Despite it's long-term benefits, new tech inherently disrupts physicians' workflows. How can health execs overcome this cultural resistance?
Artificial intelligence and other digital tools hold tremendous promise in healthcare. Thanks to emerging tech, physicians will soon spend less time wrapped up in tedious paperwork, for example, and handling other dull, unproductive administrative tasks.
But there’s one big hurdle in the way: Physicians spend about the same amount of energy demanding improved efficiencies as they do rallying against disruptions to their workflow.
How can hospital executives overcome this conflicting state of affairs?
A Healthcare Analytics News® Peer Exchange®
Kevin R. Campbell, MD: Let me change the discussion just a little bit. We talked about disruption, we talked about why and how and what we might do. But what are the biggest challenges those who enter a disruptive market as a disruptor face? I run a company that is now using AI to process device data from pacemakers, defibrillators, or any type of monitor in the world. We have challenges every day on how to enter that market and change physician behavior and get patients to accept that what we are doing is helping them. What do you guys see as a disruptor? What challenges do we all face?
Geeta Nayyar, MD, MBA: In the AI space—I’m glad you mentioned AI in particular, I think we all saw the news about Watson—there was a lot of hopes, a lot of promise built around Watson. Now we’re seeing layoffs. It’s a great example of a big nonhealthcare company gobbling up a bunch of really amazing startups and botching it, right? I think the challenge when you are a nonhealthcare company entering healthcare, one, is knowing healthcare; and two, when you’re a large health technology company or a large technology company, how do you take small pockets of innovation and actually streamline them to work? Again, I hate to use the Allscripts example again, but one of the cool things Allscripts is doing is investing in a new EMR, Avenel, which is using AI. For clinical decision support, they are wanting to predict and help physicians, smartly and rightly so. That’s one of our biggest problems with EMRs. All of us as doctors hate every EMR, right?
But if I would work with an AI technology, a machine learning technology, that actually helped me, knew what I liked, and knew that I was a rheumatologist even before I had to type anything in, that’s amazing. There’s a lot of opportunity, but there remain many challenges.
Kevin R. Campbell, MD: What about the use of AI for data processing in terms of helping us manage? Just in my space, there are data coming from defibrillators and pacemakers every day. It never stops. Half of it is trash, half of it is clinically actionable. What about using AI, machine learning, deep learning even, or teaching machines to help us process these data? What do you guys think?
Jane Sarasohn-Kahn, MA, MHSA: It’s time to do that. In particular, we talk about radiology being a prime opportunity because it’s so big. There are so much data in one image. I was just at the European meeting on radiology in February, and they had a whole section devoted to AI showcased in radiology. I’m this old, where I go back to the old juke boxes and PAC system, which was 15 years ago. AI is really important in this instance because you could actually look at a patient’s record, the longitudinal health record, bring in lab tests and pathology, mash it up with a lot of imaging data, and then get to that n-of-1 actionable moment.
But to Geeta’s point, to this question about disruption, we say culture eats strategy for breakfast. You take a big scalable solution, overlay it on this fragmented environment that I talked about earlier, and there could be some insensitivity in terms of physicians’ workflow and how things get done in healthcare. I think this is one of the concerns that I have. When we use the word disruption, doctors really don’t want to be disrupted because workflow is how they make money.
John Nosta, BA: Doctors like the same but better, right? They want a new cephalosporin that has a better resistance pattern.
Geeta Nayyar, MD, MBA: No, I disagree. Here’s the thing, we went to medical school. I’m a rheumatologist. One of the reasons I went into rheumatology is because it has cutting-edge genetic research biologics. We love innovation, actually. We’re used to it, however, on the clinical side. For a lot of us as doctors, it’s all about innovation. But when it comes to the back end of the office, that’s where our challenges are, because we didn’t go to medical school to focus on the back end. We want to see patients, I believe that. I picked out Allscripts only because for us as physicians, the EMR industry has totally disrupted our day.
Jane Sarasohn-Kahn, MA, MHSA: Absolutely.
