AI Is Already Running MSK Services. Are PTs Ready?

You talked about this tension where this rigid guideline adherence can actually be unethical practice, which that's gonna explode minds like that. There's gonna be brains that will just, that just can't be that way.

they're not human. They're non-human, and if we think of health and life cycles as being inextricably linked to all of those things, then it radically makes opportunities available for us to rethink the physical therapies in ways that can provide therapy outside of our musculoskeletal physiotherapy, ways of practice.

What if following clinical guidelines could actually be the unethical choice? That's not a rhetorical question. It's a serious argument. It's one of the first places we go in this conversation. This is part two of my conversation with Matt Lowe, who is one of the Sharper Thinkers I've encountered on what it actually means to practice well in a complex world. If you haven't listened to part one, I'd recommend starting there, but this episode does hold its own. We pick up with Matt's work on epistemic fluency and critiques of value-based care, and then we move into territory. I think every clinician needs to sit with what AI is actually going to do to this profession. Not in theory, but right now we also get into something, Matt calls the contextual architect and why the future of physiotherapy might look nothing like its past. We close with a question about ecological thinking and what it means to see the patient as inseparable from the world around them. We'll have you rethinking what is possible for future clinical practices. Let's get into it.

This is the Modern Pain Podcast with Mark Kargela

I wanna get into something you've, you spoke about in, I got a behind the scenes peak at your, uh, epistemic uh, fluency paper. You talked about this tension where this rigid guideline adherence can actually be unethical practice, which that's gonna explode minds like that. There's gonna be brains that will just, that just can't be that way. So that's a provocative claim to say the least. I mean, especially with a profession where physio, we've really. Been staunchly trying to be, you know, evidence-based and things. So that really is probably a tough pill to swallow. How do you make that argument as somebody who's pushing back on you on Yeah, because I think there's obviously some context that needs to be

Yes.

Brought out into that discussion, which I know you're gonna give me, but Yeah. Um, yeah, it, I thought I was

plane for sure.

I just, when I, when I read it, I'm like, whew, I could just feel the, I could, I, like, I loved it first off, but'cause I agree. Um, a wholeheartedly, uh,'cause I see it. Like, there's some very guideline based treatments that can be very harmful to patients. And I think there's, if, if you understand context where that exists. And I think Can you speak to that a little bit?

Yeah. I think, um, so first of all, guidelines can be shaped in the, any way that Debi, that biases the clinician to a greater or lesser degree. So if we take back pain, for example, when patients have gone through episodes of for ex, um, uh, uh, best. Best practice care, physical therapy, multimodal kind of approaches. Um, and they face an MDT. You know, one of the options that's guideline based is a medial branch block injections, plus or minus ready frequency innovation, or if it's in the case of a sacroiliac joint problem in the uk, you know. If people have failed physical therapies, they go through into, uh, an MDT situation. They've had some pain provocation tests, which are positive. They have an SIJ injection, which improves their symptoms for a short period of time, and they'll fuse this. SIJ there's so many reasons why. I'm not sure I'm very. Ambitious of that. Um, so we know that, say for example, the sacroiliac joint injection that can, um, have a general steroid effect. I know that there will be some surgeons who will push back quite strongly on this, but they'll say that yes, well, there'll be a local anesthetic effect. We know that that's the source of the symptoms. They'll also do post hoc reasoning where they say, oh yeah, so there is the, this number of patients have got better with the sacroiliac joint, um, fusion. But oh my gosh, how many patients have not done well with the sacroiliac joint fusion? That's following guideline care? That is following guideline care. That is to some concrete examples. If we go to, and that's fairly strong, hard line orthopedic ones, which are in themselves quite controversial. There are lots of people who would say that that's not, not best care per se. Um. Let's say if there are patients who we follow, uh, guideline based care, but their social circumstances are so incompatible with what they're available to access in terms of the guideline, we can't provide care according to that guideline because their particular circumstances mean that they are unable to get. Access to the guideline. Hence, we need to reshape how we help that patient. And there are all, and in the, in the paper, which is going through peer review, and we shall see where that goes. But, you know, there is a, a, a case study example of a patient who is based on a, is a, based on a real case, um, uh, where their social circumstances, without putting too much emphasis on this, uh, is just means that they, um. Uh, uh, uh, not able to follow the guideline. Hence, we have to reshape the guideline for the patient. Now, some people will come back at me and say, of course, guidelines can be reshaped. You can choose not to follow the guideline. The, the pushback that we get from clinicians, insurance companies, um, and all sorts of other things for not following guidance is pretty strong. Uh, even though we are following or trying to be the best evidence-based clinicians that we can be, where we're coalescing those three. Aspects as we talked about earlier. So, so yeah, that's, that's basically where at times following a guideline may actually cause the patient harm.

Yeah,

basically where, where, what, what that was trying to, to get to.

No, I and I, I liked it, uh, because I said I just, I, I've seen it and, you know, si joint's a great example of that. I, I, you know, I think in the UK you are much more adherent to guideline in the US there can be, you know, not as following of guidelines, uh, from just my talking to you and other, other clinicians.

Yeah. So guidelines can be seen as tram lines. That's one of the issues I think when it comes to, to, to guidelines. If there is that recognition of a broader scope of what a guideline is, then that's, that's okay. But um, unfortunately, there are times where guideline is not a guideline, it's a tramline.

