Dive overview:
- Researchers and doctors at the University of Pittsburgh have developed an automated machine-learning model that can identify patients at high risk for complications after surgery, according to the study. published Friday Open JAMA network.
- This model outperformed common tools used to calculate surgical risks. American College of Surgeons. While established calculators require clinicians to manually enter data, this model automatically reads records of people scheduled for surgery and flags high-risk patients.
- Clinicians at the University of Pittsburgh Medical Center (UPMC) 20 hospitals are using this model to identify high-risk patients and inform coordination of care and rehabilitation activities.
Dive Insight:
Complications within 30 days after surgery are the leading cause of death worldwide. Before the pandemic, postoperative complications were on the top list of causes of death, followed by heart disease and stroke. This situation creates a need for a way to identify which patients awaiting surgery are at risk of being one of her 4.2 million worldwide each year who die within 30 days of surgery. is occurring.
Many US hospitals participate in the American College of Surgeons’ National Surgical Quality Improvement Program (ACS). NSQIP) To identify high-risk patients. However, the calculator can only make predictions if the clinician manually provides all the requested information.
“Identifying high-risk patients can be challenging for busy clinicians who need to integrate the wealth of available health data and frequently perform additional tests and clinical evaluations. We wanted to build an easy-to-use model that would use existing data to quickly provide healthcare teams with automated and accurate risk assessments.” Director of Perioperative and Surgical Services at UPMC, Man Mahajan said: statement.
Researchers at the University of Pittsburgh and UPMC recognized an opportunity to automate the process and used machine learning to train a model to identify high-risk patients. The team tested the algorithm on the medical records of 1.25 million surgical patients, and using it clinically he assessed the risk of his 206,000 patients at UPMC.
On a scale of 0 to 1, higher numbers indicate better predictions, with the machine learning model scoring 0.945, while the ACS NSQIP scored 0.897. Pitt and his UPMC collaborators want to train the model to predict the likelihood of sepsis, respiratory disease and other complications that often hospitalize patients after surgery.
