Topic: The Advancement of AI in Healthcare for Veterans

Interestingly, some researchers from the Department of Veterans Affairs have found a way to harness artificial intelligence to identify and thus provide better care for veterans who are at a high risk of either being hospitalized or dying. This ground-breaking research has the potential to utilize AI to save the lives of retired service members.

The Department of Veterans Affairs and the Utilization of AI in Healthcare

A group of researchers from the Department of Veterans Affairs, including Dr. Ravi Parikh and Dr. Amol Navathe, set out with one goal in mind: to provide veteran patients with the best type of healthcare possible. To do this, veterans who were high-risk first needed to be identified.

Initially, the research team examined the records of veterans who had already been identified as high-risk, meaning they were more likely to be hospitalized or die. These veterans had already been identified via the Care Assessment Needs (CAN) calculator previously developed by the VA as a predictive healthcare tool.

The goal of the research was to improve upon the CAN calculator and cultivate a more advanced predictive healthcare system. This involved a closer look at the scoring that identified a veteran as high-risk, helping to pair them up with personalized healthcare that would meet their needs.

The study performed by the VA research team carefully examined 110,000 veterans that the CAN calculator identified as extremely high-risk (in the top 25%). Next, the researchers studied them and their electronic health records in order to identify common traits. Notably, the electronic health records stored even the tiniest of details, including blood pressure and cholesterol levels. It also took stock of many other factors, including demographics, psychological, behavioral, and social factors that could affect adequate access to healthcare.

Since these comprehensive records could take an overwhelming amount of time to go through, the research team utilized a specific type of artificial intelligence referred to as “machine learning” to go through the extensive records and condense them into a brief yet informative format.

This method of utilizing artificial intelligence helped easily identify 119 variables connected to the veterans that influenced their CAN scores. With these variables identified, the research team began to narrow it down further, identifying a total of 30 unique subgroups of high-risk veterans. Importantly, 21 of the 30 subgroups identified were traced back to medical conditions, whereas the other 9 traced back to socioeconomic and demographic data. In further refining the results from the CAN calculator, the research team from the VA intends to continue to utilize AI in identifying high-risk veterans to get them the specific care they need.

Key Takeaway

The Department of Veterans Affairs is refining predictive healthcare technology and artificial intelligence systems to help further identify veterans who belong to high-risk groups. The goal is to match veterans who are at high risk for either hospitalization or death with healthcare providers and therapies that can best suit their medical needs.

Topic Discussed: The Advancement of AI in Healthcare for Veterans