Dr Andrew Greenland: 0:02

So welcome to Voices in Health and Wellness. This is the podcast that shines a light on the innovators transforming patient care and clinical outcomes through thoughtful technology and strategy. I'm your host, andrew Greenland, and today I'm honoured to speak with Dr Glenn Loomis. Glenn is the founder and CEO of Query Health, an AI-driven platform designed to streamline patient interviews and deliver actionable clinical analytics in real time. With over two decades of leadership in healthcare, ranging from running multi-speciality practices to serving as a chief medical officer, glenn brings a unique blend of hands-on clinical insight and executive experience. Glenn, thank you so much for joining us. It's a real delight to have you on the show this afternoon. It's my pleasure to be here, thank you. So maybe you could start at the top and maybe talk a little bit about your background and what inspired your transition from family medicine into healthcare innovation. It'd be really interesting to hear.

Dr Glenn Loomis: 0:53

Sure, well, I'm a family doc by background, started off in the military and then was in teaching for about 10 years and then the last 20 years of my career was really running medical groups and health systems.

Dr Glenn Loomis: 1:09

And then, about five years ago, decided that I'd sort of burn out on health systems for a while, and technology has always sort of been my side gig, if you will. I've always been the Epic guy or the Cerner guy or you know whatever system we were using, I've always been the Epic guy or the Cerner guy or you know whatever system we were using. So I decided to try some different things. So I joined a startup, worked on an AI scribe for a while before LLM so it wasn't quite ready for primetime yet and then worked for a digital health company for a couple of years running their medical group and that was very insightful. And then decided I really wanted to go back to what I'd always wanted to do, which is sort of build a digital doctor, and when LLMs had sort of come out and it looked like it actually might be possible, I started my own company last fall and I've been working on that ever since.

Dr Andrew Greenland: 2:02

Amazing. So 20 years in clinical medicine. I'm just trying to get an understanding of what kind of made that made that switch for you, because it's quite a. I know you had a big interest running all the way through, but I'm just very curious to know what was the thing that drove you. Was there a particular gap in you spotted and you thought you have to get into this, or was there some other kind of thing that uh helped you to to switch jump?

Dr Glenn Loomis: 2:22

no, that's a great question.

Dr Glenn Loomis: 2:24

I actually I've always been a techie kind of guy, but I was working with the American Medical Association I was chairing their council on long range planning and really got into AI, really started understanding almost 10 years ago now that it was gonna really change healthcare maybe not immediately but certainly in the long run and wrote the first primer on AI for the AMA.

Dr Glenn Loomis: 2:55

And as I did that I really got involved with some other folks. So I started working with IBM on a project looking at what we could do with watson. Could we actually make it sort of think and act like a doctor? And it really wasn't up to the task yet, but that really sort of inspired me to keep up with what was going on in the ai world and as I saw that there were now tools to be able to do what I wanted to do, that inspired me to take the leap and actually go out on my own. If you'd asked me, you know, when I was in medical school or residency, if I thought I'd start a tech company someday, I'd have said you needed your head examined. But here I am.

Dr Andrew Greenland: 3:40

Amazing. So was there a particular gap that you saw in patient care or data that really stood out to you in this sort of point of transition?

Dr Glenn Loomis: 3:48

Yeah, there's really two things that have inspired me. One is in the US, we really have a crisis in terms of our ability to actually get patients in to see providers. Access is a real problem for us and I know sometimes in Great Britain it's a problem as well but we don't have enough primary care providers specifically, but all providers really, and so I was looking for a way. I've been running medical groups for a very long time, and so I was looking for a way. I've been running medical groups for a very long time, and so I was looking for a way to make doctors more efficient, and when you look at what AI can do, a lot of what it can do are things that could offload certain tasks that physicians do and maybe make physicians more efficient. So that was one.

Dr Glenn Loomis: 4:39

And then the second thing is as a chief medical officer and as a president of a medical group over the years, I get to see all the worst things that happen in medicine.

Dr Glenn Loomis: 4:48

Right, I see all the mistakes that get made and people that don't follow the guidelines and things like that, and so, again, one of the things that AI does really well is it follows the rules. It goes out and finds the guideline and says this is what you ought to do, right out and finds the guideline and says this is what you ought to do, right. And so if I could put those two things together and make providers both more efficient and more effective, I think that's a winning strategy, right. And if I could do that in a way that also made patients feel more heard, that's an even better strategy, and so that's really. What I set out to do is create an AI agent that will actually offload different parts of the sort of physician work stream if you will workflow and do that in a way that makes them more efficient and also, hopefully, more effective from a quality and patient safety standpoint.

Dr Andrew Greenland: 5:43

Amazing. So how has your clinical background shaped the way that you approach tech development, because you're probably a rarity coming from medicine into tech compared to people that go straight in and probably 20 years of experience in what you've done clinically surely has shaped the way that you approach this.

Dr Glenn Loomis: 5:59

It's funny on LinkedIn yesterday there was a post about clinician builders and how we approach things differently than perhaps the you know typical 20-something MBA. And so I think we do approach things differently. I like to say, you know, there's a lot of people that have sort of dashed themselves on the rocks of trying to reform healthcare right, or transform healthcare, because healthcare is a monolith and it's highly conservative by nature, and for good reasons, right? I mean, our first maxim is, you know, first do no harm, right? That's what we learned as doctors from the start, and so you know we're very conservative, and so if you don't understand how medicine works from the inside and where the levers of power are, where the levers are that you can push to actually make change, you get very frustrated very quickly.

