are then concomitantly treated by the different drugs compared, reducing inter-group variability and therefore requiring fewer patients than the classical parallel-group trial. Objectives The aim of this review was to provide a methodological overview of within-person randomized trials (WP-RCTs) in dermatology. Methods We searched for eligible trials published between 2017 and 2021 in MEDLINE, Embase and Central in dermatology journals and the 6 highest-impact-factor general medical journals. Two authors selected publications and extracted data independently. Results From 1034 articles identified, we included 54 WP-RCTs, mainly for acne vulgaris, psoriasis, actinic keratosis and atopic dermatitis. In most of the trials, patients had only 2 lesions/body sites. In none of the trials did we detect a potential carry-across effect (known to be the major methodological problem in WP-RCTs). Twelve studies reported a care provider applying the treatment, and in 26 studies, the patients themselves applied the treatment. Finally, we also highlight statistical issues for the statistical analysis: overall, 14 (26.9%) studies used a test for independent observations, thus ignoring the between-lesion correlation. Conclusion Our systematic review highlights that despite the publication of the CONSORT checklist extension for WP-RCTs in 2017, this design is rarely used, and when it is, there are methodological and reporting concerns.