Search
Close this search box.

Episode 10

Will AI Take Over Physical Therapy?

Vice President of Clinical Integration and Innovation at the American Physical Therapy Association Drew Contreras stops by to talk all things artificial intelligence within the Physical Therapy and Rehab space! Some of what we talk about: Drew’s background and his experience as a physical therapist assigned to the White House Should PT’s be worried about A.I. taking their jobs? The current state of A.I. and it’s possible future
Published on 07/24/2023
Play Video about Therapy Matters Podcast Video Cover

Episode Transcript

Allison Jones: Good morning, good afternoon, and good evening, and welcome to the Therapy Matters podcast, your one-stop resource for expert insights and advice on everything therapy and rehab. I’m your host, Allison Jones, and today I’m joined by Drew Contreras, the Vice President of Clinical Integration and Innovation at the American Physical Therapy Association. Drew, thank you for joining me today. 

Drew Contreras: My pleasure. Thanks for having me. 

Allison Jones: Excellent. So Drew, you’ve been in the industry for over 25 years and you’ve had a very accomplished career. Let’s start by giving our listeners a little bit of your background. 

Drew Contreras: Sure. I am a retired United States Army physical therapist. Did that for a while. In that role as a military physical therapist, the unique thing compared to a civilian therapist or your traditional clinical setting is that the privileges and skill sets are vastly different for a military physical therapist than a civilian.

You have advanced skills that you’re allowed to use and you’re allowed to order imaging studies and limited medications and things like that. But that also means you get to go to glorious and lustrous places far away from your family and friends and hang out there for a while. So, there’s that too. But during the course of that, time on active duty, I had a very interesting job where I was assigned to the White House where I was the physical therapist for the Obama administration while I was on active duty.

That was a completely different setting than your typical physical therapy setting. But it was not without its interesting things that you see and do and the privileges of doing it. So that was my clinical career as a clinician. After I retired from the military, I decided that I wanted to try and have a broader impact on the profession.

That’s what brought me over to the American Physical Therapy Association to where I’m now. 

Allison Jones: And tell me a little bit about what you do at APTA today.

Drew Contreras: Yeah. My job is really, in what we would call, the junction between clinician and practicality of clinical innovation and implementing those innovations.

So what that means in a sense is if there’s something new that could potentially change the practice and the way we deliver care, my job is to know about it and to determine if is that something that we should be doing as clinicians. Likewise to look at things and say, “Maybe that’s something we shouldn’t be doing”, and we should not be implementing as clinicians.

But the association has never been the one to lead in this space. The American Physical Therapy Association essentially lets the practice be driven by clinicians, which is how it should be. But what it never really did a great job of doing was when we found something that worked well, that the research supported, that clinicians were having great success, we weren’t great about disseminating and sharing that information amongst the profession. And that’s really what falls into my portfolio, figuring out how to do that, how to get that information to people so that we can have the best impact we can with the patients we serve. 

Allison Jones: Okay, great. Today, our focus is on the uses and impact of technology in physical therapy, and this stems from a conversation that happened a few weeks back with two of our colleagues on a previous episode of Therapy Matters.

My co-host, Scott Rongo, spoke with your boss, Justin Moore, about the challenges and opportunities facing therapy over the next five years. And they talked about three different areas, but one of them was emerging technologies. And Justin said to us, “You need to dive deeper into this topic. And to do that, we have to speak to Drew. Drew is your guy.” So no pressure, but we’re expecting big things today. When we’re talking about emerging technologies and the impact of technology, there’s one thing that is on everybody’s mind right now. And that’s AI.

AI is a hot new technology. Everybody’s talking about it. You can’t open up a webpage. You can’t turn on the news. You can’t go to a meeting practically at all today without somebody talking about AI. And lots of people are really worried that AI is going to replace their jobs.

Let’s just address the elephant in the room right up front. How worried should PTs be that AI is going to replace their jobs?

Drew Contreras: So what’s interesting is if you want to, just log in to ChatGPT and ask it. Because I did. “Hey, are you going to take my job? Are you going to get rid of me?” And the actual answer that the AI platform gave me was like. “Yeah, probably not. Because, physical therapy requires human intervention and that cannot be done, through these kinds of platforms.” And so jokingly about it, that that’s not the intent behind an AI platform. It’s not like they’re gonna take your job. I have a couple of caveats to that, though. Caveat number one, this is assuming you’re a clinician practicing the standards of care and doing good stuff.

