Risk stratification for predicting 30-day mortality of intracerebral hemorrhage

Int J Qual Health Care. 2009 Dec;21(6):441-7. doi: 10.1093/intqhc/mzp041. Epub 2009 Oct 14.

Abstract

Objective: The aim of this study was to develop a grading scale for predicting the 30-day mortality of spontaneous intracerebral hemorrhage (ICH) using initial evaluation data.

Design: Univariate and multivariate logistic regression models were used to identify independent risk factors and to construct a grading scale for predicting the outcome of ICH.

Setting: The Taichung Veterans General Hospital in Taichung, Taiwan.

Participants: Two hundred and ninety-three patients were diagnosed with spontaneous ICH between 1 January 2006 and 31 December 2007.

Intervention: Development of the simplified ICH score (sICH score) for predicting the 30-day mortality of ICH.

Main outcome measures: The discrimination of the prediction model was determined by measuring the accuracy, sensitivity, specificity and the area under the receiver operating characteristic curves (AUC).

Results: The accuracy of the sICH score was 80.5%, the sensitivity was 82.5% and the specificity was 80.2%. The AUCs are as follows: sICH score, 0.89 (0.84-0.94); ICH score, 0.74 (0.65-0.83) and ICH-GS, 0.74 (0.65-0.83).

Conclusions: The sICH score showed best discrimination among tested models. Also, it was easier for physicians without special training in neurology or radiology to use this scale. With statistical power and ease of use, the sICH score is a very suitable model for risk stratification of spontaneous ICH.

MeSH terms

  • Age Factors
  • Aged
  • Aged, 80 and over
  • Cerebral Hemorrhage / mortality*
  • Female
  • Humans
  • Logistic Models
  • Male
  • Middle Aged
  • Models, Statistical*
  • Prognosis
  • ROC Curve
  • Risk Assessment
  • Sex Factors