Head, Biostatistics & Data Science OMNY Health Atlanta, United States
Background: Plaque psoriasis (PsO) is a common chronic skin condition characterized by dry, itchy, raised skin patches covered with scales. Percent body surface area (BSA) and physician global assessment (PGA) are widely used disease activity (DA) metrics in routine clinical care. The Psoriasis Area and Severity Index (PASI) is a more laborious measurement of DA, typically used in research-based settings. Mild to moderate PsO tends to be treated with topical therapies, whereas more severe forms of disease can require advanced therapies (systemic, biologic, or phototherapy), which may carry adverse event risks.
Objectives: To evaluate the relationship between PsO DA metrics (BSA, PGA, and PASI) and initiation of advanced therapy (IAT) in the real-world setting.
Methods: PsO patients from 6 specialty dermatology networks within the OMNY Health Database were selected if they had BSA, PGA, and PASI assessments on the same day. Patients were characterized at their first DA assessment meeting these criteria. Prescription orders, administrations, and procedures were associated with DA assessments if they occurred within 7 days. IAT was defined as a prescription or administration of traditional systemic therapy (systemic corticosteroids and immunosuppressants), biologic therapy (adalimumab, brodalumab, certolizumab, etanercept, guselkumab, infliximab, ixekizumab, risankizumab, secukinumab, tildrakizumab, ustekinumab), or phototherapy. Multivariable logistic regression models were employed to assess independent and joint contributions of BSA, PGA, and PASI in predicting IAT.
Results: A total of 186 patients with 355 DA assessments were included (61% female, 45% ages ≥ 60 years, 64%/23%/13% West/Northeast/Other regions). Index BSA (mean: 31%, median: 20%), PGA (mean: 2.3, median: 3.0), and PASI (mean: 9.3, median: 6.8) exhibited skewed distributions. After controlling for patient characteristics, BSA exhibited a dose-response style relationship with IAT with odds ratio (OR) point estimates of 1.9 (BSA: 1-9%) and 2.4 (BSA ≥ 10%). A similar relationship between PASI and IAT was observed with ORs of 1.3 and 1.6 for PASI categories of 7-14 and ≥ 15, respectively. Moderate or severe PGA was strongly associated with IAT (OR: 3.1; 95% confidence interval [CI]: 1.6-6.0). When all DA metrics were considered together in the same model, the strong independent association of PGA with IAT (OR: 3.2; 95% CI: 1.4-7.6) attenuated the contributions of BSA and PASI, both losing their associations with the outcome.
Conclusions: PGA was the most predictive DA metric and retained its strong association with IAT in the presence of BSA and PASI. Given the simplicity of collecting PGA, its potential utility in the real-world setting as a single measure to characterize PsO DA could be considered.
