Prediction model of hospitalization period of patients with chronic kidney disease using hospitalization data

Abstract

We tried to establish a predictive hospitalization model in patients with chronic kidney disease.
We conducted an ordered logistic regression analysis with hospitalization period classification as an explanatory variable and patient contents as dependent variable. In the constructed probability calculation model, we identified combination among dietary intake, age, I63・I69：cerebral infarction (including sequelae), N08：diabetic nephropathy，and I50：heart failure.
The hospitalization period can be predicted by inputting the evaluation value for five items, which are patient conditions at the time of admission. Predicting the hospitalization period at the initial stage may provide valuable information for the subsequent need and organize collaboration, and suggests the possibility of a more effective bed management.