Related Experiment Video
Updated: Jun 11, 2025
12:18
A Machine Learning Approach to Design an Efficient Selective Screening of Mild Cognitive Impairment
Published on: January 11, 2020
7.5K
Development of predictive model for post-stroke depression at discharge based on decision tree algorithm: A
Guo Li1, Jinfeng Miao1, Ping Jing2
1Department of Neurology, Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology, 1095 Jiefang Avenue, Wuhan, Hubei 430030, China.
Journal of Psychosomatic Research
|September 28, 2024
Summary
Post-stroke depression (PSD) is common. This study identified key predictors like coronary heart disease history and hospital stay length to build decision tree models for predicting PSD at discharge.
Area of Science:
- Neurology
- Psychiatry
- Data Science
Background:
- Post-stroke depression (PSD) is a frequent and severe complication following stroke.
- Predicting PSD is crucial for patient management and recovery.
- Utilizing multi-predictor models offers a more comprehensive approach than single-predictor analyses.
Purpose of the Study:
- To develop practical prediction tools for post-stroke depression (PSD) at hospital discharge.
- To employ a decision tree (DT) algorithm for constructing these predictive models.
- To identify significant predictors for PSD in stroke survivors.
Main Methods:
- A multi-center prospective cohort study involving 876 stroke patients within seven days of onset.
- Independent predictors of PSD were identified using multivariate logistic regression.
- Classification and Regression Tree (CART) algorithm was used for decision tree model development.
Main Results:
- Key predictors for PSD at discharge included history of coronary heart disease, length of hospital stay, NIHSS score, and Mini-Mental State Examination (MMSE) score.
- Subgroup analysis revealed associations between hemorrhagic stroke, hypertension history, and higher modified Rankin Scale (mRS) scores with PSD in young adults.
- The study identified significant predictors for PSD, enabling the construction of predictive models.
Conclusions:
- Predictive models for post-stroke depression (PSD) at discharge were successfully constructed using decision tree algorithms.
- Identified predictors facilitate clinical decision-making for managing PSD.
- The developed tools aid in the early identification and management of PSD in stroke patients.
