FAIRFAX, Va. (WDVM) — In December, a group of high school students was invited to the National Artificial Intelligence Institute’s “tech sprint”: part of the Department of Veterans Affairs’ push to make its applications data-driven. They were the only high school group in attendance.
“It was really special to be asked to join a tech sprint with companies like Oracle that are huge — and also the government — so that was definitely a once in a lifetime experience,” said participant Neeyanth Kopparapu, a senior at Thomas Jefferson High School.
Since June, Kopparapu and friends Ethan Ocasio and Shreeja Kikkisetti have been developing a their Clinical Trial Selector, using artificial intelligence, for veterans and other users on Medicare.
Director of the Department of Veterans Affairs’ National Artificial Intelligence Institute Gil Alterovitz invited the group to the tech sprint. “There are many unique circumstances that we see with veterans, which include different exposures, higher cancer rates, and other conditions, that lead to a point where potentially a more personalized medicine may be useful,” said Alterovitz. “Veterans may have different demographics and needs that [the group was] able to see and kind of get those personal stories.”
Kopparapu says speaking with veterans changed his approach from technology-centered to human-centered. Some of the veterans they spoke with had experience with using clinicaltrials.gov; a government website that connects a Medicare user to clinical trials. Once the list appears, the Medicare user has to contact each trial or do their own research to determine whether they’re eligible or not.
“For every single clinical trial there is something called an ‘eligibility criteria’ and it’s just this pretty lengthy description of exactly what a patient needs, the requirements that patient must have, to be able to be eligible for this clinical trial,” said Ocasio. The Clinical Trial Selector uses artificial intelligence to only match the patient to clinical trials they’re eligible for.
Using a sample user with cataracts, Kopparapu and Orcasio compared the clinicaltrials.gov results to the Clinical Trial Selector’s. The government website churned up 341 trials the patient might not even be eligible for. The Clinical Trial Selector? Just two in a matter of minutes.
Kopparapu and Ocasio hope to launch the app in the next couple of months. They’re working on the app’s security and making sure it’s HIPAA compliant. The patient’s personal information is never stored in the application; only linked to a government database.
In December, the group was recognized at the Creating the Future of Health competition by the Executive Office of the President.