Karen C Johnston1, Guofen Yan. 1. Departments of Neurology, University of Virginia, Charlottesville, VA, USA. kj4v@virginia.edu
Abstract
BACKGROUND AND PURPOSE: The Acute Physiology, Age, Chronic Health Evaluation score for critically ill patients has provided a method of predicting outcome using major physiological variables. We hypothesized that a physiology score for stroke patients (Acute Physiology of Stroke Score [APSS]) when added to a validated clinical prediction model would improve outcome prediction. METHODS: The APSS was developed and validated using multivariable logistic regression. It was added to a previously validated clinical model to assess for increased area under the curve in predicting 3-month outcome. RESULTS: The bootstrap-validated bias-corrected area under the curve for just the APSS predicting alive/dead at discharge was 0.753. The clinical model area under the curve ranged from 0.77 to 0.88 and the addition of the APSS resulted in areas under the curve of 0.77 to 0.89. CONCLUSIONS: These data suggest that the APSS is related to 3-month clinical outcome in patients with ischemic stroke. However, the APSS adds no clinically relevant additional predictive value when added to our previously validated clinical prediction model.
BACKGROUND AND PURPOSE: The Acute Physiology, Age, Chronic Health Evaluation score for critically illpatients has provided a method of predicting outcome using major physiological variables. We hypothesized that a physiology score for strokepatients (Acute Physiology of Stroke Score [APSS]) when added to a validated clinical prediction model would improve outcome prediction. METHODS: The APSS was developed and validated using multivariable logistic regression. It was added to a previously validated clinical model to assess for increased area under the curve in predicting 3-month outcome. RESULTS: The bootstrap-validated bias-corrected area under the curve for just the APSS predicting alive/dead at discharge was 0.753. The clinical model area under the curve ranged from 0.77 to 0.88 and the addition of the APSS resulted in areas under the curve of 0.77 to 0.89. CONCLUSIONS: These data suggest that the APSS is related to 3-month clinical outcome in patients with ischemic stroke. However, the APSS adds no clinically relevant additional predictive value when added to our previously validated clinical prediction model.