|
KMID : 0613620190390010390
|
|
Health Social Welfare Review
2019 Volume.39 No. 1 p.390 ~ p.427
|
|
|
A Study on the Development of Predictive Model for Severity-Adjusted Length of Stay in Nervous System Patients Using Machine Learning
|
|
Park Jong-Ho
Kang Sung-Hong
|
|
|
Abstract
|
|
|
|
This study aims to develop a severity-adjusted length of stay predictive model according to comorbidity index by using machine learning and propose a algorithm of severity-adjusted length of stay (LOS) predictive model. The dataset was taken from Korea Centers for Disease Control and Prevention database of the hospital discharge survey from 2006 to 2015 and the severity-adjusted length of stay predictive model was developed for the nervous system patients to need a urgent management for length of stay. when it comes to the severity-adjusted length of stay predictive model about nervous system discharging patients, three tools were used for the severity-adjustment of comorbidity: the CCI, the ECI, and the CCS. The models using Regression, Decision Tree, Random Forest, Support Vector Regression, Neural Network as a Machine learning analysis methods were developed and then evaluate. As a result, Severity-adjusted predictive model using CCS as the severity-adjustment of comorbidity and Neural Network method has the highest R-square and has the most excellent prediction capability. In conclusion, there is a need to develop a severity-adjusted predictive model using CCS as the severity-adjustment of comorbidity and make use of severity-adjusted predictive model to has high prediction capability by using various machine-learning analytics.
|
|
|
KEYWORD
|
|
|
Length of Stay, Diseases of the Nervous System, Comorbidity Index, Machine Learning, Severity-Adjusted Predictive Model
|
|
|
FullTexts / Linksout information
|
|
|
|
|
|
Listed journal information
|
|
|