ntibiotic guideline concordance and area deprivation in the US emergency departments, 2015-2024.

BACKGROUND: Older adults (≥65) often receive antibiotics in emergency departments (ED), where guideline concordance is low with both overuse and underuse; whether community disadvantage drives these patterns is unknown. METHODS: We analyzed electronic health records from 1,318,281 ED encounters among 790,562 adults ≥65 (2015-2024). Outcomes were guideline concordance and components (overuse, underuse). To address omitted variables and reverse causality, we used two-stage instrumental-variable models with Social Vulnerability Index (SVI) and Social Deprivation Index (SDI) as exposures, instrumented by maximum state Earned Income Tax Credit, adjusting for patient, facility, state, and time covariates. RESULTS: 16.8% of encounters were non-concordant (overuse 9.5%; underuse 7.2%). Concordance improved to 83.8% in 2024. In IV models, each 10-point increase in SVI reduced concordance by 6.37 percentage points (pp) (SE 0.69) and each 10-point increase in SDI by 1.84 pp (SE .20; P < .0005). Declines reflected higher underuse (+0.43 pp per 10-point SVI; +0.12 pp per 10-point SDI) with less overuse (-0.10 and -0.30 pp). Concordance was higher for Black (+12.90 pp vs White) and Medicaid-insured patients (+5.40 pp vs commercial). Overuse was higher in rural (+7.90 pp vs metropolitan) and academic EDs (+5.00 pp vs non-academic), whereas underuse was more common in metropolitan (+42.50 pp) and non-academic EDs (+32.70 pp). Spatial analyses found high deprivation/low concordance in the South and low deprivation/high concordance in the Midwest. CONCLUSIONS: Community disadvantage may causally predict lower guideline-concordant antibiotic prescribing for older adults, primarily via underuse. Stewardship should address underuse and overuse and prioritize disadvantaged regions where gaps are greatest.