Abstract Rationale Emerging evidence, guidelines, and the pandemic experience may change corticosteroid use in severe pneumonia. This study integrated a quantitative analysis of national practice data with a qualitative analysis of clinician interviews to understand 10-year trends in practice across the VA healthcare system. Methods We assessed the use of systemic corticosteroids within 24 hours among all emergency department (ED) admissions to intensive care units at 119 VA Medical Centers from 1/1/2015-12/31/2024 with an initial pneumonia diagnosis and positive chest imaging. To distinguish true practice change from changes explained by patient risks, we developed two mixed effects logistic regression models to predict marginal probability of treatment using patient data from two periods: early(2015) and late(2024) using 60 patient factors (demographics, comorbidities, vital signs, and laboratory results). We visualized trends in observed versus model-predicted treatment. We compared patient factors that were most predictive of treatment in the early versus the late model. Qualitative interviews with 29 clinicians from eight VA facilities (2023-2024) explored experiences and influences on pneumonia management and were coded by two trained qualitative analysts. Results Among 38,383 ICU admissions across 10 years, the use of corticosteroids increased from 41% in 2015 to 46% in 2024. Trends were nearly completely explained by patient risk (Figure 1). Strong positive predictors of corticosteroid use consistent across both periods were COPD/asthma (early adjusted odds ratio 3.0495%CI 2.47-3.74; late 3.102.50-3.84), immunocompromised status (early 3.042.34-3.96); late 2.26[1.74-2.93), and severe hypoxemia (early 1.661.26-2.17; late 2.191.71-2.82). SARS-CoV2 became a new positive predictor in the late period (5.192.79,9.64). Influenza was not predictive in the early period (0.790.40-1.60) but was a strong positive predictor in the late period (3.041.59-5.81). Conversely, sepsis became a strong negative predictor in the late period (0.54 0.39-0.75), while congestive heart failure remained a weak negative predictor in both early (0.700.56-0.87) and late periods (0.790.63-0.98). Interviewed clinicians reported several barriers to corticosteroid use, including mixed evidence, recommendations, and peer practice, concerns about efficacy and safety, and an importance in but challenges to individualizing treatment based upon patient factors including stage of illness and type of host-pathogen pattern. Conclusion Corticosteroid use for severe community-acquired pneumonia has increased only slightly over the past decade and was more associated with changes in patient risk than changes in practice. However, factors influencing treatment have shifted. Providers reported mixed attitudes and challenges to individualization as important barriers to use. This abstract is funded by: US Department of Vetaran Affairs
Ju et al. (Fri,) studied this question.