Predicting outcome after acute ischemic stroke: an external validation of prognostic models.

Abstract

To externally validate two prognostic models predicting functional outcome and survival 100 days after acute ischemic stroke. Using prospectively collected data from 1,470 patients, the authors evaluated two previously developed models. Model I predicts incomplete functional recovery (Barthel Index <95) vs complete functional recovery with 11 variables, whereas model II predicts mortality vs survival with 3 variables. On admission to a participating hospital, patients were registered prospectively and included according to defined criteria. Within 72 hours, predictive variables under investigation were assessed. Follow-up was performed 100 days after the event. Model I correctly predicted 68.1% of the patients who had incompletely recovered or had died and 85.7% of the completely recovered patients, model II 46.9% of the patients who had died and 95.9% of the surviving patients. Both models performed better than the treating physicians' predictions made within 72 hours after admission. The resulting prognostic models are useful to correctly stratify treatment groups in clinical trials and to accurately predict the distribution of endpoint variables.