It's

like if you step outside that you're not practicing properly. In fact, actually the, the pink finger gets point to you as the clinician saying that we are practicing and ethically, and that that's, it's a difficult pill to follow fo to swallow for some where you are. I'm trying to help the patient in the best possible way.

another thing you said in this paper that you had kind of argued that these four pillars, uh, we're gonna get into value-based care. Um, you had, you had mentioned, I think Chad Cook's paper on value-based care. Yeah. Pretty nice

paper.

Yeah, good paper. Um, you argue that those four pillars of value-based care contained some inherent frictions rather than being naturally synergistic as they're kind of portrayed in that paper. I'm, I'm wondering which of those tensions you think is the most kinda. Maybe underappreciated in day-to-day practice.

So in the paper, it, uh, I hope it gets published because you'll be able to articulate it far better than I can from memory now.'cause we submitted it some time ago. Um, but, uh, there was, um, so basically comes round to the idea of what value is. So, um, this is a bit of a sideline I'm gonna talk about, because if we look at evidence-based practice, it's a values-based model. Um, and then I'm gonna come back to cook's. Um, paper. So if we look at, and people are probably thinking what you want about evidence-based practice being a values-based based model in practice, evidence-based in practice, the EBM or EBP, uh, model. Um, best of research, best of the clinical experience and best of patient values, uh, are all value laden. That is not an obvious claim, but it is a claim that I stand by. So when the patient, sorry, when the therapist selects the appropriate evidence, they're making a values-based claim on what they think is legitimate evidence to apply for that patient, that's a values-based claim. Uh, someone will come back to me and say, yeah, well, no, there's. That that's not the case, that's not the case. Uh, because you're following guidelines that's being predetermined. It's sitting outside of that. I'd still say that how the patient, how the therapist applies their practice will come from a value on what research evidence has informed their practice. And it may have been an inherited form of practice from their supervisor or so on, but I'm still saying it's value-based. Then you've got, um. The, uh. Uh, yeah, clinician's experience, of course, that's values based. And then of course, you've got the patient's values. So value is the, is the challenging and tricky concept here, and that creates the tension. So in, in, uh, Chad's paper, which I think is a really nice paper, in, in, uh, it, it, it talks about values-based care, talking about low values, uh, low value care, uh, a cost effectiveness. Um. Gosh, there's four dimensions of it and I can't quite remember which, how they coalesce now, but essentially, um, what, uh, all of those factors does depend on the context under which the, the patient and the therapist encounter each other. It is not set that, uh, the, the value that that person, um. The term value is going to be contested. There's going to be an inherent friction between what is best practice and what is low cost, and what is perceived by value from the patient and from the therapist and from the system. So if we see the systems value in terms of patient experience, or in terms of economics, that may be very different from the patient's value. Um. From their own perspective as opposed to a self-reported measure versus the, uh, clinical guideline, uh, or, or the clinical research that they're coming together. I think in principle it sounds good. It sounds like they all come together quite nicely, but as soon as we talk about the social determinants of health, as soon as we talk about the environmental contexts and the social political environments that these people come from, now we have a very different, um, ground. I've con contested, uh, ideas about what value is, and this is why I'm, I'm, I'm not making any negative claim against Chad Cook's paper. I think it's pretty good. I just think it requires a form of, um, fluency of understanding of knowledge and an understanding of what value is and how you apply that know. In practice.

In context. In context, yeah. Where, where we don't know how those four factors or four pillars are realistic for the context that this person sits in. Um, and the context you may sit in, in your, in your situation with, with that patient. Um, the other thing you spoke about in this paper that I, I, I noted is it mentions AI evolves to handle. Population level pattern recognition, the clinician's irreplaceable value will shift towards situational understanding. And you used the word, the phrase, and we talked about this earlier, the contextual engineer. Um, do you think the profession understands what that shift actually demands of Its of our identity. Like, it's kind of a shift a little bit right. Of, right of, of that. I'm wondering, I'd love if you can speak to that.

Okay. So I'm gonna go back a step and make a suggestion about what might happen with, when AI really takes a grip of the healthcare context. Um, so I think that, uh, healthcare is gonna hit an irreversible change. Um. And I'm not saying this good, bad, or indifferent, but I think it's gonna change. One of the things that I think AI is gonna do extremely well is manage musculoskeletal conditions in certain ways. Not for all, but in certain ways. I think it will. Um, in particular, how certain musculoskeletal pathways are managed in a very algorithmic way. This is gonna be ai, right? A piece of cake for AI to, to consume where we've reduced health care in a musculoskeletal context, in such a way that it's follow step by step rules that AI is, is just, that's bread and butter for ai. AI can do it way faster than us, and it can be a 24 7 therapist. Um, I'm not saying this is bad. I'm saying it provides the physiotherapists with quite a lot of opportunity for the future, um, but in very different ways, not in the ways that we're currently practicing. So, so what does that, what does that mean if AI kind of comes in and is able to manage populations using evidence-based approaches and following? Um, and pretty dynamically I'd suggest, uh, those, um, evidence-based pathways. What place does the physiotherapist have, and I think the physiotherapist has a place as a contextual architect. Basically it's following on from Michael Rose work and, and, and he's. Published a book I think coming up soon. I haven't read it and I don't think it's published yet, but he's written some really nice articles and he's spoken at conferences including ifu, where the physiotherapist, um, will have a place either to interpret some of the information and knowledge that the patient has come to them that they have gleaned from AI and to situate it for them. Or to be able to support and or to be able to support the patients to use the AI in such a way that it will help them. So, um, that's one of two ways. There's a, there's gonna be at least. Three or four different ways. But essentially the idea of context being a situated form of practice, that under that, that sees the patient as more than just a single entity, to see them as, um, in a multidimensional, uh, element, uh, multi multidimensional context, will provide physiotherapists with a way to interact with certain forms of technology in a way that will support. That patient I think in a, in a, in, in a fit for the future, I'd say in, in ai. So that's probably where,

mm-hmm.

