(R) Agreement and Correlation Between Different Topical Corticosteroid Potency Classification Systems

Publication Number: 642

Background: Topical corticosteroids (TCS) are used to treat inflammatory skin diseases and are available in a wide range of potencies that influence their effectiveness and safety. The absence of a universal potency classification system hinders pharmacoepidemiologic studies of TCS by limiting comparisons across studies, robust exposure classification, and clinical interpretation.

Objectives: To classify TCS formulations using three potency classification systems and evaluate their agreement and correlation to inform future pharmacoepidemiology research.

Methods: We compiled a comprehensive list of TCS formulations using sources published in the literature, the Ontario Drug Benefit Formulary, a Cochrane Review, and the Anatomical Therapeutic Classification (ATC) of the World Health Organization. TCS were classified using (1) the 7-category US classification system; (2) a 4-category classification from a recent Cochrane Review largely based on the UK formulary; and (3) the 4-category ATC classification. To facilitate comparisons, we consolidated the 7-category US classification system into 4 categories and computed Cohen’s weighted kappa (kw) to assess agreement. We assessed correlation between the systems (without consolidating the 7-category classification) using Spearman rank correlation coefficients (rs).

Results: We included 232 unique TCS formulations (ATC: N = 231; US Classification: N = 232; Cochrane Review: N = 89). Overall, there was low-to-moderate agreement but strong correlation between the classifications. Comparisons with the consolidated US classification demonstrated weak agreement with the ATC (N = 231; κw 0.53, 95% confidence interval [CI], 0.45-0.60) and moderate agreement with the Cochrane Review (N = 89; κw 0.60, 95% CI, 0.48-0.73); there was weak agreement between the ATC and Cochrane Review (N = 88; κw 0.58 (0.46-0.71). The US classification was strongly correlated with the ATC (N = 231; rs = 0.77, 95% CI, 0.71-0.82) and Cochrane Review (N = 89; rs = 0.74, 95% CI, 0.62-0.82). There was also a strong correlation between the Cochrane Review and ATC (N = 88; rs = 0.71, 95% CI, 0.58-0.80).

Conclusions: We used multiple resources to classify TCS into three potency classifications. Our findings demonstrate that while correlated, the systems are often incongruent and could yield different results in pharmacoepidemiologic studies. We offer a comprehensive list of TCS potency classifications to support future research; investigators should leverage this resource to contribute robust evidence on the use of TCS, be transparent in their classification approach, and consider alternative potency definitions in sensitivity analyses.