Development of Stroke Diagnosis Algorithm Through Logistic Regression Analysis with National Health Insurance Database

Abstract

This study purpose is to derive a model equation for developing a stroke pre-diagnosis algorithm with the potentially modifiable risk factors. In this study, logistic regression analysis technique was used for model derivation. It is one of the methods employed in the machine-learning field of statistics. Korea’s National Health Insurance Service (NHIS), one of the largest administrative health care databases around the world, has been used widely in academic studies. From the NHIS Corporation, 500,000 enrollees’ databases were collected. For the regression analysis, 367 stroke patients’ data were selected from the NHIS database. The control group consisted of 500 patients who were followed up for two consecutive years and who had no history of stroke. As a result, the separation accuracy with the modifiable risk factors was 64.7%. The results of this study are expected to be useful for the development of stroke pre-diagnosis algorithms.