Related Experiment Video
Updated: May 12, 2026
Implementation of a Real-Time Psychosis Risk Detection and Alerting System Based on Electronic Health Records using CogStack
Published on: May 15, 2020
Design and predictive performance analysis of an early readmission risk index in a mental health hospitalization
Yaiza García-Illanes1,2,3, Vera Carbonell-Aranda1,2,3, María Traverso-Rodriguez1,2,3
1IBIMA Plataforma BIONAND, Instituto de Investigación Biomédica de Málaga, Málaga, Spain.
Abstract:
Early hospital readmission in mental health inpatients negatively affects patient outcomes and increases healthcare costs, making early risk identification crucial. This study aimed to develop a 30-day readmission risk index using a retrospective case-control design in the Mental Health Hospitalization Unit of the Hospital Regional Universitario de Málaga, including 196 admission episodes (98 early readmissions and 98 controls). Univariate and multivariate logistic regression analyses identified the following predictive variables: previous hospital admissions, presence of a personality disorder, high social risk, having used emergency services in the past year, having a legal capacity restriction, aggressive behavior at admission, and marital status. The resulting index demonstrated good predictive performance (AUC = 0.764), with internal validation showing an optimism-corrected AUC of 0.727 and acceptable calibration (slope = 0.796, intercept = -0.023). This tool may be useful for risk stratification and post-discharge planning, but should be used cautiously due to the exploratory nature of the study.
Clinical Trial Registration:
ClinicalTrials.gov, identifier NCT06604780.
