Development of a Machine Learning-Based Patient Classification System for a Gastroenterology Ward
Hong-Fei Ren1, Ming-Fang Wei1, Ming Shen1
1Department of Gastroenterology, West China Hospital, West China School of Nursing, Sichuan University, Chengdu, Sichuan, China, scu.edu.cn.
Journal of Nursing Management
|July 30, 2026
Summary
A new patient classification system (PCS) accurately predicts nursing care needs for gastroenterology wards. This validated tool supports evidence-based nurse staffing and resource allocation.
Area of Science:
- Nursing Informatics
- Healthcare Management
- Machine Learning in Healthcare
Background:
- Existing patient classification systems (PCSs) lack validation for gastroenterology wards.
- Accurate nursing workload estimation is crucial for budgeting and staffing.
Purpose of the Study:
- To develop and validate a novel PCS specifically for gastroenterology wards.
- To quantify nursing workload and inform evidence-based staffing decisions.
Main Methods:
- Utilized Orem's self-care and Henderson's human needs theories.
- Applied machine learning (decision tree model) to HIS data (2019-2020).
- Validated prospectively with 357 patients (2022-2023).
Main Results:
- Developed a 2-category, 5-subcategory PCS based on patient characteristics and 24-hour nursing time.
- Category 1 (surgery days) ranged from 1.66 to 4.15 nursing hours.
- Category 2 (other days) ranged from 0.96 to 3.42 nursing hours.
- Demonstrated good model fit and predictive performance via internal and external validation.
Conclusions:
- The PCS enables rapid patient classification and accurate prediction of nursing care requirements.
- The system shows strong internal consistency, stability, and generalizability.
- Provides a scientific basis for data-driven nurse staffing and workforce allocation policies.