Geeta Nayyar, MD, MBA: It is incumbent on every EMR company—Epic, Cerner, Allscripts, NextGen, eCW—that they have an opportunity to disrupt, that is my point. With Allscripts, I have watched them change their flavor in the market because it’s so easy to sit back on their laurels and say, “Hey, we got the outpatient market,” right? Whether it’s Epic or Cerner, I think we should be looking at them and putting pressure on them and asking, how are you disrupting?
Kevin R. Campbell, MD: I agree.
Geeta Nayyar, MD, MBA: You guys have the entire industry, so what are you going to do with it?
Kevin R. Campbell, MD: I will say that, to your point of disruption, athenahealth has something called More Disruption Please, where they want startup companies to basically pitch to them. They will put you on their interface if you show value to patients, physicians, workload, and customers. I think some are doing it, but we need more to do it. I’m sorry I interrupted you.
Colin Hung, BaSC: No, I totally agree. To hit on some of the points being made here, we talk about workflow, and to me that’s the biggest barrier. People will create these amazing technologies and then fail to accommodate for the disruption in the workflow in the clinical practice.
I look at AI and I’m more encouraged by the nipping at the edges. What Nuance is doing with their AI is help radiologists manage by exception. Instead of having to read through every single image, they can use some AI and say, “Well, we’ve identified these 10 images that potentially look like there’s a problem. Now please go and take a closer look at this.” They still need to look at all the others, but at least the AI is helping the workflow there. Instead of looking at a stack in chronological order—first/in, first/out—they’re now able to look at it prioritized order, which is something new and it helps the workflow.
John Nosta, BA: It’s fascinating. I think the fundamental, startling reality is that humanity is overrated, and that artificial intelligence is now judged in the context of humanity, which is the big mistake. We’re looking for an AI, a chatbot interface, that is as good as a human. That’s such a mediocre goal. It’s going to be better than a human, which puts in question humanity, in that broader context. I think that’s where the interesting part is.
Geeta Nayyar, MD, MBA: I love where you’re going John because one of the things that we’ve actually done in our practice, and you mentioned this, is patient education. We’re actually using AI from a very small company, a startup company called Chirp, and that’s exactly what it is. It’s Chirp. We are having this chatbot talk to our patients. What is a mammogram? What is a PAP Smear? What do these cells mean, 24 by 7? It’s not replacing a person, but when it’s 3 in morning, you can’t sleep, and you want to interact with one of my doctors, you interact with the chatbot. Hopefully, you bring that piece of the conversation in and are better informed when you actually do come meet one of my physicians. Does that make sense?
John Nosta, BA: Yes. I think it’s a bit of Alan Turing and the Turing Test. I think that’s the old modality. The new modality is that I can use AI to create a super human construct that I can speak to my patient with, perhaps from a genderless perspective, on demand; that I can listen to the voice of that patient and quantify word choice and say, “Last time we spoke you said pain 7 times. Today you said it 37 times.” It’s an intangible factor the physician might not get. I can listen for breath and I can listen for voice and I can check for things like Parkinson’s disease. To me, AI-mediated physical exams and all those clinical constructs are extraordinarily important. If I were a physician I’d duck. Because AI is coming at you that fast.
Colin Hung, BaSC: The example I use when people get that fear in their eyes around AI taking over their job is the auto industry. When you look back to the old days, mechanics had to know everything about the car and they were the only ones able to fix it. Now, what do mechanics do? They plug in a device and they look at the device and they’re fixing your car. Do we have fewer mechanics today than we did back in the day? Nope. They’re still using their intelligence, they’re just using tools to do things faster, quicker, better, cheaper. To me, that’s where AI can definitely help us.
John Nosta, BA: I would quantify the number of people on the line. I think it’s actually decreased. But again, I don’t mourn the blacksmith.
Kevin R. Campbell, MD: I would agree though. Medicine is very different, and with physicians, there’s still a doctor-patient relationship that’s absolutely critical to healing and caring. But you should use all the tools. As one of my mentors said, the problem is that the digital world is over here in a separate orbit away from physicians. It takes doctors so long to adopt change, and that’s unacceptable and that’s what’s holding us back. What do doctors change? They change based on data and papers published in medical journals. They base change on what the elders in their society are doing, the thought leaders, and what their buddies and colleagues are doing. We have to lead the change. All of us together have to lead that change.
Get the best insights in healthcare analytics directly to your inbox.