I would say things are gonna go, we we're gonna be involved in kind of contextual architecture, supporting the patient in their own individual context,

which I love. And I, I, I, I hear clinicians and not just in pt, I mean, you see this in other. Especially the ones that are more algorithmic, you know, triage based, you know, primary care I think has some understandable concerns, um, that I would have too. Uh, with, with what we're seeing with ai, I, I'm a firm believer that if we leverage it the right way, it allows us to really leverage our humanity components, right? It can let us lean in and, and have some of these. You know, the more human humanity components of our care that, uh, you know, it's hard for AI to, to sit with a patient and just share a room with their suffering and, and empathize and validate and do things. Of course, it can validate and, and some we all know ai, sometimes it can be very syncopathic with it's validation and, oh, yes, mark, that's the greatest idea ever. You should do that with everything you tell it to, you're gonna do. Um. But I, I'm wondering what your view is on, on that piece as far as like AI's ability to help us, you know, maybe remove some of the administrative burden from us.'cause I'm now, I'm thinking more as like a clinician aid to help us, like get to the. I guess the meat and potatoes of like what it is that truly separates us as a profession that maybe it is touch, maybe it is, you know, other things of what we do. I'm wondering where you sit with with that.

Yeah. Yeah. I think, um, it's interesting we're saying things about the kind of narrative aspect of, of what AI can and can't do. I mean, there's research papers to show that it demonstrates empathy. AI demonstrates empathy better than doctors. I'm, I'm not sure if that's. You know, uh, I know we're not doctors, but I think it points to something. Um, so, so I think that AI will get to the point where it'll be able to take a history. It will be able to, uh, which I think you can do already. Um. And pretty accurately, I'd say, uh, it could demonstrate empathy. Um, I'm not saying it's gonna be perfect. I'm not actually gonna say it's better than, than physical therapists. Um, but I think, you know, it will be able to do pretty much everything that physical therapist can do. Uh, albeit at currently within large language models, it won't be able to do the physical aspects of course. Uh, and there's all sorts of talk speculatively about where this is going in terms of treatment, robots and so on. Apparently in Japan, they have. Massage robots and so on. Anyway, that's a different conversation. Um, so. So I think where, um, where we are already seeing some of the benefits in, uh, physio physical therapy departments is the use of, um, although you have to change your practice when you use it, is automating of record keeping. Uh, I think that is a game changer, particularly for burnout collections who just go behind for every single patient.'cause they're putting everything into their patient and then running behind on their notes. They know they've gotta get their notes done by the end of the day and they spend like an hour or two after the clinic. Like we all know we've all been there. So I know that I'm pretty confident ambient technology is gonna be, and that's already here and it, that's. It does change the way you practice, of course, because it can only pick up on stuff that you say so much. Like the work that you currently do, mark, when you've got a students in front of you and you are talking through what you are doing and what you are thinking with the, with, with the student, you'll have to adopt that. With the patient, of course, you'll have to have their consent and they'll have to know that all of that, um, you know, have you used ambient, um, technology for record keeping?

We, we've experimented with it a little bit and just like recording a transcript and then, oh, we don't have it like live with it, but it basically, it recorded the entire thing and I just shoved it in and it spit out a, uh,

good.

Pretty,

so, so there are ambient technologies out there, which you does it real time. Like there's no kind of recording stop, you know that kind of, and it will put that all of the context and the content. The difficulty is it's audible content, so there's all sorts. You remember I was talking about the different forms of knowledge? Yeah. Epitomy technique for Anis. It won't pick up the. Implicit forms of knowledge, the stuff that you see implicitly, unless you make it explicit. And that can be challenging.'cause there are gonna be times where you've seen something, you've understood something, but you've not said it. And that's the essis. There are times where we. There are just, it's just not the right time to say something. Um, so there are gonna be things that, but anyway, my point is we've got ambient technology, ai, and that is now I, that's gonna be taking away some of that administrative burden. I think triage, you know, referrals coming in. I think that, gosh, how much time do we take doing triage, doing it fairly manually. There will be places who have already got that automated pretty well. But for a lot of us, it's a long, drawn out process. Um, there will be some safeguarding challenges that we'll need to overcome, but I don't think it's going to be too long before all of the triaging process goes in. I think that there will be, um, a stratification of care where, say for example, on a start back measure, which. It's been taken up in different parts of the world to various degrees or use of repro measuring tools and stuff like that. If we stratify patients, the, the lower risk patients could be managed with AI and then if they get into trouble, they can then come in to see a physical therapist and so on and so forth. So there are different ways in which we could stratify care. Um, so I do think that that is going to take away. Full positive or negative, it's gonna change the way we practice. And that I'm not talking too far in the distant future. This is now, this is already happening.

