Analyzing AI for Healthcare

Analyzing data from doctors’ notes may seem unrelated, but the Department of Energy’s (DOE) Oak Ridge National Laboratory (ORNL), is using Artificial Intelligence (AI) tools known as machine learning algorithms and deep learning algorithms to analyze information from doctors.

Researchers at ORNL are now able to analyze patient data from medical tests, doctors’ notes, and other health records. These techniques use language processing to identify patterns among notes from different doctors and are able to extract previously inaccessible insights from mountains of data.

When combined with results from x-rays and other relevant tests, these results may be able to improve healthcare providers’ ability to diagnose and treat problems ranging from post- traumatic stress disorder to cancer.

For example, Gina Tourassi Director for the ORNL Health Data Sciences Institute, uses AI to automatically compile and analyze data to determine which factors are responsible for developing certain diseases. Her team is running machine learning algorithms and scans millions of medical documents to gain better insight on certain diseases.

In a collaborative project with the VA, the team at DOE’s ORNL Administration is working hard to prevent suicide as more than 6,000 veterans died by suicide in 2016. The VA has started using predictive models and advanced informatics to identify at-risk veterans.

One model being used is called the medication possession ratio algorithm which creates individualized summaries of veterans’ medication patterns, to see which medications are prescribed for a veteran, and how often the prescriptions are filled. This model helps clinicians pinpoint veterans with inconsistent medication usage patterns as these veterans are known to have a higher risk of attempting suicide.

Until now, the medication possession ratio calculations have been limited in scope. This model has typically included only active psychotropic medications such as narcotics or mental health medications that only covered a narrow class of the total veteran population in the Veterans Health Administration’s (VHA) database.

The ORNL team was able to speed up the algorithm to cover all current medications, recent past prescriptions, and data on all nine million veterans in the database. Without the speedup, the expanded version of the model would have taken 75 hours to run. With the speedup, it runs in only 15 minutes.

As part of the program, the VA implemented a Crisis Line connecting veterans who call in to a hotline. To provide brief and targeted overviews of veterans’ medication histories, the VHA and the Crisis Line have worked together to establish the correct medication possession ratio algorithm for use in the Crisis Center.

According to Jodie Trafton, Director of the VA Program Evaluation and Resource Center in the VHA Office of Mental Health and Suicide Prevention, “We worked with the crisis line staff to build a report that summarizes all the key elements the staff would look for in a medical record so the patient’s information is provided quickly in a simplified format.”

Share Button