While generative AI brings the promise of great advances in many sectors, health care may benefit more than most according to Ravi David Yarid, DO, assistant professor in the Doctor of Osteopathic Medicine (DO) program at 黑料传送门 Georgia.
Yarid, who currently leads the AI task force for the and is also spearheading efforts to incorporate AI into medical education at 黑料传送门, believes the technology can dramatically improve the practice of medicine.
鈥淭he job is going to change from reactively treating everything to proactive health,鈥� Yarid said. 鈥淲e are going to be the gatekeepers of amplifying health and improving the quality of life in our population as opposed to mitigating the reduction in quality of life as we age and in reacting to disease states with medications in treating symptoms.鈥�
One aspect of being a good healthcare provider is the ability to pick up on things the patient does not say, or is unable to tell you. Yarid envisions a future in which patient encounters are monitored by AI systems that will assist physicians in detecting meaningful nuances.
A patient, for instance, might have an inflection or shift in tone of voice when answering a question. The AI system can detect that subtle change and give the physician a virtual 鈥渘udge,鈥� Yarid explained.
鈥淚t will give you an opportunity to start to go down pathways that you weren't aware of,鈥� he said.
It will also serve as a learning tool, Yarid added.
鈥淵ou reflect upon that and say, 鈥業 did hear that. I didn't realize it was important.鈥� And then the next time you hear it, you don't need the tool to tell you,鈥� he said. 鈥淚t brings an awareness to yourself that you can grow from.鈥�
Another potential use of AI is looking at large amounts of patient data and identifying anomalies.
While a patient may have a lab result within a normal range, that result could still be abnormal for that patient. Yarid predicts AI will be used to assist providers in identifying health issues at earlier stages.
鈥淵ou can get ahead of a potential disease process rather than reacting to it when it's already become a significant issue in their lives,鈥� he said.
An example of this would be a patient whose blood pressure has typically measured around 105/68鈥攁 normal reading. Over time, the readings creep upward.
鈥淭hey're measuring at 125/84 and when you just look at it, you think that's normal but if you look at the trend you know something's happening,鈥� Yarid explained. 鈥淲e're gonna intervene at a much earlier place and we're going to start discovering that we can prevent most of our suffering in these illnesses through this intense awareness.鈥�
鈥淚f I had AI assistance, I could probably see twice as many patients,鈥� Yarid said.
Currently, the system is structured in a way that someone with a chronic condition has to come in several times a year. In a world where AI is helping monitor the patient, the number of visits could be reduced, freeing up patient slots for others.
In the case of someone with diabetes, for example, AI would assist the physician in assessing how well the condition was being controlled.
鈥淲e're following your blood sugar鈥攎aking sure it's all good. I need to see you once a year again. And now I have three more appointments,鈥� he said.
If something changes with the patient鈥檚 condition, the physician is notified immediately and can schedule an appointment to see the patient.
With AI systems ultimately reducing the need for some positions and increasing efficiency in healthcare delivery, costs for providing health care should go down.
Yarid estimates these AI-facilitated cost reductions could occur within a relatively short period of time.
鈥淵ou're talking about huge reductions in the cost of health care, over 50%鈥� he said. 鈥淚t is possible within 10 years.鈥�
In the near future, according to Yarid, tests such as EKGs will no longer resemble their current form.
鈥淵ou're going to walk in and someone is either going to throw something around your chest or you鈥檒l put this vest on and it's going to measure your heart from 360 degrees and give you a three-dimensional view of your heart and all the electrical activity as opposed to interpreting these lines on paper,鈥� he predicted. 鈥淚t's gonna give you an overwhelmingly accurate view of the heart compared to what we currently have.鈥�
Instead of an Apple Watch or Fitbit or other wearable device, Yarid believes the health technology of the future will be more integrated.
鈥淚t'll probably be embedded,鈥� he said. 鈥淚t probably won't be wearable. Honestly, chances are soon enough, it'll be like a tattoo.鈥�
The device, he expects, will monitor your blood sugar, pulse and other vital signs.
鈥淚t'll be constantly monitored whether it's on you or just under your skin,鈥� he said.
Not only will physicians have access to the collected data, but the patients will be able to view the information as well.
鈥淚t'll be glasses for a minute, but then it'll go straight to something even more profound,鈥� he said. His best guess as to what shape that might take鈥攍ens replacement.
鈥淥ur lens replacement is going to have a built-in interface that we're gonna be able to look at the world and it's going to tell us all kinds of things about ourselves and what we need,鈥� he projects.
According to Yarid, electronic medical record (EMR) systems have failed miserably in living up to their promise of making entries from disparate systems easily accessible to patients and providers.
Instead, he said, you have resources allocated to glorified data entry and patients end up with hundreds of pages of essentially useless information.
With AI, years of physician notes and laboratory results could be summarized and provide physicians with key information in one place and allow them to interact with that data.
This could help improve patient safety and outcomes by ensuring important information is not overlooked.
鈥淚t's a safety net and enhances our ability to be efficient, effective with our time鈥攁nd deal with the onslaught of information we have to address every day with expected perfection,鈥� Yarid said.
What if AI could be used to direct robotic surgery? This is another potential use Yarid sees for generative AI.
鈥淭here will still be surgeries, but it'll be very different,鈥� he said. 鈥淚 can see having a master surgeon in a room. Anywhere in the world with a whole bunch of screens up.鈥�
If the AI surgeon detects an issue, then an alert can be triggered notifying the master surgeon to intervene.
鈥淵ou need some human input, and it'll learn from that and then at some point you won't even need that,鈥� Yarid anticipates.
Today鈥檚 medications come in a set range of dosages. Patients are prescribed a dosage that best matches their needs within the limitations of the available dosage amounts. Yarid foresees a future in which medications become person-specific.
鈥淭heoretically, AI applications can be utilized across fields. You'll basically have medications printed for you,鈥� he said. 鈥淪o you'll have a medication of 37.2 milligrams and that's your specific medication dosage tailored to you and that's what you are going to get from the pharmacy.鈥�
Yarid also expects AI to make medication management easier.
鈥淪ay you take eight medications鈥擜I will actually turn it into just one pill a day,鈥� he said.
Breakthroughs such as this will come from AI鈥檚 ability to process large numbers of variables allowing highly customized treatments.
The ability to relate variables to each other is something humans can do on a limited basis. For AI, the capacity is essentially unlimited.
鈥淲e're going to discover associations we never dreamed of, never thought of, and it's going to happen really fast,鈥� Yarid said. 鈥淚t's already happening.鈥�
As for how long it might take before we see these types of changes in clinical practice, Yarid admits it is hard to predict.
鈥淭here are so many wild cards that if any one of them shifts in a direction, that changes the entire course,鈥� he said.
Yarid believes having the ability to process and analyze such large amounts of data will open the floodgates.
鈥淚 do know that within a decade society鈥攁nd our living鈥攚ill be radically different than anything we're looking at today,鈥� he said.