Machine learning analysis of lab tests to predict bariatric readmissions

Area of Science:

• Medical Informatics
• Surgical Outcomes Research
• Machine Learning in Healthcare

Background:

• Bariatric surgery is a common procedure for obesity treatment.
• Identifying patients at high risk for readmission is crucial for improving patient care and reducing healthcare costs.
• Predictive models can aid in proactive interventions.

Purpose of the Study:

• To develop and evaluate machine learning models for predicting 30-day readmission after bariatric surgery.
• To identify key laboratory test indicators associated with readmission risk.
• To compare the performance of various machine learning algorithms in this prediction task.

Main Methods:

• Utilized data from 1262 patients undergoing bariatric surgery (2018-2023).
• Analyzed preoperative, postoperative day 1, and postoperative day 3 laboratory test indicators.
• Employed Least Absolute Shrinkage and Selection Operator (LASSO) regression for feature selection.
• Constructed and compared Support Vector Machine (SVM), generalized linear model, multi-layer perceptron, random forest, and extreme gradient boosting models.
• Evaluated model performance using the area under the receiver operating characteristic curve (AUROC).

Main Results:

• A total of 7.69% of patients were readmitted within 30 days.
• The Support Vector Machine (SVM) model achieved the highest predictive performance with an AUROC of 0.784 (95% CI 0.696-0.872).
• The SVM model outperformed other evaluated machine learning algorithms.

Conclusions:

• Machine learning models utilizing laboratory test data can effectively predict 30-day readmission risk post-bariatric surgery.
• The SVM model demonstrates significant potential for identifying high-risk patients.
• These findings support the integration of predictive analytics into postoperative care pathways.