Background: Empirical antibiotic use is common in hospitalized patients with COVID-19 despite the low prevalence of bacterial coinfection, raising concerns about antimicrobial resistance and inappropriate prescribing. However, the comparative safety of commonly used antibiotic classes in this context remains unclear. Methods: We conducted a retrospective cohort study using real-world clinical data standardized through the Observational Medical Outcomes Partnership Common Data Model from 1 January 2020 to 31 May 2025. Adults with confirmed COVID-19 who were administered empirical antibiotics on the admission day were included. Empirical antibiotic exposure was categorized as third-generation cephalosporins (3GCs), fluoroquinolones, or aminopenicillins with β-lactamase inhibitors (PEN–BLis). Results: Compared with 3GCs, fluoroquinolone use was associated with significantly higher risks of mechanical ventilation (hazard ratio [HR]: 1.50; 95% confidence interval [CI]: 1.12–2.00), ICU admission (HR: 1.54; 95% CI: 1.10–2.15), vasopressor use (HR: 1.35; 95% CI: 1.11–1.63), all-cause in-hospital mortality (HR: 1.55; 95% CI: 1.22–1.96), and the composite outcome (HR: 1.32; 95% CI: 1.10–1.60). PEN–BLis showed no significant differences from 3GCs across outcomes. Conclusions: Empirical fluoroquinolone use at COVID-19 admission may be associated with greater risks of critical care interventions and in-hospital mortality compared to those of 3GCs. These findings highlight the need for careful patient selection and clinical judgment when initiating empirical antibiotic therapy for viral respiratory infections such as COVID-19.

Background/Objectives: Real-world antidepressant use for depressive disorder varies across clinical settings. This study aimed to characterize prescribing pathways and evaluate the comparative safety of frequently prescribed antidepressants in a large multicenter cohort with newly diagnosed depressive disorder. Methods: We conducted a retrospective multicenter observational study using electronic health record data standardized to the OMOP Common Data Model (1999–2026) in adults initiating antidepressants. Prescribing patterns among 13,381 patients were visualized using sunburst plots and Sankey diagrams. Comparative safety, escitalopram vs. sertraline and trazodone, was evaluated across 18 hospitals (n = 44,542) for nine clinical safety outcomes using propensity score-matched Cox proportional hazards models and empirically calibrated random-effects meta-analysis. Results: Selective serotonin reuptake inhibitors, particularly escitalopram (21.4%, n = 2868), were the preferred antidepressant. Common adjunctive medications included trazodone (14.6%, n = 794) and quetiapine (14.0%, n = 761), though utilization varied across hospitals. No clear differences were detected for most safety outcomes between cohorts. However, empirically calibrated QTc prolongation risk was significantly increased with sertraline relative to escitalopram (HR 1.75, 95% CI 1.10–2.78), a signal requiring further confirmation. Conclusions: This study demonstrates substantial institutional variation exists in real-world antidepressant treatment pathways and combination strategies. While comparative safety profiles were broadly comparable across evaluated outcomes, the isolated QTc prolongation signal for sertraline warrants targeted external confirmation. Furthermore, trazodone and escitalopram cohorts should not be regarded as clinically interchangeable first-line options, supporting individualized, evidence-based antidepressant selection in routine clinical practice.