Credibility and Generalization of the Minimally Important Difference Concept in Dermatology: A Scoping Review
Authors:
Speeckaert, R., Belpaire, A., Herbelet, S., Lambert, J., and van Geel, N.
Abstract:
IMPORTANCE: The minimally important difference (MID) represents the point at which a difference in an outcome measure (eg, Dermatology Life Quality Index) is important enough that it warrants a change in treatment, and, to the authors' knowledge, the robustness and limitations of MIDs have not been thoroughly evaluated in skin diseases. The MID is increasingly used in clinical trials to demonstrate that an intervention is worthwhile for patients; furthermore, MIDs also contribute to sample size calculations in clinical trials, influence treatment guidelines, and can guide clinicians to modify treatment.
OBJECTIVE: To evaluate the credibility and generalization of MIDs for patient-reported outcome measures (PROMs) in skin disorders.
EVIDENCE REVIEW: A systematic search was conducted in PubMed and Embase for all original articles using the MID concept for skin disorders from inception to December 29, 2021. The credibility of MIDs obtained via an anchor-based approach (eg, global rating of change scale) was assessed with a previously developed credibility instrument. The validity of generalizing established MIDs to other patient groups was evaluated based on the diagnosis and the patient characteristics.
FINDINGS: A total of 126 articles were selected, and 84 different MIDs were identified for PROMs. A total of 13 of 84 MIDs (15.5%) for PROMs displayed acceptable credibility. The anchors used had varying capacity to assess minimal important changes from a patient's perspective and were deemed inappropriate for this purpose in 52 of 84 cases (61.9%). Correlations between the anchors and PROMs were frequently not determined (39 of 84; 46.4%). The time interval for anchor questions assessing a change in the experienced disease burden was not optimal for 10 of 32 transition anchors (>3 months), introducing potential recall bias. Previously reported MIDs were widely used to examine relevant changes in other study populations. However, the diagnosis and disease severity were different from the original MID population in 39 of 70 (55.7%) and 45 of 70 (64.3%) cases, respectively.
CONCLUSIONS AND RELEVANCE: In this scoping review, only a minority of MIDs for PROMs demonstrated sufficient credibility in dermatology. Inappropriate generalization of previously reported MIDs to patient populations with different disease characteristics was found to be a major concern. Furthermore, the study supported the use of multiple anchors and encouraged consistent reporting of the correlation between changes in the anchor and changes in the outcome measures.
