Using Machine Learning to Predict Opioid Prescriptions After Spine Surgery
Author Information
Author(s): Bouterse Alexander, Cabrera Andrew, Jameel Adam, Chung David, Danisa Olumide
Primary Institution: Loma Linda University
Hypothesis
Can machine learning algorithms identify clinical variables that predict increased opioid utilization after spine surgeries?
Conclusion
The study identified key risk factors for increased postoperative opioid use, including shorter hospital stays, younger age, and longer operative times.
Supporting Evidence
- 59% of patients were prescribed additional opioids at their first outpatient appointment.
- Hospital length of stay was a significant predictor of opioid prescriptions.
- Machine learning algorithms achieved an average AUC of 0.743 in predicting opioid prescriptions.
Takeaway
Doctors used computer programs to find out which patients might need more pain medicine after spine surgery, helping to prevent too much medicine from being given.
Methodology
The study used six machine learning algorithms to analyze data from patients who underwent three types of spine surgery between 2013 and 2022.
Potential Biases
The use of a deidentified database may introduce confounding errors and bias in variable importance due to correlated features.
Limitations
The study is limited to a single patient population and lacks access to some relevant perioperative variables.
Participant Demographics
The study included adult patients undergoing ACDF, PTLF, or lumbar laminectomy, with a mix of genders and various comorbidities.
Statistical Information
P-Value
p<0.001
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website