Listen to this article 2 min
Vanderbilt University Medical Center is looking into how to use artificial intelligence to manage computerized alerts for medical professionals.
And the results are promising, according to a study conducted by a Vanderbilt research team.
The study’s lead author Siru Liu, assistant professor of biomedical informatics at Vanderbilt, said in a press release that many alerts — which can include warnings about drug contraindications to gaps in patient care documentation — are missed by medical professionals due to the hectic nature of their jobs.
"Across health care, most of these well-intentioned automated alerts are overridden by busy users,” Liu said. “The alerts serve an essential purpose, but the need to improve them is clear to everyone.”
Vanderbilt University Medical Center is Nashville’s largest hospital with $12.82 billion in gross patient revenue and 15,369 total employees in 2022, according to Business Journal research. This includes Monroe Carell Jr. Children’s Hospital at Vanderbilt.
Gross patient revenue
| Rank | Prior Rank | Name/Prior rank (*not ranked)/URL |
|---|---|---|
1 | 1 | Vanderbilt University Medical Center and Monroe Carell Jr. Children's Hospital at Vanderbilt |
2 | 2 | TriStar Centennial Medical Center |
3 | 3 | TriStar Skyline Medical Center |
The problem is the lack of specificity, resulting in up to 90% of those alerts being ignored, according to the release. This contributes to what is known as “alert fatigue” and throwing human experts at the problem is slow, expensive and somewhat hit and miss.
Which is why they turned to AI.
The use of artificial intelligence has boomed across the health care and business world over the past year. As AI evolves, more business leaders across Nashville are embracing the technology and figuring out how it can improve efficiency. The health care space specifically has been more cautious with the implementation of AI and has focused on its advancements on the administrative side.
As part of the study, Liu developed a machine learning approach to analyze two years of data to determine how often medical professionals interacted with the alerts they received at Vanderbilt University Medical Center. The team learned when specific alerts would be dismissed by users based on patient characteristics.
She then used the AI she’d built to understand why it made the conclusions it did and generate suggested improvements to the hospital’s alert system. Basically, it told Liu why users would ignore certain alerts.
For example, "if the patient is a hospice patient, then the user is less likely to accept the breast cancer screening alert."
"The alignment of the model's suggestions with manual adjustments made by clinicians to alert logic underscores the robust potential of this technology to enhance health care quality and efficiency," Liu said. "Our approach can identify areas overlooked in manual reviews and transform alert improvement into a continuous learning process."
The team found about 9.3% of the nearly 3 million alerts they analyzed could have been eliminated based on the patient’s circumstances.
And in the process, Liu said they found areas where there were issues with workflow, education and staffing.
“The transparency of our model unveiled scenarios where alerts are dismissed due to downstream issues beyond the alerts themselves,” Liu said.
Liu and her colleagues now have a list of ways they can use AI for studies and changes in processes at Vanderbilt.
Liu, senior author Adam Wright, professor of biomedical informatics and director of the Vanderbilt Clinical Informatics Center, and a research team reported the study in the Journal of the American Medical Informatics Association.
