Powering the Future of Precision Medicine
By Laura Freeman
The more you know, the more you realize there is to learn.
Research is expanding the boundaries of medical knowledge so rapidly that no one human brain can learn, retain and recall it all. So how do we sort through this sea of data to access the best information? How can we help patients benefit from new knowledge that could offer a better
Fortunately, as medical science has advanced, so has the science of managing information. Informatics, machine learning and artificial intelligence have made tremendous progress in analyzing data. By focusing on the tasks that machines do best, they free humans for the creative thinking people do best.
"Informatics is the enabling technology in precision medicine," Matt Might, PhD, said. Before becoming Director of the Hugh Kaul Precision Medicine Institute at UAB, Might spent three years at the White House as a strategist for the Precision Medicine Initiative. He was also a faculty member in the Department of Biomedical Informatics at Harvard Medical School.
Might's background in computing took a personal turn toward bioinformatics when his son was born with a rare enzyme deficiency that had never been reported before. Using his computing skills, he searched for a diagnosis and anything in the literature that might suggest an approach to treatment. His search found other patients with the same deficiency and an over the counter medication--Pepcid--that helped to ease some of the symptoms.
"Precision medicine is about optimizing health with data, so it's really about fusing high performance computing with big data," Might said. "At the Precision Medicine Institute at UAB, we've developed in-house artificial intelligence tools that use bioinformatics databases to predict potential treatment strategies tailored to the individual characteristics of patients.
"As medicine becomes increasingly personalized, we'll see a surge in the need for informatics tools that can help provide this mapping from patients to best possible treatments."
Instead of spending days or weeks searching through medical texts and research to find the one golden needle in a vast haystack, a well written informatics program can search through enormous databases in seconds and return answers ranked by the most likely fit to the question.
"For the most part, academic medical centers are taking the lead in using informatics in precision medicine, although in precision oncology, it's becoming more common to acquire genomic sequencing," Might said.
Now that genomic sequencing has become more affordable, being able to search the patient's genes for clues offers the ultimate database for personalized precision medicine. Oncology is taking that a step farther by sequencing the genome of the tumor itself. That has resulted in some surprising answers. One UAB patient with prostate cancer resistant to treatment had a tumor that genetically had much more in common with ovarian cancer. Treating him with an ovarian cancer drug dramatically improved his response.
The information that bioinformatics searches goes well beyond genomics. It can be programmed to look at vital signs, labs and medical history notes to seek out factors that could help with diagnosis or predict where a patient's health is heading. It can also look through the back shelves of medical information to seek out remedies that might offer another option for treatment when nothing else has worked.
One patient with persistent vomiting to the point of being near death in spite of every known treatment was saved when a search of old remedies found a suggestion for nasal alcohol. It worked. She went from less than 80 pounds to more than 120.
Moving to electronic medical records came with a learning curve. What needs to happen for bioinformatics to move from academic medical centers to become a tool physicians can use in a medical office?
Most primary care providers are continually pressed for time. After dealing with acute problems, they have little time to address chronic conditions and review trends that might have an effect on the patient's long term health. In the right form, bioinformatics could help to change that.
"Perhaps the biggest hurdle is user interface and user experience," Might said. Physicians have to be able to use and trust these tools. Tools that can't explain their recommendations in terms that physicians understand will never be used."
A great deal of current work is going into designing informatics program to ask the right questions, incorporating data from current electronic records and making room for new categories of data that may be added later.
Interface designers are also making progress toward paring electronic records technology with bioinformatic dashboards that can take on some of the health monitoring tasks to show at a glance where patients are doing well and what areas need help to optimize their health. They can also monitor remote blood pressure, blood sugar and other signs to alert physicians that patients are headed for trouble between visits.
As we move toward precision medicine, bioinformatics will empower physicians to go beyond treating the average patient and then fine tuning. It will enable us to treat the person in front of us--the right medicine for the right person at the right time.