Provider-Level and Specialty Variation in ARNI and SGLT2i Use After Heart Failure Hospitalization

Background: Angiotensin receptor-neprilysin inhibitors (ARNI) and sodium-glucose cotransporter-2 inhibitors (SGLT2i) are incompletely utilized in heart failure with reduced ejection fraction (HFrEF). The degree to which use varies at the individual level and within specialties is unknown. We aimed to characterize provider-level and specialty-level variation in ARNI and SGLT2i prescribing after HF hospitalization. Methods: This retrospective cohort study included Veterans Affairs (VA) providers seeing patients within 6 months of HFrEF hospitalization 2017-2023. We used descriptive statistics and site-adjusted multilevel logistic regression to examine prescribing by medical specialty, provider type, and provider sex. Adjusted median odds ratios (mOR) and empirical (unadjusted) provider-level rate ratios (RRs) by percentile were used to describe provider-level variation. Clinical pharmacy visits were assessed as an independent factor and as an interaction with specialty. Results: Among 7,946 providers caring for 76,795 patients across 211,267 outpatient encounters, provider-level variation remained substantial after adjustment for patient characteristics and site-level clustering (mOR 1.45 for ARNI; 1.46 for SGLT2i). Within every specialty, providers ranged from near-zero to near-100% use. Relative to the median provider in the same specialty, RRs ranged from 0-0.31 (5th percentile), 0.37-0.71 (25th), 1.38-1.66 (75th), to 1.8-3.2 (95th). Cardiology visits carried the highest adjusted odds of ARNI (OR 2.30, 95% CI 2.18-2.44) and SGLT2i use (OR 1.99, 95% CI 1.88-2.11) and most new starts. Pharmacist visits were independently associated with greater use of ARNI (OR 1.14, 1.06-1.23) and SGLT2i (OR 1.15, 1.06-1.24), with no effect modification by specialty. In early 2023, 58.4% (ARNI) and 40.2% (SGLT2i) of post-hospitalization HF visits had untreated patients. Conclusions: Provider-level variation in ARNI and SGLT2i use after guideline incorporation was large within every specialty and independent of site effects. Though specific specialties have higher use, individual providers, not specialty alone, may be an important target for efforts to close gaps in therapy.