Background: A better understanding of the clinical features associated with prolonged hospitalization in acute respiratory failure may allow for better-informed care planning.Research Question: What are the incidence, mortality, cost, and clinical determinants of prolonged hospitalization among patients with acute respiratory failure (ARF)?Study Design and Methods: Using the National Inpatient Sample data from 2004 to 2014, we identified adults 18 years and older with International Classification of Diseases, 9th edition (ICD-9) codes for ARF requiring mechanical ventilation for at least 2 days (ICD-9 518.81 or 518.82, 96.7 or 96.04, and 96.05). Outcomes studied included incidence, in-hospital mortality, cost of hospitalization, and associated patient-level and hospital-level characteristics. Trends were assessed by logistic regression, linear regression, and general linear modeling with Poisson distribution.Results: Of the 5,539,567 patients with ARF, 77,665 (1.4%) had a prolonged length of stay (pLOS), defined as ≥ 60 days. Among those with pLOS, 52,776 (68%) survived to discharge. Over the study period, the incidence of pLOS decreased by 48%, in-patient mortality decreased by 18%, per-patient cost of care rose, but the percentage of the total cost of ARF care consumed by patients with pLOS did not significantly decrease (P = .06). Prolonged LOS was more likely to occur in urban teaching hospitals (OR, 6.8; 95% CI, 4.6-10.2; P < .001), hospitals located in the northeastern United States (OR, 3.6; 95% CI, 3.0-4.3; P < .001), and among patients with Medicaid insurance coverage (OR, 2.1; 95% CI, 1.9-2.4; P < .001).Interpretation: From 2004 to 2014, incidence and mortality decreased among patients with ARF and pLOS, and although per-patient costs rose, the percentage of total cost of care remained stable. There is substantial variation in length of stay for patients with ARF by US region, hospital teaching status, and patient insurance coverage. [ABSTRACT FROM AUTHOR]