Taking care of patients, and providing quality evaluations and treatment plans. Now, if you’re a terrible clinician, and my favorite example is everybody gets yellow theraband clamshells no matter what their problem is. If you go into an AI platform and say, “Hey, give me a rehab program for somebody six weeks post-op from an ACL”, it’s going to give you a nice program.

So if your whole clinical essence is just programming of exercises and has nothing to do with progression, regression, evaluation, or diagnosis, you’re not going to be doing that much longer. Because it can do it better than you. The take that we have is that technology will never replace a physical therapist.

But a physical therapist who is technologically savvy and educated will replace a physical therapist who is not. And that I am certain of. 

Allison Jones: All right. Give me some examples of how AI can enhance physical therapy. 

Drew Contreras: Sure. As you said, the news is everywhere on AI.

I think that a lot of healthcare providers, for lack of a more sophisticated word, pooh-poohed AI until it passed the medical licensing exam. Then all of a sudden a bunch of physicians started paying attention, “Hey, wait a minute.” Now we’ve also seen an AI there was an AI platform that went up against a team of a subspecialty of radiologists, but very bright men. And the humans beat out the AI platform, but just barely recognized diagnosing imaging studies. They were better, but not tremendously. And let’s also be honest. Four or five radiologists, at whatever a radiologist costs a day, look at pictures versus an AI platform, which runs for pennies. The scaling and the termination, the platforms are going to be good enough to handle a lot of screening.

And that’s the truth. It’s going to be able to recognize the patterns and the things as you set it up. Like, hey, this is what a carcinoma looks like. I’m looking for a medial tibial stress fracture. It will pick up the patterns, but it will never be able to correlate that with a physical examination.

So the point is that these things are coming and those pattern recognitions of what they excel at, they’ll be able to get in there and change the way that we do things because you won’t have to spend the clinical time in some of these things where it could be quickly screened for you, and then it can you can take out 10 or 15 diagnosis fast.

And we could just move on to the things that are most likely to be worrisome and then really focus on the plan of care as opposed to that taking a few visits, and a few weeks to get to that point. 

Allison Jones: Yeah. When we were doing the prep for this, you had talked about an AI tool, you could use on your phone to evaluate the movement of your patient.

Drew Contreras: Yeah, definitely. So here’s the conversation I love to have with clinicians who are skeptical of this. I say, what is your gold standard for measuring motion right now? And the answer most clinicians will tell me is, “Oh, I use a goniometer.” I say, “Great. You’ve got a piece of plastic with a metal rivet and a couple of Sharpie marks on it. And what’s your plan for measuring joint motion.” Or I can get out essentially a microcomputer with the screen in my hand, take a picture or a video and get reproducible things, the variances are minuscule and it can give me exact measurements to use with my patients using an AI platform that’s doing motion capture. 

So why would I not fold that into it? It’s faster. It’s easier. Its inter-reader reliability is astronomical compared to standard goniometer measurements. So why wouldn’t I use that? It’s just because I’m not familiar with it. Or it hasn’t been woven into my clinical practice yet, but it’s there now.

We have them now. It’s available. And what you’re going to see is things like this will very quickly get folded into our practice. And the reason is not always the obvious thing. So obviously more accurate measurements are better. That’s a good thing. Us being able to deliver high-quality care should be something we always strive for. But as the business of medicine continues to operate in the face of dwindling reimbursement, money’s given to clinicians for the care they’re provided, we’re going to continue to have to provide true evidence that what we do is effective. That the care that we give translates to results that mean patients live more independently, that they can safely return to work quicker than if they had not seen us. That they’re available to do those things.

And having high-quality measurement options available will support that as well as support improvement and reimbursement. These things inevitably come into practice because it’s not just for the side of the clinician in the quest for a cleaner and better measurement of our patient, but also in the management and administration side of being able to provide quality data that we can use to support not only that clinician but the health care outcomes that are demanded of us as clinicians. 

Allison Jones: Supporting the shift to value-based care. 

Drew Contreras: It’s inevitably coming. The difference as a clinician is what’s the physical therapy HEDIS measures. Whereas, other disciplines have measures of lab values or other outcomes that are very easy to measure. We’re not as tangible in some of those aspects. So, this will allow us to provide an equivalent measurement that can demonstrate the quality of care.

Allison Jones: So we’re talking a little bit about the benefits of AI, and how it can help. There’s also a lot of news out there about the potential drawbacks of AI and the warnings of AI.

There were some news articles the other day about some big CEOs and tech heads warning about the downfalls of AI. What drawbacks could there potentially be for AI? 

Drew Contreras: There are. And there’s a work group that is being put together now, the grandfathers of AI, if you will, the three wise men who spearheaded this. And you’re right. On a large scale, I think you will see that elected officials and leaders will be waiting for this. Because it’s uncharted waters. And there needs to be some sort of regulation monitoring it. But as far as, what are the warning things that PTs should be concerned about, this just happened and it was really interesting. 

A lawyer used an AI platform to prepare a legal defense. Just jammed it in there, spit out the legal defense, and gave it to the judge. Judge came back and said, “Hey, the references in this legal defense can’t find some of them. Can you go back over this?” So put it back through AI, generate another one, and send it back to them.

They pulled through it. All the references and all the things that the AI platform had referenced were made up. It’s completely fabricated. Because it’s an AI platform. It doesn’t know. It’s just doing its thing. I think that the biggest initial problem that we can potentially see is people using AI platforms to try to work around the drudge work of research.

That’s the initial thing that I see. Whether that’s intentional or not, I’m not here to say whatever. But I think that the potential is there where, you could say, this looks good. I’m getting great results on it. We’re just gonna provide some more evidence to support it.

And that evidence, if you’re using an AI platform to generate it could be completely fabricated. So by not going through a true process and understanding truly scientifically that this works out where it doesn’t work out, you run the potential of whatever that intervention was being branded as unacceptable if you’ve gone on that road too fast. I think that’s the first and foremost. That is the thing that I potentially see right away, struggling with that. I think that those things that require a lot of tedious churning are going to have the biggest temptation to leverage these sorts of platforms to do that. Now, another one that I would say has a lot of potential, let me also throw this out there AI is not new. The truth is there hasn’t been a whole lot of scientific advancement in the last few years.

It’s stagnant where it’s at. People are just hearing about it now. There are those companies that will use an AI tool to generate recommendations. Maybe a patient does a thing. Whatever that thing is. Maybe it’s a handwritten test.

Maybe it’s a functional test. Maybe it’s a stand on one leg and touch your nose. Then they will say, “Oh we’ve run it through the super algorithm AI. And here’s the recommendation. If you do these things, your problem will go away.” There’s a temptation there to allow platforms to take some of that clinical decision-making away and get to the result faster.

Hey, look, if I’ve got 15 patients on my schedule that day, and I’m churning and burning on the plinth all day long if it’s just another ankle sprain, we’ll just run through the thing, I do these things and it’s good to go.

That’s going to be a big temptation. But the question is going to be, when it’s wrong, how are you going to defend that? You were like, I thought it was just an ankle sprain. You miss this, you miss that. 

That’s gonna be indefensible because it’s not the standard of care, and that’s the next place that I see clinicians potentially getting stuck with. The temptation to let it expedite your workload in the defense of the clinician, I got a bunch of patients to see, and I got a bunch of work to do. But when you take that without going through the true process, you inevitably will be stuck trying to defend the indefensible. 

Allison Jones: But you could use it after the initial evaluation and you understand what the problem truly is. You could use it post that.

Drew Contreras: Yes. And you can use it to trim it down, to make your evaluation more precise, get down to a more precision level, to rule out a bunch of things that make sense. But that means you still have to do the work. You just are enabled with more tools, which is excellent. That’s what we should be striving for as healthcare providers. But if you let it run the system, that’s where you’re gonna run into problems.

Allison Jones: So, it’s workflow efficiency. It’s about reducing the burden. But it’s not about taking over. 

Drew Contreras: Yeah, exactly. There’s another AI platform that is out there where you can essentially audit a clinician’s notes. And what you start to realize is that PTs do two things historically as a professional. Number one, we over-document. We read a whole bunch of stuff in areas of a note that doesn’t matter. And a lot of that has to do with kind of the way we were brought up in the system. And the majority of people are type A control freaks. And then you just really want to do a good job.

