Prognostic indices in stroke.

Abstract

Two hundred and six patients with acute stroke admitted consecutively to District General Hospitals, were studied for a period of six months. Significance tests conducted singly detected 21 factors present during the first 48 h of stroke, which were related to outcome six months later in terms of both mortality and functional recovery. Among these significant factors were various measures of perceptual dysfunction, including Albert's Test. Multivariate statistical analysis which included discriminant analysis and linear logistic modelling, revealed six factors (Albert's Test Score, leg function, level of consciousness, arm power, weighted mental score and ECG changes) which were significantly and independently related to outcome. A statistical model based on these factors predicted functional outcome with an overall accuracy of 67 per cent and mortality with an accuracy of 83 per cent. This model provides a useful basis for stratification in future randomized controlled trial in stroke, and may have a role in the management of the individual stroke patient.