Hospital Readmission in Stroke Survivors in Social Vulnerability: Predictive Modeling with Machine Learning from the Perspective of the Chronic Conditions Care Model
This model demonstrates predictive modeling for hospital readmission in stroke survivors with social vulnerability, indicating significant complications associated with increased risk.
Key Points
The study aims to develop a predictive model for hospital readmission among socially vulnerable stroke survivors using machine learning techniques.
Applied machine learning algorithms including decision tree and logistic regression.
Data was split into training (70% and 80%) and testing (30% and 20%) sets.
Analyses conducted using Python, evaluated with ROC curves, AUC, and confusion matrix.
The decision tree model achieved an accuracy of 92.45% with an 80/20 data partition.
Falls, time since the first stroke, caregiver presence, and difficulty sleeping were key variables associated with readmission.
Falls increased readmission risk by 235%, ischemic stroke by 155%, and COVID-19 by 132%.