Everything I hear you say is, is something that is, the capability is there at the moment. So, uh, it, it should have us take pause of like, we need to be having some serious considerations of this. We, we spoke about this, but I wanna bring it into obviously our, our listeners. Do you feel like there's any. Movement.'cause you know, we've had this neoliberal self-empowerment, autonomous creating autonomous patients that can self manage. Yeah, that seems like AI would be really good to to, to help that happen. So do, do you feel like that might, you might see a retreat of, of clinicians back into more hands-on touch based care and, and

I think there will be some clinicians who will double down and they will go, I'm digging in. I'm a hands-on physical therapist. We're gonna be going back to old traditional values and we're gonna, you know, this is very, very similar. If you look at a paper that, uh, Dave Nichols and his colleague wrote, I've forgotten his name, his colleague's name wrote back in 2006, right about the future of physiotherapy. Um, and he divided into four different characteristics. The, the kind of, and I'm gonna use the wrong terminology, but essentially the, they are. The whole lists, the ones that try to take on everything. But then ultimately it's very difficult to teach in a curriculum. You've got those who are going to be back to the bodies machine and double down on that. Uh, you've got the people who are gonna go, you know who, sorry, these are the four directions. If you, if you were to go down in one direction, that would be doubling down as the body and the machine go in another direction, you could go whole list and then you lose your identity, go into another direction, and you could basically become kind of a radicalist in a sense, um, uh, where the, uh, you practice in a way that's almost unrecognizable. From a physical therapy perspective. And then there's the other one, which is, uh, I think it's kind of like you are gonna add so much to the curriculum that it's gonna be impossible to teach. Um, and we think there's gonna be an element of where ai, uh, I think a lot of physical therapy practice is gonna be relatively the same, but for not that long. Maybe one or two years after that, you're gonna have various forms of practice, which you're gonna double down. Could be double down is the middle bodies machine. But I think that there will also be the opportunity for physical therapists to practice physical therapy in a broader way. There's an opportunity, and that might mean that we practice in a much more social. Uh, context or it might be that we think social ecologically we might be looking at our own environments, become part of neighborhood therapy teams, be able to look at communities of practice or communities, uh, where we work with third sector parties and, um, work much more closely with vulnerable parties for vulnerable people. Um, and I think that'll be a good thing. Whether or not it'll be funded, that's a different question, but does it open up the possibility for that far more legitimately? Yes. I think as well ai, because of the way that it's gonna be practiced, there will be those that can afford therapy, and there's more of a possibility that therapy becomes a a com a commodity. So, uh, there is the possibility that only the rich will be able to afford physical therapy. That's another way that this could go. So AI will take other majority and if you want to have physical therapy with a human, you're gonna have to pay for it. Um, so I'm not saying that's a good thing. I think that's, um, a possible direction that things could go.

Yeah. And that kind of goes within my next question I was gonna ask you. I like, I mean, that's obviously one of them would be like, we all of a sudden lose access for, you know, and it becomes something the wealthy, you only have access to. Any other concerns you see out there or any other like kind of things you would be concerned about, you know, AI not helping or moving us in a positive direction, or maybe being used in a way that is not maybe what you would consider ideal? Like what are the like maybe, I guess, worst case scenarios you would see? I mean, we don't have to go doomsday here or anything like that. Like, you know, the world will continue hope, you know, obviously I. Maybe we just wanna stay bliss, be aware of like some of the risks that are out there. I'm, I'm hearing things about Mythos Anthropics new model. They won't even release it to the public because they're worried it's too good. It's hacking everything. They're bringing a consortium of all these big companies to say, you need to figure out how not to let this, how to protect your software against it. So there is some understandable concern out there. Yes. With AI and, and that, I mean, maybe that goes beyond this discussion, but I'm wondering. Any concerns you have of maybe the misuse, again, you've just brought up a huge one with maybe a, you know, uh, you know, AI for the masses and true human centered, you know, one-on-one human care for those who can afford it. Anything else that you see that might be a concern as AI is adopted?

I mean, that's a already in. Built in, in some AI models, and the way that it comes out is, is inherent bias. Yeah. So it's been, you know, so it's, it's whatever data is fed into the system becomes amplified potentially. So the way that it treats certain populations may have that inherent bias baked in. Okay. So that's one of the. One of the risks and challenges, I guess the other is, is kind of safety features and red flag in gar, in safety rails with how, uh, that all who, who takes ownership? Who takes responsibility for that? Uh, is it the, um, at the moment if we deploy ai, if. Ambient technology or otherwise, the clinician takes the, takes, the takes the can, holds the can for that the clinician takes responsibility. Uh, in Cambridge, in the uk, um, there is a procurement that's probably coming to a close or if not finalized for musculoskeletal services. They're gonna be taken over by ai, or AI is going to be running some, not all, but some of its musculoskeletal services. So that's, this is already happening. And I guess, um. I think the pilots of that have shown pretty good. Um, we can put a, uh, uh, uh, a, a link in the show notes for that.

Yeah.

I think some of the pilot studies have shown fairly good management of red flags and management of safety. But I think again, you know, I dunno what clinician, how sensitive that is and if it's oversensitive, what clinician burden that has. Uh, if it's not sensitive enough, what then goes on, uh, after the fact and then the legal ramifications, who takes responsibility for not just the deployment, but the ownership and the responsibility of that patient care? Now, within the uk my assumption is that as soon as you accept, as, as soon as your service accepts responsibility, accepts the referral, you accept duty of care for that patient. So therefore it would reside. On the part of the service and not the AI company that's created it. Um, so there's plenty of things to think about in that kind of very proximal aspect of ai. Not even just the kind of more speculative aspects,

anything that's got you excited about ai. I mean, we've talked, I mean, but what's got you like generally excited of like, possibilities with it?'cause I think we all like. To me, I get nerdy with the coding and the tech and their ability to bring, like, you know, clinicians having like, no need to hire a software developer. At least, you know, they can code a lot and do this, um, with some basic software development. And then obviously maybe consult and, and make sure we're, we're doing it all by the book, but anything that's got you. Like genuinely excited about.

