Predicting all-cause 90-day hospital readmission for dental patients using machine learning methods

Area of Science:

• Health Informatics
• Machine Learning in Healthcare
• Predictive Analytics

Background:

• Hospital readmission rates are key indicators of healthcare quality.
• Machine learning (ML) offers potential for identifying high-risk patients for readmission.
• Few studies have utilized ML for predicting hospital readmissions.

Purpose of the Study:

• To assess ML as a tool for developing prediction models for all-cause 90-day hospital readmission in dental patients.
• To identify key predictors of hospital readmission for dental patients.
• To compare the performance of various ML algorithms for readmission prediction.

Main Methods:

• Utilized the 2013 Nationwide Readmissions Database (NRD) with 9260 dental patient admissions.
• Implemented five ML classification algorithms: decision tree, logistic regression, support vector machine, k-nearest neighbors, and artificial neural network (ANN).
• Evaluated model performance using AUC, accuracy, sensitivity, specificity, and precision.

Main Results:

• 18.9% of dental patients were readmitted within 90 days.
• Top predictors included total charges, number of diagnoses, age, chronic conditions, length of stay, procedures, payer, and illness severity.
• ANN models slightly outperformed other algorithms with an AUC of 0.743.

Conclusions:

• ML, particularly ANN, can effectively predict dental patient readmissions.
• Preventing all 90-day readmissions could save over $500 million annually across 21 states.
• Further research with ANN can refine risk factor identification and targeted interventions.