Predictors of recovery in US patients hospitalized for viral lower respiratory tract disease: an analysis of linked electronic health records and claims data

<bold>Background:</bold> Patients hospitalized with viral LRTD may have a long recovery period. Predictors of recovery are poorly understood. <bold>Objective:</bold> Quantify predictors of recovery in US patients with severe viral LRTD. <bold>Methods:</bold> This retrospective cohort study used Optum’s de-identified Market Clarity Data which deterministically links medical and pharmacy claims with EHR data from providers across the continuum of care. Included adults had a viral LRTD hospitalization diagnosis (2015–2023), 12 months of continuous pre- and post-admission enrolment and 12 months of post-discharge survival. A modified PICS definition was used to model recovery across physical and mental domains. A multivariable ordinal logistic regression model quantified ORs (95% CI). <bold>Results:</bold> 42 800 patients were included. 50.2% and 37.6% of patients met PICS definition in one or both domains, respectively. Discharge to long-term care/nursing (OR:1.61, 95% CI:1.54–1.68), pre-admission diagnosis of anxiety/depression (OR:1.51, 95% CI:1.48–1.55), pre-admission high-dose steroid use (OR:1.38, 95% CI:1.10–1.73), and IMV/ECMO (OR:1.31, 95% CI:1.24–1.39) were associated with increased PICS score (Fig). <bold>Conclusion:</bold> Awareness of patients’ medical history and reducing invasive hospital interventions may help optimize recovery and prevent downstream sequalae from critical illness after viral LRTD hospitalization. <fig><object-id>erj;66/suppl_69/PA3134/F1</object-id><object-id>F1</object-id><object-id>F1</object-id><graphic></graphic></fig>