What the study found
The study found that anomaly-based alerting for unusual patient-management actions can have reasonably low false alert rates, and that stronger anomalies are linked to higher alert rates.
Why the authors say this matters
The authors suggest that patient-management actions that are unusual compared with past cases may reflect a potential error, so it may be worthwhile to raise an alert when such a condition is encountered.
What the researchers tested
The researchers developed and evaluated a data-driven approach for detecting unusual patient-management actions using past patient cases stored in an electronic health record (EHR) system. They tested the approach on electronic health records from 4,486 post-cardiac surgical patients and based the evaluation on a panel of experts' opinions.
What worked and what didn't
The results supported the idea that anomaly-based alerting can have reasonably low false alert rates. The findings also indicated that stronger anomalies were correlated with higher alert rates.
What to keep in mind
The available abstract does not describe additional limitations, and the evaluation was based on expert opinion in a specific group of 4,486 post-cardiac surgical patients.
Key points
- The study evaluated anomaly-based alerting for unusual patient-management actions in an electronic health record system.
- The authors suggest unusual actions may indicate a potential error worth flagging.
- In 4,486 post-cardiac surgical patients, the approach showed reasonably low false alert rates.
- Stronger anomalies were associated with higher alert rates.
- The evaluation was based on a panel of experts' opinions.
Disclosure
- Research title:
- Anomaly-based alerting showed low false alert rates
- Image credit:
- Photo by Charlss GonzHu on Pexels
Get the weekly research newsletter
Stay current with peer-reviewed research without reading academic papers — one filtered digest, every Friday.