Machine Learning Models to Predict Readmission Risk of Patients with Schizophrenia in a Spanish Region

  Currently, high hospital readmission rates have become a problem for mental health services, because it is directly associated with the quality of patient care. The development of predictive models with machine learning algorithms allows the assessment of readmission risk in hospitals. The main ob...

Full description

Saved in:
Bibliographic Details
Published in:International journal of mental health and addiction 2024-08, Vol.22 (4), p.2508-2527
Main Authors: Góngora Alonso, Susel, Herrera Montano, Isabel, Ayala, Juan Luis Martín, Rodrigues, Joel J. P. C., Franco-Martín, Manuel, de la Torre Díez, Isabel
Format: Article
Language:English
Subjects:
Algorithms
Codes
Community and Environmental Psychology
Datasets
Disease prevention
Drug abuse
Health Psychology
Hospitals
Length of stay
Machine learning
Medicine
Medicine & Public Health
Mental disorders
Mental health
Original Article
Patients
Psychiatry
Psychology
Public Health
Regression analysis
Rehabilitation
Risk factors
Schizophrenia
Support vector machines
Tumors
Variables
Citations: Items that this one cites
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:  Currently, high hospital readmission rates have become a problem for mental health services, because it is directly associated with the quality of patient care. The development of predictive models with machine learning algorithms allows the assessment of readmission risk in hospitals. The main objective of this paper is to predict the readmission risk of patients with schizophrenia in a region of Spain, using machine learning algorithms. In this study, we used a dataset with 6089 electronic admission records corresponding to 3065 patients with schizophrenia disorders. Data were collected in the period 2005–2015 from acute units of 11 public hospitals in a Spain region. The Random Forest classifier obtained the best results in predicting the readmission risk, in the metrics accuracy = 0.817, recall = 0.887, F1-score = 0.877, and AUC = 0.879. This paper shows the algorithm with highest accuracy value and determines the factors associated with readmission risk of patients with schizophrenia in this population. It also shows that the development of predictive models with a machine learning approach can help improve patient care quality and develop preventive treatments.
ISSN:1557-1874
1557-1882
DOI:10.1007/s11469-022-01001-x