I think the way that it can bring together different forms now, I'm not quite sure how it does the analysis of this, but how it can bring, um, different forms of, uh, data, um, might be quite empowering. So, um. The example that I gave in regards to how it handles heart rate, variability, heart rate, resting heart rate, blood pressure biometrics, uh, how you can program in or, or at least be able to work with the architecture of AI to be able to support recovery programs so that, uh, the. AI could become a physiotherapy assistant who works independently with that ai. Um, sorry. The patient works with the ai, like a physical therapy assistant, provides supportive advice, um, enter into, um, and, and enter in like a collaborative, a genuinely collaborative form of care for the patient using AI that can guide by the side and work with the patient over weeks. Uh, and. Alert the patient if they need to contact us for some. Uh. Up to date, some updates maybe for an update, physical examination assessment or, uh, reorientation of the AI with support. So there would be three people involved in that care, sorry, three entities involved in that patient's care, the patient therapist and ai, and all of them working co you know, together. I think that's a really exciting possibility. Again, how we shape the context, um, is gonna be of utmost important. Hence, I return back to this idea of epistemic fluency, the different forms of knowledge that we are trying to work here, and what, what are we trying to orientate this toward? What's our, what's our direction? Totally agree.

I think there's some. Exciting times ahead and, but using due diligence and, and really having these considerations in mind when we're deciding how we're gonna apply it. Because I think, I'm sure there will be those that rush to market and, you know, and that's what we see with AI development and generally everybody wants to get their model front and center and be and win the race and, and safety rails will be damned at the moment. So hopefully. Um, that doesn't, uh, cause any issues, but I, I, I feel, and

that's, that is the issue right now, you know?

Yeah.

Safety.

And I just feel the same with, with, with our adoption of it too. I think we just need to be careful and make sure there is no, you know, harm being done and, and really spot checking. You here? Uh, I just watched an episode of The Pit not too long ago. It's the, the Great show. Um, it's, it's, it's kinda like the modern day er, but, um, it, it had a AI scribe that kind of. Had an error in it and it ended up having some interdisciplinary miscommunication and some older clinician came in storming, you see this ai. And so I mean there's, I think there's definitely issues we need to be aware of, but I think some, again, exciting times ahead is, is fun. Making sure we, again, don't let it lose sight. To me, I think it's exciting opportunities to, to maybe provide some coherence with that knowledge and, and, and maybe, you know, de-bias some of our things. Yes.'cause I think we can de-bias things with ai. What are your thoughts on that? Do you think it can. Help us. Help us maybe limit our human nature of, of bias. And, and you mentioned some good points though.'cause I spoke about this before in, in a talk i, I just did recently. Um, about how a lot of the data that forms and, and trains this as bias data, it's, it's not, it represents, you know, a predominantly western, you know, Caucasian, you know, population. There's women are often underrepresented. There's a lot of issues with that data that may make it, um. Challenging. I'm wondering what you think.

I absolutely agree. I think, um, I think when it comes down to the opportunities for ai, um, that that ability, there is so much potential, but it, it really does come down to. How the AI is trained, it's gonna be, that's gonna be critical. Um, because you, you would want the AI to push back on you as a clinician and say, are you sure this is the right direction? Or maybe with the patient not to be too s ahan, you're doing a great job, you're doing da, da da da dah. That's fine. If it's encouraging and. Situated in the right time, in the right place. There's times where that's not the right thing to say and it's not the right thing to do. So it's trying to get that, again, it comes back to my point around epistemic fluency. The timing, the essis, when you dis, when you deploy knowledge, when you deploy information, when is the right time and how do you do it? Um, but I think it's got the possibility for, as a clinician, for example, to push back at us and ask us. Serious questions. I mean, this should happen in peer review practice anyway, but we know how challenging practice can be and how peer review sometimes doesn't happen as much as we we'd like where people should be observing us and giving us some critical feedback. But let's imagine that you're using AI as a third entity between the, the kind of partnership and then you ask the AI. Do you think this is the right direction? Am I missing anything? Tell me, tell me why this isn't the right thing to do. That's quite an interesting thing to do a, with AI models, because they come up sometimes with some very insightful things. When you ask it to tell you why you, you know, why, why isn't this the right thing to do? And you're like, oh God, I would never have thought of that. So that is something that I think AI can sometimes just. Poke is in the side and say, hold on a second, there is something else that you probably need to consider here. And a and a completely, a blind spot that we would never have necessarily explored, becomes available.

Yeah, I could see that being like just a great clinical companion. Right. And not just a yes man, syncopathic, uh, voice that's just reaffirming all your decisions. One that just pushes back, like you said, uh, and, and rightfully so. And then. I mean,'cause I've even had it with students, like, you know, just, yeah, I didn't even think about that. Like, or just I oversaw overlooked some, some, some things that they rightfully brought up and, and had me, you know, cha change course and things. So I think having that ability and kind of jealous, like if that's what's happening for a younger clinician, I would've loved to have that voice, you know, of like reason and whoa, are you thinking of everything? Uh, I could see how this. You know, being trained as almost, I don't wanna say a mentor, like an ai, you know, I guess maybe, um, with like, uh, but I, I, that becomes a slippery slope. Of

course it does. That's a really interesting idea, isn't it? Yeah.'cause first of all, we've bestowed rationality upon AI there, which is an interesting idea. Um, but essentially it's predictive. And it's trained on, um, uh, uh, it's trained on repeated forms of knowledge. Like it's, uh, gosh, I'm forgetting what the word is. Oh, it's, it's been trained on volumes of repeated data. It's very good at, um, like pattern

recognition.

