한국인을 위한 뇌졸중 발생 예측모형 개발

Abstract

Background: Assessing an individuals risk of stroke can be a starting point for stroke prevention. The aim of this study was to develop a stroke prediction model that can be applied to the Korean population, using the best available current knowledge. Methods: A sex- and age-specific stroke prediction model that is applicable specifically to Koreans was developed using Gails breast cancer prediction model, which is based on competing risk theory. Results: The relative risks for major stroke risk factors, including hypertension, diabetes, hypercholesterolemia, atrial fibrillation, ischemic heart disease, previous stroke, obesity, and smoking status, were obtained from a recent systematic review of stroke risk factors among Koreans. The results were incorporated into the concept of a proportional hazard regression model. For baseline age- and sex-specific hazard rates for stroke, we employed Jees 10-year stroke-risk prediction model with its reference categories for predictor variables. Death-certificate data from the Korea National Statistical Office were used to calculate competing risks of stroke in our model. Conclusions: Our prediction model for stroke incidence may be useful for predicting an individuals risk of stroke based on his/her age, sex, and risk factors. This model will contribute to the development of individualized risk-specific guidelines for the prevention of stroke.