Consistency of the benefits of stroke units over years of operation: an 8-year effectiveness analysis.

BACKGROUND: The efficacy of stroke units (SU) has been amply demonstrated in randomised trials. However, no long-term studies analysed their effectiveness in daily practice over several years of operation.
METHODS: Observational study from the stroke data bank of our neurology ward that includes consecutive stroke patients hospitalised since the SU was established in 1995 until the year 2002 (8 years). Clinical effectiveness was analysed in terms of mortality on day 7, in-hospital mortality, functional state at discharge, length of stay, in-hospital complications and long-term hospitalisation rates on a yearly basis using for comparisons chi2 or Student t tests between the first and last 4 years of SU operation. Case-mix adjustments for baseline imbalances and stepwise multivariate logistic regression were also performed for comparative purposes.
RESULTS: 3,986 consecutive in-patients were included. No significant differences in the proportion of independent patients at discharge (73.9 vs. 74.5%; n.s.), length of stay (11.2 vs. 11.3 days; n.s.), in-hospital complications (25.0 vs. 25.9%; n.s.) or long-stay hospitalisation (6.3 vs. 6.4%; n.s.) were found comparing the first and last 4 years of SU operation. No significant differences in in-hospital mortality were found after adjustment by case-mix and length of stay.
CONCLUSIONS: SU effectiveness, in terms of length of stay, death or dependence and long-term hospitalisation, is sustained over the period of operation. Copyright 2006 S. Karger AG, Basel.