Predictive Models for In-Hospital Stroke Mortality

Abstract

Development of models that can predict potential stroke mortality is important for research purposes, but also to inform families correctly. Using data from Brasov Stroke Registry, we created predictive models for inhospital acute stroke mortality that include demographic (age, gender), history (arterial hypertension, diabetes mellitus, atrial fibrillation) and clinical information (NIHSS at admittance, type of stroke, arterial territory affected by stroke). Our models have statistical significance and are easy to use, therefore may aid clinicians in their daily decisions, after further validation.