(105) Hospital-level Variation in COVID-19 Treatment Among Hospitalized Adults in the United States: A Retrospective Cohort Study

Background: In the first 2 years of COVID-19 pandemic, studies have noted substantial variation across health centers in dexamethasone and remdesivir utilization. Little is known about hospital-level variation in COVID-19 drug treatments in the large healthcare organization in the United States (U.S.).

Objectives: To characterize variation in dexamethasone and remdesivir use over time between hospitals.

Methods: This retrospective cohort study used data from COVID-19 Consortium of HCA Healthcare and Academia for Research Generation (CHARGE), an electronic health records repository from a network of community healthcare facilities in the U.S. The study cohort included individuals hospitalized for COVID-19 across 163 hospitals between February 23, 2020 and October 31, 2021. Our primary outcome measure was the receipt of dexamethasone, remdesivir, and combined use of dexamethasone and remdesivir during hospital stay. We used two-level logistic regression models to determine the intraclass correlation coefficient (ICC) at the hospital level, adjusting for patient (sociodemographic, clinical and medication use) and hospital characteristics (size, teaching status, area deprivation index and region). The ICC shows the proportion of total variation in drug use accounted for by hospitals.

Results: Among 161,667 individuals hospitalized for COVID-19, 73.0% were treated with dexamethasone, 49.1% with remdesivir, and 45.0% with both dexamethasone and remdesivir. The proportion of variation in dexamethasone use was 12.7% (adjusted ICC, 0.127), 8.5% for remdesivir, and 11.3% for combined drug use, indicating low interhospital variation. In the fully adjusted models, between-facility variation in dexamethasone use declined from 34.1% in February-March 2020 to 11.3% in January-March 2021 and then increase to 17.3% in July-October 2021. The variation in remdesivir use remained relatively stable during the study period.

Conclusions: During the first two years of the pandemic, there was relatively consistent use of dexamethasone and remdesivir across the hospitals examined. Consistent adoption and implementation of treatment guidelines across the hospitals examined may have led to relatively low interhospital variation in drug use.