Development and validation of a prediction model for extended hospitalisation in patient with ischemic stroke
10.3969/j.issn.1671-8283.2024.10.002
- VernacularTitle:缺血性脑卒中患者住院时间延长预测模型的构建与验证
- Author:
Yang YU
1
;
Shengqiang ZOU
;
Huihua BO
;
Xiaojuan ZHU
;
Cong WANG
;
Yue CAO
Author Information
1. 江苏大学医学院护理系,江苏镇江,212013;常州市第一人民医院,江苏常州,213003
- Publication Type:Journal Article
- Keywords:
ischemic stroke;
length of hospitalisation;
prolong;
prediction model
- From:
Modern Clinical Nursing
2024;23(10):7-14
- CountryChina
- Language:Chinese
-
Abstract:
Objective To develop and validate a prediction model for extended hospitalisation in patients with ischemic stroke.Methods A total of 318 patients with ischemic stroke hospitalised between November 2021 and May 2023 in a Grade ⅢA hospital in Changzhou were selected as study objects with a convenience sampling method.The modelling group consisted 212 patients and the validation group included 106 patients.The patients in the modelling group were divided into a group of extended hospital stay and a group of normal hospital stay.Binary logistic regression analysis was conducted to develop the prediction model,and data from the 106 patients in the validation group were then incorporated into the developed prediction model.The prediction performance and goodness-of-fit of the model were accessed using the area under the curve(AUC)of the receiver operation characteristic(ROC)curve and the Hosmer-Lemeshow test.Results Multivariate logistic analysis showed that comorbid diabetes,number of complications during hospitalisation,a Braden score less than 18,and a white blood cell count greater than 3.5×109/L or more than 9.5×109/L were the risk factors of extended hospitalisation in patients with cerebral ischemic stroke.Based on the factors,a prediction model was developed with following formula:P=1/[1+exp(-Z)].Hosmer-Lemeshow test for the prediction model yielded χ 2=7.430,P=0.191.AUC of the prediction model was 0.818(95％CI:0.754-0.883,P<0.001)with Jordon index of 0.51 and the optimal cut-off value of 0.268,sensitivity of 78.9％and specificity of 72.3％The results of the validation of independent data model showed a sensitivity of 75.0％,specificity of 74.4％,and accuracy of 74.5％.Conclusion Comorbid diabetes,Braden score at admission,white blood cell count and the number of complications during hospitalisation are the significant factors that affect the length of hospital stay in the patients with ischemic stroke.The prediction model for extended hospitalisation of the patients with ischemic stroke exhibits good predictive value and can provide reference in clinical decision-making.
