Project Summary Acute respiratory failure and sepsis are the leading causes of critical illness and death in hospitalized patients. Patients who become septic during hospitalization are at high risk of developing acute respiratory failure: up to 50% require mechanical ventilation. Sepsis patients with lung injury and respiratory failure have more than double the mortality of those without respiratory failure, but little is known about how to prevent this important complication in hospital-acquired sepsis. Guidelines recommend initiating antibiotics within one hour for all septic patients, but these recommendations rely on evidence primarily from patients with community-acquired sepsis, a population known to differ in characteristics, risks, and outcomes from patients with hospital-acquired sepsis. Furthermore, patients with hospital-acquired sepsis have well-documented delays in sepsis care and infrequently achieve quality targets for antibiotic timing. Critical knowledge gaps exist regarding determinants and implications of antibiotic timing among patients with hospital-acquired sepsis, limiting our ability to prevent the development of acute respiratory failure and other adverse outcomes in this high-risk population. The goals of this study are to identify the characteristics of patients’ clinical presentation associated with antibiotic timing in hospital-acquired sepsis, and to evaluate the role of antibiotic timing in reducing the risk of respiratory failure. First, using a retrospective cohort of approximately 1,800 inpatients with hospital-acquired sepsis, we will test the association of patient-level clinical factors at time of sepsis onset with time to antibiotic initiation. Second, we will test the association of antibiotic timing with the development of acute respiratory failure, using applied regression and instrumental variable methods. This project is supported by the Palliative and Advanced Illness Research (PAIR) Center of the University of Pennsylvania, which has an outstanding track record of supporting and advancing the careers of health services researchers in training. The candidate will be intensively mentored by a team with expertise in acute respiratory failure, sepsis, resource strain, and causal inference methods. Experiential training through this project will be complemented with coursework in applied regression and time-to-event analysis. Findings from this work will directly inform the development of an application for a K23 Mentored Career Development Award that aims to identify important sepsis clinical subgroups in order to improve the timeliness of sepsis care and our ability to prevent progression to respiratory failure.