This month’s blog is by Fast Layne Solutions, Inc. CEO Christopher Hughey.
When I say the terms ‘Artificial Intelligence’ (aka AI) and ‘Medicine,’ what images spring to mind for you? For a lot of people, especially nerds like me, combining those two terms conjures up images of android doctors or Emergency Medical Holograms of the kind we were introduced to in Star Trek. For other people, though, it conjures up images and feelings that are considerably darker and far less wondrous: healthcare professionals unfamiliar with this cutting-edge field of technology often fear that AI-driven medical diagnostics are nothing more than an attempt to replace them with advanced technology designed to supplant human healthcare workers.
Nothing could be further from the truth. First of all, the technology simply isn’t there yet, and we’re at least decades away from even the most rudimentary attempts to achieve such a goal, a goal most humans would likely reject anyway, given the very human touch required to establish that vital element of trust between healthcare workers and their patients. But even if humans one day decide they prefer the lifeless touch of an android or holographic doctor, the future physician whose job would be threatened by such a machine has not only not been born yet; it’s likely her grandmother hasn’t even been born yet.
So what are the goals of AI-driven medical diagnostics? In a nutshell: to clear the path between patients and healthcare workers in order to facilitate, not inhibit more human interactions when patients are in need.
Let’s take WellAI as an example. This international start-up is at the bleeding edge of AI-driven diagnostics. Consider a use case for their diagnostics tool.
First, let’s look at how things work without WellAI: A patient calls into a practice’s nurse triage line with a medical issue. Mary calls the practice’s switchboard with a question about swelling in her ankle. Staff is very busy dealing with patients in the office and all the busy work that so encumbers modern medical practices, such as chasing medication prior authorizations or doing eligibility checks (two things they have no business doing, by the way, but that’s another blog article). She finally gets through to a staff member, who takes down her information, looks her up in the EHR, then puts her on hold again while a nurse is found who can take the call. Finally, 15 minutes later, she is talking to a nurse, John. John verifies all her information, then listens to Mary’s symptoms. He determines she has at most a minor sprain and recommends RICE (rest, ice, compression, elevation). Total time spent getting to this point: 20 minutes. And while Mary is happy she doesn’t have to go to urgent care, she’s frustrated at the long wait. Staff, meanwhile, have had yet another thing added to their plate, adding to their own frustration and stress. And while of course they were all happy to help a loyal patient of the practice, note that the practice received no compensation for this call, adding further strain to their financial burden.
Let’s rewind and look at how things could work with an advanced machine learning algorithm (aka the AI) helping out all these humans. As a patient of your practice, Mary has the WellAI app on her smartphone because it was recommended at her last visit. She enters her symptoms into the app, and as she does so, she is prompted for more information. Within a few minutes, the AI has determined she has a simple, minor strain and lists the top three likely, unbiased diagnoses based on literally millions of medical studies the AI has read (leaving it to the doctor or nurse to make the final determination). Total time expended: less than 5 minutes. Total time taken up by the staff: close to zero. Threat to their jobs: none, since this does not reduce revenue for the practice, and in fact helps drive it, because Mary is now a more loyal patient than ever.
But let’s suppose that in that second scenario, Mary mentions a symptom that makes the AI suspect that there may be something more here. Let’s say that despite her mentioning a stumble involving only the left foot, she tells the AI that both her ankles are quite swollen and sore, and other information given suggests there may be something more serious involved. The AI doesn’t play physician: it flags her case and suggests she connect with the doctor’s office immediately, and she schedules either a visit or a telehealth session immediately. Further examination upon her visit reveals an underlying health condition that only a trained doctor can diagnose. And that doctor and her nurse had more time to dedicate to helping Mary because they were spending less time and effort on the scenarios the AI could handle. The path to the doctor was cleared by an AI that was sorting the ‘busy work’ from the very work these irreplaceable healthcare professions were trained to handle.
This is the future of healthcare: AI and healthcare heroes working together to do what each is best at: the AI handling the busy work, the humans taking on the challenging tasks they spent so many years in school and training to learn to handle.
And that is how all healthcare technology should be focused today: getting doctors and nurses away from the clicks and back to the care, back to the human element of medicine that is increasingly absent in a broken system that drowns our healthcare professionals in minutiae and a constant buzz of clicking and typing.
Want to find out more about WellAI and the future of AI-driven diagnostics? Click here to visit their website.
(Corporate Social Responsibility Disclosure: Christopher sits on the Board of Advisors for WellAI. To connect with Christopher, follow him on LinkedIn.)