Abstract P192: Provider And Health System Level Variability In Initial Antihypertensive Intensity: A Mixed Modeling Approach In Newly Treated Patients With Hypertension

Abstract

Background: Initial antiHTN regimen intensity varies across patients, presumably due to factors like age, comorbidities, or baseline BP. However, the extent to which initial regimen intensity is influenced by patient factors, versus provider characteristics or preferences, and health system contexts, is unknown. Methods: We conducted a cross-sectional study of adults with newly-treated HTN in the OneFL+ Consortium with Medicaid or Medicare claims linked to EHR data from their treatment initiation visit. We used linear mixed modeling with prescriber nested within health system as random effects to estimate variability in total therapeutic intensity score (TTIS) – a patient’s total daily dose (TDD) divided by the recommended max TDD for a drug, summed across the initial regimen. We modeled patient- (demographics, BP, BMI, comorbidities) and provider-level (sex, specialty) factors as fixed effects and calculated intraclass correlation coefficients (ICC), to estimate relative variance in TTIS attributable to provider preference (ICC P ) and health system (ICC H ). Results: A total of 3970 patients (mean ± SD age, 59 ± 18.4 y; 58% women; 38% Black) were included. In the full model, 11.4% of the observed TTIS variation was attributed to provider (p<0.0001) and 3.9% to health system context (p=0.17), leaving 84.7% of TTIS variation within patients. The most influential factors in TTIS variation were Black race, male sex, and comorbidity profile (T2DM, CKD) (Table). Conclusion: Provider preference accounted for a significant but modest amount of variation in initial antiHTN regimen intensity. Most of the variance in TTIS was explained by patient factors.