Identifying key influencers of patient satisfaction using an explainable machine learning approach

Area of Science:

• Healthcare Quality Assessment
• Health Informatics
• Machine Learning in Medicine

Background:

• Patient satisfaction is a critical indicator of healthcare quality, impacting health outcomes and patient experiences.
• Identifying factors that influence patient satisfaction is essential for improving healthcare services.

Purpose of the Study:

• To identify key factors influencing patient satisfaction in healthcare facilities.
• To evaluate the predictive performance of machine learning models for patient satisfaction.
• To interpret the findings using SHAP value analysis.

Main Methods:

• A cross-sectional survey was conducted with 312 patients from two private hospitals in Rangpur, Bangladesh.
• Machine learning models (LightGBM, Random Forest, XGBoost, CatBoost) were employed to predict patient satisfaction.
• SHAP value analysis was used for model interpretability.

Main Results:

• The LightGBM classifier demonstrated superior performance with high accuracy (0.85), MCC (0.69), and ROC-AUC (0.83).
• Significant factors influencing satisfaction included treatment plan, age, appointment ease, waiting time, and medication details.
• Shorter waiting times, clear communication, and structured interactions correlated with higher satisfaction.

Conclusions:

• The LightGBM model effectively predicts patient satisfaction, offering valuable insights for healthcare providers.
• Healthcare facilities should focus on enhancing communication, minimizing wait times, and providing clear treatment plans.
• Further research can explore additional factors to refine predictive models for patient satisfaction.