At a lab at Rutgers University, Dr. Zeeshan Ahmed works on a field of research, called precision medicine, that exudes all the usual promise of a really big idea. In this case, it bears health care-revolutionizing potential.
In more practical terms, the real-world impact of this area of study on the practice of medicine has been … imprecise, to put it kindly.
Although it faces some ongoing hurdles to implementation, Ahmed is confident that the insights of precision medicine — and the artificial intelligence and machine learning emphasis his team puts on it at his Rutgers lab — are increasingly useful to a pandemic-stressed health care system.
“The time has never been more critical for these innovative solutions,” he said. “AI and machine learning have emerged as key tools for the discovery of new diagnostics and drug combinations. These technologies are also very powerful clinical management tools for delivering efficient care to patient populations.”
AI and machine learning, however, have so far not been widely applied to the data side of the health care space — which happens to be the forte of the researchers at the Ahmed Lab at Rutgers Institute for Health, Health Care Policy and Aging Research.
Their field of research, precision medicine, involves an examination of factors, including genetics and the lifestyle of an individual, to help predict health issues through the discovery of disease biomarkers. The analysis of statistical patterns also could help guide medical practitioners in coming up with more personalized treatment plans.
“Precision medicine started getting attention when former President Barack Obama put a spotlight on it in 2015,” he said, referring to the former president’s 2015 State of the Union address and the launch of the nationwide Precision Medicine Initiative. “And, so far, it has seen good progress, but the actual implementation is not satisfactory.”
Ahmed came to health sciences from a background in computer science, taking into the sector with him a belief that machine learning and pattern-detecting AI methods could do a lot more for the study of biological data.
“And, when I jumped into this bioinformatics field in 2009, I was one of the few people who had published an article regarding use of AI in bioinformatics,” he said. “At that time, it was not as widely accepted in the community. But, in the last few years, AI and machine learning has become popular in this field, which is a great sign.”
Last year, Rutgers’ Ahmed Lab had a research article published in Oxford University Press that Ahmed said remained one of the publication’s most-read articles for a long time.
Still, in his 10 years of experience in this field, he said change, at least in the early stage of a development, “is usually not welcome.” New approaches take a long time to find acceptance in the health sciences, he added.
One of the ways that reluctance makes itself known is the lack of trust health care organizations have with sharing electronic medical record data with researchers.
“Availability of data is currently one of our biggest hurdles,” he said. “The right data is not yet available in a huge quantity.”
In order to do something like use an AI system to transform genetic or metabolic data into details that a patient can do something about, precision medicine researchers need the data to begin with. Even getting data that’s been stripped of identifying details and rendered anonymous has proved difficult for them.
Even with that as a barrier, the spread of COVID-19 presented Ahmed’s team the opportunity to demonstrate how useful data analysis might be in optimizing decision-making when machine learning algorithms are applied to health information.
“This was a very sad situation that (confronted) us,” he said. “But, during the pandemic, we started a new project with the goal of implementing AI and data analysis to support better diagnosis of patients.”
One of the major issues with the novel infectious disease on the diagnostic and treatment side was classifying and prioritizing COVID-19 positive patients into those that needed immediate care and those who could safely quarantine during their recovery.
The Rutgers researchers have been studying the clinical presentation of the disease and using high-tech data analysis methods to try to link it to predictive biomarkers, such as particular genes. They’re also looking at the long-term COVID-19 effects and how those biomarkers might be at play there, too.
It’s just one of the ways Ahmed hopes to illustrate the power of precision medicine.
“With collaboration with the right people, development of understanding and the increased acceptance of new change at all the labs working to help the field of health care with this approach … we believe we can achieve truly personalized treatment for earlier, more effective disease detection and prevention,” he said.
