Stanford develops data-mining algorithm to identify safety of implantable medical devices

Stanford (Calif.) Medicine researchers created an algorithm that uses artificial intelligence to sift through EHR data for implantable medical device surveillance, according to the medical school's blog post.

While there are methods to report medical device safety issues to the FDA, these measures can sometimes introduce bias into data analysis and even go unreported due to drawn out processes, said Alison Callahan, PhD, a Stanford research scientist and co-lead author on the research published in npj Digital Medicine.

The widespread amount of patient information across various databases is also an issue for detecting a device's safety profile or success rates. To combat this, researchers developed an AI-based monitoring algorithm that can access de-identified medical records from former Stanford patients and link up medical characteristics, such as infection rates, how long an implant lasts before needing replacement and patient pain levels, with specific implantable devices.  

To test the algorithm, the researchers examined EHR data of former hip replacement patients. When applied, the algorithm accurately identified complication events with each patient's implant and also recognized pain as a strong predictor of complication.

"Previous studies have likewise shown that pain is a useful predictor for more serious complications later on," Dr. Callahan said. Analyzing patient pain levels over a period of time helped the algorithm to identify a patients' quality of life with that specific implant.

Researchers concluded that while the algorithm can denote the "safest" medical devices, the team hopes to further develop it to provide more information like which devices are most effective based on patient population.

More articles on EHRs:
HITAC releases draft of interoperability standards report
Allscripts Developer Program adds interoperability startup founded by former Epic engineers
Keck Medicine of USC implements Cerner PDMP access for EHR

© Copyright ASC COMMUNICATIONS 2020. Interested in LINKING to or REPRINTING this content? View our policies by clicking here.

 

Featured Webinars

Featured Whitepapers