Pattern recognition, yeah. Yeah. So, so it's gonna be picking up on the things that you haven't seen based on pattern recognition and forms of prediction,

which is quite

rational.

Talk about in clinical reasoning, right? A lot of. Yeah. Pattern recognition.

It's a part of it.

Yeah.

And then here's the other thing, you know, so you've got gut feeling, haven't you? The, the embodied, tacit forms of knowledge that makes you think, actually, look, I know this patient says they haven't lost weight. I know that they've told me that, you know, they don't have, uh, these red flags, but there's something not, right.

Yeah.

And you pick that up. Ask for further investigations and something serious has been picked up. Right? Yeah. And you know, it it, not quite yet. I'm not saying it will never happen, but at the moment that that's that kind of tacit form of knowledge, it's nowhere near.

Yeah. And it just seems like there needs to be so much context that how do you, that you, that you as a clinician. Used to, to form that gut feeling, right? That, you know, even with that pattern recognition and training of all this data, I just don't know.

Well, at that time we might find that we, we can provide biometric kind of scans, biometric data, so that all of those things, the iterative stuff that we would've otherwise seen clinically. The AI may have picked up instrumentally. I guess that's what some of the transhumanists would say. The transhumanists are the people who basically think that technology's gonna save us. They're gonna improve our longevity, their lifespan, and so on. They would say, actually AI is gonna get to a time where we just basically augment the human with various biometric scanning forms of information and data such that you don't need that iterative for, sorry, that, uh, tacit form of knowledge anymore. The data will tell us. We'll just basically we'll have an EC, G or square up. We'll have eeg. It's not a problem. Well, we've already got their pulse heart rate. We've got all their bloods, no problems.

Yeah, I mean this just, I, it, it bends my brain a a bit'cause it's almost that like hyper determinist way of looking at like the way things work in the world. And I just. I just, and, and again, maybe I'm just, you know, the, a determinist, you know, volley going on myself where I just don't know how, you know, and it goes into a lot of questions philosophically, I know, but how we. AI can just, you know, just see how these ping pong balls are all bouncing off each other and predict perfectly. I just think there is, and again, it's my philosophical position that there are just things that are not the, the, the black box up here consciousness. Right. It just, it just doesn't seem to me that that is something that is, and, and again, I'm, maybe I'll be proven wrong and I'm been, it's happened many times before and it'll happen again, but. I, it just doesn't, that kinda just, it just bends my brain to think that AI is gonna be able to step into that and just, it's all just the data, right? We're all data points and this is this hyper determinist world. You know, Polsky's book was another one that was a real Bri, you know, determined, I think was determined. It is a great book and I'd highly recommend it. It definitely had me thinking, which is a good sign of a great book. What do you think? Like,

yeah.

Is it, do you, because in a determinist world, AI just like could do it. Could do it, yeah. Yeah,

yeah,

yeah. Like they just need to know all the variables at play that. And AI has a massive capacity to understand.

Yeah.

A immense amount of variables, more than we can hold in our brains and even when we can hold in some of our research models that we're using.

Yeah, a hundred percent. But the, you know, I, I, I don't hold that philosophical view. Um, I think if we look into physics and we look at quantum mechanics and we look at quantum fields and fluctuations, it's indeterminate. Even at the base levels of what we believe is reality, it's indeterminate. We cannot predict. Um, we can predict with a greater or lesser degree of, uh, predict, we can predict to a greater, lesser degree of um, uh, determinacy, but it's not determinate. It is indeterminate. So as soon as, for example, uh, uh, an observer comes into that situation. It changes the, the game. Um, uh, and you know, there are effects. The whole idea of course cause and effect becomes completely, uh, up the wahoo as well. Yeah. When you go into the kind of basic levels of reality. So do I think that the world is determinant? No. Uh, do I think that, uh, reality is processual? I think it is. It's not the start point and end point, and it's, uh, a line between the two. Um, I think it's, it, uh, I think reality changes and evolves, right? So even, uh, as we sit to this conversation. Everything's in motion. Your bone turnover is happening. Uh, you'll be shedding skin. A what Part of Mark is. Mark yesterday. As is Mark today. Uh, you are going to be physically different and it will have a tendency toward Mark, but it won't be the same Mark. You'll, you'll, as time goes on, you are not gonna stay the same. Nothing stays the same and everything moves and it is indeterminate.

