Preoperative Prediction of Postoperative Infections Using Machine Learning and Electronic Health Record Data
Authors' conclusion
Parsimonious preoperative models for predicting postoperative infection risk using EHR data were developed and showed comparable performance to existing American College of Surgeons National Surgical Quality Improvement Program risk models that use manual chart review.
Study context
- Population
- Surgical patients, 5 hospitals of the University of Colorado Health System, 2013–2019, electronic health record (EHR) data linked to ACS NSQIP data
- Participants
- 30,639 patients · 5 hospitals
- Outcome
- Surgical site infection, urinary tract infection, sepsis/septic shock and pneumonia
- Follow-up
- Up to 30 days postoperatively
Methodological detailsretrospective, multicentre, lasso and knockoff filter variable selection
- Retrospective
- Multicentre
- Lasso and knockoff filter variable selection
Performance
As reported in the abstract
Model discrimination
AUC range across models
0.73 to 0.89
Predictors selected for surgical site infection model (of >15,000 candidates)
7
Predictors selected for each of urinary tract infection, sepsis and pneumonia models
6
Original publication
Preoperative Prediction of Postoperative Infections Using Machine Learning and Electronic Health Record Data. Ann Surg. Published 27 September 2023. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →