In-Patient Hospital Costs of Stroke: A Focused Literature Review.

status. Methods: New ischemic stroke (ICD-10 code I63) patients (no previous stroke within one year) between April 1, 2006 and March 31, 2009 (N= 4,350) were followed for one year using hospital Discharge Abstract Database. The severity of the stroke was obtained from the ambulatory care database. Median hospital costs by CMG+ group were obtained from Alberta Health. Hospitals were classified as teaching, community large, community medium, and community small hospitals. Hospitals were also classified as comprehensive stroke centre, urban and rural primary stroke centres, and other urban and rural hospitals. The adjusted risk factors in Bayesian Model included sex, age, all disease-specific co-morbidities, and disease severity. The results for four hospital types and five stroke center categories were calculated using the observed/expected approach. Results: The 30 days mortality rates (95% CI) were lowest for teaching hospitals 10.1% (9.0%-11.2%) and large community hospitals (10.0%; 8.3%-11.8%), and the small community hospitals had the highest mortality rates (12.8%; 9.9%-15.8%). The mean LOS (95% CI) varied from 21.7 (20.9-22.6) days in teaching hospitals to 34.2 (28.6-41.0) days in community medium hospitals. The community medium hospitals had significantly higher costs ($62,400; $49,900-$78,000) than the community large hospitals ($32,900; $29,900-$36,200) and teaching hospitals ($37,000; $34,900-$39,200). Both comprehensive stroke and urban stroke centers had lower 30 day mortality rates (95% CI): 9.9% (8.8%-11.1%) and 9.7% (7.3%-12.0%); shorter LOS 21.6 (20.7-22.5) and 25.0 (22.7-27.6) days; and medium levels of costs $39,300 ($36,100-$40,700), compared to other hospitals. ConClusions: The study shows the hospital type and stroke centre had limited effects on the mortality but significant impact on LOS and costs.