Yeah. And that gets into what we spoke about earlier when we weren't on camera, but is this whole thought of stasis, right? This summit of like, this belief that we, we hit. And that's always been my contention with re with like these. This, this, like precision biologics is another one that I'm like. That you're, you're, you're trying, you're aiming at a moving target. Like the biology is ever changing, like at every moment. It's like you said, bone turning over electrochemically, who knows how much complex stuff's going on in immune and different things are shifting. You offend me. I get all sorts of chemicals going on in my system. You make me ha. I, I just, this, this belief that there's some just like we're gonna, and the physical list, and that was where a soft, I think it feels like we're going like, yeah, we're just gonna need is once we get the data and we can just precisely target it. But that's making some assumptions in my book that, that we can have a static human that that is gonna respond to a. Treatment consistently, and that's where I, I just don't see it Where, where you,

yeah. Yeah. I think Asaf may be, in his defense, might be saying, look, okay, even if you were to say that everything's processual, everything moves, there is so much familiar repeated patterns in the nociceptive apparatus that we'd still be able to target it, even though it's not the same. This familiarity of iteration, which keeps happening from a kind of nociceptive apparatus would be so. Uh, similar even at an individually level, if not a population level, that we'd still be able to target biochemically what is driving a patient's pain and therefore be able to augment it, change it, dissipate it, whatever word, what you wanna use, use to be able to interact with it. And I think, again, know, that's fine, but it's not my, it's not my view. I would say I am quite happy to, but I'm equal. I'm quite happy to have my mind changed. Um, so yeah, I. I think that, uh, I, I agree with you. If we look at, if we look at biological, sorry, if we look at biology, even biology, but I'd say this is true of any, uh, whichever system or sphere that we look at. Um, I. It's process orientated, it's relational, it's historical. So epigenetics, uh, gets involved in this. Um, uh, it, yeah, it's process based relational and it's indeterminate. Those are the key kind of as key, key concepts I think that we've gotta take, take forwards. That doesn't mean that data isn't helpful. We've gotta recognize that these data points just don't, don't tell us an absolute. Answer for everything, for all time. Um, I think we'll be able to collect information. Um, but even how we interpret that information is gonna be somewhat limited. It's not gonna predict everything. It won't,

yeah.

Might tell us a tendency, but that's as far as you can go. I think.

And to me that just becomes like, there, there comes a point where, you know, all the data and work can inform, but in, in, in the end, you have to have, you know, something to a human to, to try to help move somebody in a direction with, and maybe hopefully be informed by as much of the data as as possible. But to me, you know, there needs to be a human in the loop,

I think at the moment, for for sure. It seems that, uh, you know. When it comes, and, and I see this as an opportunity, um, that if we conceptualize, I, I agree. I think we will, certainly for the foreseeable, there will always need to have, uh, degrees of human interaction. But what I hope this. Offers us is to look at health in a radically different way. Such, um, that involves understanding health, not residing just as humans, but in terms of, um, ecosystems thinking, ecologically thinking about climate crisis, the planet, planetary health. Um, because at the end of the day, all living life goes through a cycle. We, we, we, we, we born. We grow, we die, we return to the earth and we give our, um, energy and life sources to other non-human entities. And we consume non-human things for our own life. Cut microbiome, oxygen, and so on, so forth. So I would be really interested in how physical therapy could really be a much more ecosystem driven earth profession.

That brings me to some discussions on your Physiopathology paper. This will be broken up into some episodes because we're, we're going deep.

Yeah.

Um, you had proposed in there in your physiopathology paper that. A pedagogy that orients a student towards ecological composition and experimentation rather than diagnostic protocol. I'm wondering what would that curriculum look like for a, a physio student to maybe orient them that way? Um,'cause that's a tough, they're so, we're so ingrained in the Yeah. The way we train. I love the thought and I agree with you. We need to look beyond. The clinic walls to understand, to really get a good context of a human we're serving. And it could look, you know, social determined ecologically, environmentally planetarily, like you said. Uh, what would that, what do you think that kind of would look like for educational experience? Specifically of

curricular? Yeah. It's gonna, it's gonna look very different. You wouldn't look, um, uh, at human centered. Um, importance. So you wouldn't, for example, focus on the anatomy and physiology of a human. Um, let's not suggest that anatomy and physiology isn't important. Um, but it's to suggest that the i in the context of life and death, uh, say let's, let's call it life cycles. Life cycles require a ecosystem perspective. Um, and that basically what I suggest in the paper is that we have to. Oh, sorry. Have to, it's too strong. Is, is, in order to, to orientate us that way, I think we probably need to decenter the human in that system. So by focusing a lot of models, even when we think of ecological care or ecosystem-based care, you still have the human at the center. So the human at the center, the communities, uh, you know, the environment that sits out outside that and the ecosystem and so on, planet and so on. But the, the human's still in the center. Well, what that does is that still means that any kind of curricular needs to focus back onto the human. And so rather than focusing on human centered health, we'll think, we think of health, what consists of health, what provides life? And what diminishes life and how do those systems interact? And what we're going to be looking at is a theoretical system that's far more relational. In other words, thinking of the connections between entities. Um, we'll be gonna be thinking about, um, uh, the interactions between, between those entities. You're gonna be thinking about life cycles. You're gonna be thinking about. All of the porous barriers that interface between them. You're gonna be thinking about the interaction between human and non-human entities. You're gonna be thinking about oxygen, carbon dioxide, entering gaseous exchange in lungs, which we accept, but we suddenly, when does the human begin and where does the human end? Gut microbiomes, for example. We know that that has an effect on, or we believe that has an effect on immune systems, and also perhaps with the persist, uh, uh, health engagements and relationships with persistent pain, for example. But that's a non-human entity. So in what ways do these human and non-human elements interact? Um, and so, so I think that that's where the curriculum starts. It looks at life cycles. It looks interactions. It looks at the broader. Kind of, uh, relationship between biology, physics, chemistry, um, in a far more broader sense. And then it can focus on the physical whatever we would like to kind of, if we need to narrow it down, then we can of course narrow it down to aspects of physical therapy. And then we start thinking about what does therapy mean? That opens up another conversation. What is therapy?

