Derivation and validation of the prolonged length of stay score in acute stroke patients

• Published in: Neurology Vol. 74; no. 19; p. 1511
• Main Authors: Koton, S, Bornstein, N M, Tsabari, R, Tanne, D
• Format: Journal Article
• Language: English
• Published: United States 11.05.2010
• Subjects: Activities of Daily Living; Acute Disease - economics; Acute Disease - therapy; Aged; Aged, 80 and over; Atrial Fibrillation - rehabilitation; Cerebral Hemorrhage - rehabilitation; Cohort Studies; Confidence Intervals; Critical Pathways; Emergency Medical Services; Female; Health Care Costs - statistics & numerical data; Health Care Costs - trends; Hospitalization - statistics & numerical data; Humans; Intensive Care Units - economics; Intensive Care Units - statistics & numerical data; Intracranial Thrombosis - rehabilitation; Israel; Length of Stay - statistics & numerical data; Length of Stay - trends; Male; Middle Aged; Odds Ratio; Outcome Assessment (Health Care); Palliative Care; Patient Discharge; Prognosis; Recovery of Function; Severity of Illness Index; Sex Factors; Stroke - economics; Stroke - therapy; Time Factors; Treatment Outcome; Israel
• ISSN: 1526-632X
• DOI: 10.1212/wnl.0b013e3181dd4dc5

Length of stay (LOS) is the main cost-determining factor of hospitalization of stroke patients. Our aim was to derive and validate a simple score for the assessment of the risk of prolonged LOS for acute stroke patients in a national setting. Ischemic stroke (IS) and intracerebral hemorrhage (ICH) patients in the National Acute Stroke Israeli Surveys (NASIS 2004 and 2007) were included. Predictors of prolonged LOS (LOS > or =7 days) in the NASIS 2004 (n = 1,700) were identified with logistic regression analysis and used for the derivation of the Prolonged Length of Stay (PLOS) score. The score was validated in the NASIS 2007 (n = 1,648). Median (interquartile range) LOS was 6 (3-10) days in the derivation cohort (42.3% prolonged LOS) and 5 (3-8) in the validation cohort (35.7% prolonged LOS). The derivation cohort included 54.8% men, 90.8% IS and 9.2% ICH, with a mean (SD) age of 71.2 (12.5) years. Stroke severity was the strongest multivariable predictor of prolonged LOS: odds ratio (95% confidence interval [CI]) increased from 2.6 (2.0-3.3) for NIH Stroke Scale score (NIHSS) 6-10 to 4.9 (3.0-8.0) for NIHSS 16-20, compared with NIHSS < or =5. Stroke severity and type, decreased level of consciousness on admission, history of congestive heart failure, and prior atrial fibrillation were used for the derivation of the PLOS score (c statistics 0.692, 95% CI 0.666-0.718). The score performed similarly well in the validation cohort (c statistics 0.680, 95% CI 0.653-0.707). A simple prolonged length of stay score, based on available baseline information, may be useful for tailoring policy aimed at better use of resources and optimal discharge planning of acute stroke patients.