Variation in isotretinoin use for acne in an integrated health system: an ecological study

Abstract Isotretinoin is highly effective for treating severe nodulocystic acne but is associated with serious adverse effects and complexity in its use. Understanding variation in isotretinoin prescribing patterns is limited. We conducted a cross-sectional ecological study to measure variation in and predictors of isotretinoin prescription rates within a large integrated health system in Northern California. Patients aged 13–40 years old with ≥ 2 visits with an acne diagnosis from 2016–2020 were identified. Individual patient-level data were aggregated at the level of 15 administrative “service areas” for analysis. We examined 20 candidate predictors potentially associated with isotretinoin prescription rates. The primary outcome was the proportion of patients with acne who received ≥ 1 prescription for isotretinoin during the study period. Analyses showed that rates of isotretinoin prescription varied from 7.7% in the lowest-prescribing service area to 17.3% in the highest-prescribing area (RR = 2.2, 95%Confidence Interval: 2.0-2.6). Three candidate predictors were found to be independently and statistically significantly associated with higher isotretinoin prescription rates: a shorter elapsed time between the index date and the first isotretinoin prescription (beta=-0.07, p < 0.006), a higher rate of prior oral antibiotic prescriptions (beta = 0.40, p < 0.004), and a higher percentage of patients whose initial isotretinoin prescription was written by advanced practice practitioners (beta = 0.36, p = 0.05). In sum, we found large variation in rates of isotretinoin prescription among the healthcare system's 15 service areas. Three predictor variables were independently and significantly associated with isotretinoin prescriptions and provide potentially actionable information that may help improve appropriateness and consistency in providing this effective but potentially burdensome therapy.