That that is, and we, we'll have to touch, touch upon that.'cause we're getting into some deep philosophical things. And I mean, and this is why I like these discussions. In that paper you kind of talked about how our boundaries need to be more ame Boyd than fortified. Right. Where they, we may need to be able to, to move our boundaries. Like how do you, you know. See that? Is that something you think, think that's happening kind of right now, or is that something more aspirational in your, in your view or, um, how do you, where, where are you actually seeing those boundaries maybe dissolve in productive ways?

Yeah, it's a really good, it's a good question. Again, I think. It's gotta start from a position that kind of extricate ourselves from our current institutional barriers. I want us to, we may have to speculate to start with, to imagine what might be possible before it becomes actualized. Um. And so one could quite imagine how a physical therapy, as I use an example in the paper, recognizes a health related impact that really impacts their, their community and their society of where they're situated. I use the example of homeless people near, uh, in, in a seaside area. I live near the seaside. There are lots of hu uh, homeless people in Roma area, and they seek, um. They seek, uh, uh, shelter underneath the local peer. So there is something around how physical therapy. What is therapy? Therapy? Well, I, I draw from Greek, so I'm gonna come back to this. So. Physi, what does physiotherapy mean? Physiotherapy comes from the Greek root words of nature and therapy. And so it's not something that's specific to a human. Um, and we can go back to that. It's been an institutionalized name and there's lots of good reasons from that. And we, and we, and we, and as I mentioned before, we are, we, we, we bear the privilege of that, but it also limits us. And I think if we were to imagine. If that was a real problem in our environment, in our community, in what ways could we help those vulnerable people? So for example, if you were residing in, uh, a town with a high level of, um, veterans, um, we use this as an example in some of our conversations where veterans were having a rough deal on how, uh, their, their health outcomes. They have poorer, uh, engagements and access with healthcare. They may have poor outcomes with their, um, uh, interactions with healthcare systems. How can we go out to them and serve them in the way that makes sense to them? Um, and we don't have to, and that's a human based example. Um, but you could use that in environments that give them calm, give them solace. You can go out to nature, for example. Um, and, uh. So that is a form of physical therapy to be able to build relations with people and their interactions with their environment and all of the various things that en enhance their lifecycle in multi, uh, multiplicity of ways. And in the paper I try to talk about how we analyze these aspects of what gives life and what doesn't, um, through longitude and latitude. And so I use. Philosophy from Al Lewis and Felix Ery, who are French philosophers, um, to build up this kind of map. How can we become, for one of a better description, cartographers of care? How can we map out the possibilities in using physical therapy to support. Not just humans, but are co-constructed cohabitating entities that provide support for health in a much more broader scale. And that would also help support our environments, which may also support things like planetary health'cause we depend on the planet for our health.

Yeah. No, I, I like that line of thinking as far as, you know, looking health beyond just our human centered way of looking at health, what is health, and looking at a big picture, maybe that also equals planetary health and a lot of the things, especially when we talk climate change and, and different topics. I'm gonna leave it, I want, I wanna kinda maybe land the plane with the last, uh, question for you.'cause we could probably talk for another few hours and we're already at multiple episodes at this point. Um. You close with the loses line. An animal, a thing is never separable from its relations with the world. If you had to maybe translate that into a a single clinical principle, a physiotherapist could carry into their next encounter.

Yeah, I mean, I think Ulus is already describing something that is an organism which is never separable from the world. Any entity is not separable from the is is inseparable from the world. But let's use the human animal as a, as an example. Is that. The ways in which we interact with everything. It is not just based on our, if we look at the bio-psychosocial model, that's a very anthropocentric model. It focuses on human biology, it focuses on human psychology, and it relates to human sociology. But it's common sense for us to go actually the, uh, we are inseparable from the. Sitting here with the electricity that's going through this house that's powering everything. We are connected to the wind circulation that comes from the fan, from the air conditioning that's drying at our pauses as we speak. And, um, from the energy resources that that is going around us from the, um, uh, distributions of water that come through and outta through the, through the house. Um. All of those elements, just talking about a human environment is inseparable from the world. So the oxygen, the carbon dioxide, the, you know, the, the inhabitants, the non-human inhabitants of my own body, inside of me and externally. So all of those things are in relation to me. Um, they're not human. They're non-human, and if we think of health and life cycles as being inextricably linked to all of those things, then it radically makes opportunities available for us to rethink the physical therapies in ways that can provide therapy outside of our musculoskeletal physiotherapy, ways of practice.

Yeah. Outside those four walls you described were the, the sterile non paintings, um, little plinth sticking out there. It's just kind of that view we all tend to have of a physical therapy clinic. Well, Matt, I just wanna thank you one for, for blessing the studio for the first annual, uh, in-person recording. Hopefully we have more of these. We're gonna hopefully get some guests that will, uh, come into the studio. Hopefully we'll deliver some courses here, both virtually and then, um, in some space here in Arizona. Uh, but thanks for for joining us for the first, first go round.

This has been another episode of The Modern Pain Podcast with Dr. Mark Kargela. Join us next time as we continue our journey to help change the story around pain. For more information on the show, visit modern pain care.com. This podcast is for educational and informational purposes only. It is not a substitute for medical advice or treatment. Please consult a licensed professional for your specific medical needs, changing the story around pain. This is the Modern Pain Podcast.