3056. A machine learning model to predict the risk factors causing feelings of burnout and emotional exhaustion amongst nursing staff in South Africa
Invited abstract in session TB-13: AI and Machine learning in healthcare, stream OR in Healthcare (ORAHS).
Tuesday, 10:30-12:00Room: Clarendon SR 1.01
Authors (first author is the speaker)
| 1. | Derya Demirtas
|
| Industrial Engineering & Business Information Systems, University of Twente | |
| 2. | Maria Van Zyl
|
| School of Industrial Engineering, University of Twente and North-West University | |
| 3. | Jacoba Bührmann
|
| School of Industrial Engineering, North-West University | |
| 4. | Alwiena Blignaut
|
| North-West University | |
| 5. | Siedine Coetzee
|
| North-West University |
Abstract
Nurses in South Africa face high demands and limited resources, leading to burnout and emotional exhaustion. Identifying key predictive factors is crucial for support and policy decisions. This study explores whether demographic data alone can predict burnout and uses machine learning to analyze contributing factors. Using PyCaret 3.3, supervised machine learning models were developed on 1165 survey responses from nurses in medical-surgical units. Models were evaluated by accuracy, AUC score, and confusion matrix. Predictions based on demographic data alone were compared to full survey data, and key predictive factors were extracted for analysis. The gradient booster classifier (GBC) achieved the highest accuracy: 75.8% for burnout and 76.8% for emotional exhaustion using full survey data, while demographic data alone yielded 60.4% and 68.5%, respectively. Fatigue was the strongest predictor of both outcomes. Confidence in management was the second highest predictor of burnout, while management’s willingness to listen was the second highest for emotional exhaustion. We conclude that machine learning effectively predicts burnout and emotional exhaustion from full survey data but not from demographics alone. Addressing fatigue, confidence in management, and management’s responsiveness may help mitigate these issues among South African nurses.
Keywords
- Machine Learning
- Health Care
Status: accepted
Back to the list of papers