����ɫ��

Patient safety is a critical component of a quality health-care system and is often measured by the occurrence of adverse events such as hospital falls, IV-site infections, and accidental punctures during surgery. Capturing these adverse events is crucial to understanding the safety of health care and ultimately improving it.

A World Health Organization advisory group identified 153 adverse events that may be used as potential patient safety indicators (PSIs). But measuring these PSIs isn’t easy. All it takes is a misspelled word or a wrong code for the PSIs to be missed when analyzing the data. That’s why PSIs are largely under-reported in hospital data and health surveillance systems.

Goal: Improve collection of accurate data

But an international team based at the Cumming School of Medicine (CSM), led by Dr. Yuan Xu, MD, PhD, is tackling the problem head-on. 

“Our goal is to better count adverse events in health data so we can potentially identify them earlier, even while individuals are still in the hospital,” he says. “Our goal is to improve patient safety and experience.”

The team recently received afor a four-year project that will pair computer algorithms and artificial intelligence (AI) techniques with patient data from electronic medical records (EMR) in Alberta, to capture the true numbers. 

The team’s first step is developing the EMR-based algorithms. Step two is using AI to automate the detection of the adverse events. Finally, the team will assess their work using clinical expert reviews conducted on a dataset of 10,000 EMR charts from Calgary to ensure their findings translate to the clinic.

Best possible experience for patients

Following completion, the project should allow improved monitoring of adverse events and patient safety surveillance, giving decision-makers the information they need to improve health-care systems.

Dr. Cathy Eastwood, PhD, a co-principal investigator of the study and a former front-line nurse, is glad to be part of the project. 

“Medical professionals all seek to help patients, but sometimes harm can occur,” she says. “This project will allow us to accurately identify and count the number of adverse events in electronic medical record data for the safest care possible for patients.”