But to your point on the workflow, if you’re spending an extra 25% of your time writing a note, how much does that equate to? How many extra notes have you written that day? So using AI platforms to help you understand that you didn’t need this, you got all the things you needed, yes, that’s great. But you also did a bunch of work you didn’t need to do. Just helps with your efficiency. And then the other thing PTs tend to be cautious of is overbilling. Because nobody wants to trigger an audit. And if you work in a place that has compliance departments of lots of people who do all that, awesome. But I think the vast majority of people do not.

They don’t have like, “Oh yeah, we got a compliance department of 20 people going through everything.” No. It’s, “Hey, I hope you didn’t mess up that Medicare note.” That’s what it is. So, the tendency is to underbill because you don’t want to trigger that. You want to make sure to just be on the safe side.

I don’t know. Was it eight minutes? Was it seven minutes? How many units was that? Is that 20 minutes? The biggest challenge of a clinician’s life is how many minutes and what things am I checking for a treatment plan. So if a platform can help you understand that and make sure that you’re not leaving reimbursable codes on the table, then that can help. Then maybe you don’t have to see 15 patients. You can only see 10 or 11.

You can make it a little more reasonable or increase the salary of clinicians because they’re bringing in more revenue. There’s great potential for this to have a huge impact on the profession if folded in properly and we understand what it is and what it’s doing. 

Allison Jones: Absolutely. From a billing perspective, I feel like the AI aspect has been there for a while. You have a lot of automation in billing systems today which is using that AI to help flag where there might be coding errors or denial problems or issues that might cause some of those red flags or cause some of that under billing or over billing and some of those issues.

Drew Contreras: Right. And let’s be clear. On the other side, the payers are using an AI system as well. Let’s just call it what it is. So, it’s on both sides of it. Who better to learn from them than the payers? It’s the same system, the workflow is just going the other way.

Allison Jones: Right. So, use it to your advantage.

Drew Contreras: Yes, exactly. Use the powers for good.

Allison Jones: Right. Instead of evil. Excellent. What else can AI be used for to reduce the administrative burden? Is there anything else that we haven’t touched on from an AI perspective?

Drew Contreras: I think the one thing that the profession will have to deal with is where does it sit in the education of clinicians?

Is it a tool that you can use as you’re going through your educational process? Is it to be banned completely? It’s everywhere. It’s in my kid’s high school. I don’t have the answer. I feel like it sometimes like, can I use a calculator? Is it gonna be any different than all these things along the way where a technological advance was given and you were like why can’t you use the technology? Because it’s there and it’s pervasive. It seems silly. Try doing calculus without a calculator. That’s insane. But they used to not. They used to use beads on a stick to count things. So at some point, when does the technology become accepted into it? And that’s going to be sticky. That’s going to be, in my opinion, harder to delve through because you’re going to have to get serious conversations about dogmas, educational philosophies, and all those sorts of things.

And I don’t have those answers at all. But it’s inevitable. Because then it’s part of the upbringing. It’s woven into the fabric of the profession and it will be there. We’ll get there. I don’t know what that’s going to look like. I think that we’re quite a bit off from that before we get to that place.

But someday, I’ll be talking to somebody who’s like, “Oh, AI platformed a third of my lectures.” 

Allison Jones: Okay. Let’s switch gears to APTA. When we’re looking for more information on AI technology topics, or any other topics that relate to the therapy industry, APTA is a great resource for people.

Tell our audience a little bit about why they should be joining as members of the organization and what they can gain out of joining. 

Drew Contreras: Yeah. There are two points about being a member of the APTA that just need to level set.

If you’re looking for just a straight transactional relationship, it’ll be difficult for you to decide if this is something I should do. I work there. It’s not the same as when I go to the grocery store, I pay this much money and I get this.

There are a lot of things in APTA that are valuable for your membership, but they don’t always apply to you as a person. So you have to decide maybe I don’t care about the magazine or I’m not into the CEUs that the APTA offers.

