Objective: To determine predictors of in-hospital and long-term mortality and length of stay in patients admitted to the neurosciences critical care unit.
Design: Retrospective analysis of a prospectively collected database.
Setting: Neurosciences critical care unit of a large academic tertiary care hospital.
Patients: Adult patients (n = 2381) admitted to our neurosciences critical care unit from January 1997 to April 2000.
Interventions: Introduction of a specialized neurocritical care team.
Measurements and main results: Data obtained from the database included demographics, admission source, length of stay, neurosciences critical care unit and hospital disposition, admission Acute Physiology and Chronic Health Evaluation (APACHE) III score, and principal and secondary diagnoses. The introduction of a neurocritical care team in September 1998 was also collected, as was death at 1 yr after admission. Univariate analysis was carried out using Student's t-test, Mann-Whitney U test, or chi-square test (significance, p < .05). A logistic regression model was used to create a prediction model for in-hospital and long-term mortality. A general linear model was used to determine predictors of length of stay (after log transformation). Independent predictors of in-hospital mortality included APACHE III (odds ratio, 1.07 [1.06-1.08]) and admission from another intensive care unit (odds ratio, 2.9 [1.4-6.2]). The presence of a neurocritical care team was an independent predictor of decreased mortality (odds ratio, 0.7 [0.5-1.0], p = .044). Admission after the neurocritical care team was implemented was associated with reduced length of stay in both the neurosciences critical care unit (4.2 +/- 4.0 vs. 3.7 +/- 3.4, p < .001) and the hospital (9.9 +/- 8.0 vs. 8.4 +/- 6.9, p < .0001). There was no difference in readmission rates to the intensive care unit or discharge disposition to home before and after the neurocritical care team was established. The availability of the neurocritical care team was not associated with significant changes in long-term mortality. Factors independently associated with long-term mortality included female gender, admission from another intensive care unit, APACHE III score, and being moderately disabled before admission.
Conclusion: Introduction of a neurocritical care team, including a full-time neurointensivist who coordinated care, was associated with significantly reduced in-hospital mortality and length of stay without changes in readmission rates or long-term mortality.