Maybe this is not for me. Maybe that doesn’t work for my practice setting. But I think that if you ask yourself, am I part of the profession in the community? When people ask you, what you do, what do you say? And if you say, “I’m a physical therapist” or “I work in a physical therapy clinic”, or whatever that is, then what community are you a part of? Who are your people? Where is that at? Is it just the two people in the clinic you work with? Is it that or is it more? And I think if you want it to be more, if you want more than just, I come in, I punch the clock and I go home, there’s nothing wrong with that. There are a lot of people with a lot of things going on in their lives and that’s just fine. But if you’re, if you want to be more of a professional. And if you want to live up to the title ‘doctor’, you should do some doctoring.

And you should be a part of a community that exposes doctoring and exposes the exchange of professional ideas and knowledge and find those people that make you strive to do good clinical care or find the next treatment that you’re stuck on. You can’t figure out what to do with these hip patients.

And maybe somebody will be able to help me. Or maybe it’s, you know what? I want to start working with Athletes and kids because now my kids play soccer and they’re nine years old. Then how do I connect with the people that do that? That’s what the APTA does That’s what the APTA is for.

It builds that networking community for you to be part of the doctoring profession. And that’s the value of it. The other portion of that aside from finding your people and finding the community that you belong to is that whether you’re a member or not, if you have the initials after your name, the association’s working for you. There was just a large bill that was passed to make sure that telehealth reimbursement continued past the end of the COVID extension that was done. And the APTA pushed for that, lobbied for it, and made sure it happened on behalf of physical therapists.

And at the beginning of the pandemic, physical therapists weren’t even included in telehealth reimbursement. The APTA made sure that they were. The APTA is going to continue to do those things whether or not you’re a member or not. It’s going to continue to do that work because it’s important and it’s the right thing to do to advocate for the profession.

But if those sorts of things excite you or if making sure that your profession is represented, for example, right now the country’s stuck on what’s long COVID? People are like, “I don’t know if it’s a diagnosis.” People that are arguing over all these things, but APTA is working with the Office of the Secretary of Health and Human Services to be a voice at the table.

Working in conjunction with the PM&R Society, with AOTA to make sure that we’re providing those things for clinicians who are working in these areas and trying to address the bigger problems of our society. That’s what APTA advocacy will do. And those are the things that you can see yourself in.

It’s the only place that does it for our profession. So, that’s another place that you just can’t get it anywhere else. That’s really what it’s about. 

Allison Jones: Excellent. Thank you for that, and we appreciate everything that the APTA does for the therapy industry and all the practitioners out there.

Wrapping it up, any final thoughts or takeaways for our listeners on today’s topic? 

Drew Contreras: I think if I just had one message to say, it would be that the profession has hit a pivot point in the last few years, especially coming out of the pandemic where Society as a whole is accepting of technological advances, especially in medicine.

And then our profession has the opportunity to decide what’s part of how we practice and what we fold into how we do that and how we deliver that. And I would tell people to not shy away from it. It’s very easy to slide back into the comfortable practice patterns of circa 2009. It’s easy. We already know how to do that.

But it’s not sustainable. It’s not going to help us as a community deliver better care or even stay relevant, to be honest. So my advice would just be, don’t be afraid to get involved. Be curious. Look at the things that are out there, find out how you can get involved in these things, and how you can fold them into your practice.

Because the change is coming and it’s coming fast. We had another project we were looking at where we realized that in 2030, kids who are graduating high school this summer 2030 will be the new PTs. So the question is, what does that look like for them?

How will they practice? Will they even ever touch a goniometer? Will they be looking at it as some archaic thing like, “I can’t believe you guys did this? I can’t believe you guys did manual muscle testing and thought it worked.” What will they look at and make fun of? And what’s it going to be like?

I think that we should be open to embracing those things with a skeptical eye. Don’t ever lose that. Don’t ever lose your critical thinking and decision-making. But also understand that, that the inevitable change is being accelerated and it’s just here to help.

Allison Jones: I like that message. Embrace change. Great. We could probably go on for quite a while, but we are out of time for today. I want to thank you so much for joining us today, Drew. Appreciate your insights. I want to thank our audience for tuning in to this episode of Therapy Matters, your one-stop resource for expert insights and advice on everything therapy and rehab. And we look forward to seeing everybody on our next episode. Thank you.