Dr Glenn Loomis: 6:58

And so I think that's where I am different. I've run big systems, I've been a part of organized medicine. I know where the levers of power are and how we can change things for the better. And I believe that we can change that from the inside and make it better for both patients and for physicians and all providers by doing it sort of within the system rather than trying to revolutionize. You know, create a revolution, I'm not sure revolutions will ever work in healthcare.

Dr Andrew Greenland: 7:34

Okay, and from your vantage point, what major shifts are you seeing right now in healthcare delivery or health tech? Obviously, we're in a very fast moving situation with AI, but what are?

Dr Glenn Loomis: 7:43

you seeing? Yeah, I think the first one is just the use of scribes, the digital scribes that listen and transcribe what you're saying with the patient. I think that was the sort of first use case for AI, and it's done a lot of good for physicians' sort of mental health, if nothing else. A lot of good for physicians' sort of mental health, if nothing else. Unfortunately, if you look at the studies, they would say it doesn't really save time for most providers. If it does, it saves just a small amount, and so we need to actually go beyond that to other tools that are actually going to take out parts of the visit, take out parts of the time that patients are spending with physicians, or physicians are spending on paperwork, et cetera, and be able to actually give physicians back some of the time that they've lost with all of the administrative sort of trivia that we have to do nowadays.

Dr Glenn Loomis: 8:39

Likewise, I feel like patients have gotten squeezed right. We physician visits have gotten shorter and shorter over the years, and because of that, patients oftentimes feel shortchanged, like they don't have time to tell their story, and so, again, if we could use AI AI doesn't care how long you talk to it. You can talk for an hour or 10 minutes or two minutes it doesn't care right and so we can give patients, I think, time to tell their story in a better way, gives us better data as providers and at the same time, then allows the provider to save a large amount of time in the visit because they don't have to do that. Data extraction Data extraction from patients takes us about 50% of our time. About 50% of a visit is taken us trying to pull that out. If we could get that done by AI ahead of time, you could make us much more efficient, and so that's really sort of where I've been focused.

Dr Andrew Greenland: 9:35

Cool, and in terms of the clinicians that are using your tools at the moment, how are they finding them and how are they sort of most benefiting from the AI power that you have behind them?

Dr Glenn Loomis: 9:45

Sure, I mean, we're still sort of in beta with it, and so we've been doing a lot of testing, refining, working a lot on our sort of EQ, if you will, on the bedside manner of it, so that you know, the AI learns to take turns in a proper way and talk to patients in a way that's very reassuring and here's when the patient is frustrated, and those kind of things and so we've been working a lot on that. In terms of the actual just sort of data piece, data piece, I think all the providers that have used it have been really amazed at how good it actually is, even though it's our first generation of it and it'll certainly change and improve as we go along. Out of the box, I'd call it a good, solid A-.

Dr Andrew Greenland: 10:41

Brilliant. Compliance and data privacy is a huge thing around AI. How are you balancing the usability of the tools that you're involved in creating with this whole issue of compliance and data privacy?

Dr Glenn Loomis: 10:54

Yeah, that's a hard one, right? Because this new world, there's a lot of data going in a lot of directions and so trying to make sure that we very much honor the patient's need for privacy. One of the things we've done is actually put all of the patient's data in the patient's hands. So part of our app is a personal health record that they get all of their data, they have access to it. They have control over who is gonna be able to access that data. So if a new physician wants to access their data, we send a two-factor authentication to say, hey, this person is asking to see your data. Do you want them to have access or not? Things like that.

Dr Glenn Loomis: 11:37

So we're trying very much to comply with things like, you know, hipaa and GDPR and make sure that patients feel like their data is protected. At the same time, we need that data to be able to make our application better all the time, to train the application. So we've done a good job, I think, of trying to de-identify the data so that we can use it in the background to make our application work better. But it's a dance for sure, you know trying to make sure that patients feel their data is secure, protected but at the same time, use the data in a way that is effective. Likewise, when we send the data out to the large language models, to the AI, we strip all the identifiers out so that it can't be traced as well. So yeah, very much a dance, but trying very hard to protect patient data, because that's extremely important.

Dr Andrew Greenland: 12:39

It's very important. It's a very sensitive area. I think most clinicians it's one of the things that they fear most about the whole AI thing. Very important, it's a very sensitive area. I think most clinicians is one of the things that they fear most about the whole ai thing. Um, what about the, the tools themselves? I mean, is there a learning curve to these things or is, if you try to make them or your people who are behind them feel as intuitive as possible for busy providers, you just want to basically pick things up and run with them, without getting bogged down in sort of manuals and training and all this sort of thing.

Dr Glenn Loomis: 13:03

Yeah, we're trying to make it as low lift as possible, you know, so that you basically can pick it up as a patient and just use it. We will have a couple of screens that say you know, push here to do this, push here to do that, but it's pretty intuitive On the provider side. What we are attempting to do is make it so that you can basically the patient talks to it. You come in the office, you talk to the patient and then at the end it's all. The whole note is available for you. We'll actually have the scribe piece of it done so that you can just keep going with that same visit that the patient did with the agent. And then, as a provider, the only time you need to touch the EHR will be to actually enter orders and push send. That really is our goal is to get physicians off the keyboard as much as possible.

Dr Andrew Greenland: 14:06

Our goal is to get physicians off the keyboard as much as possible. Cool, and you mentioned the patient experience. Well, what is the patient experience? Do patients have a reluctance to be talking to bots and robots and AI tools? How do they feel about all of these things?

Dr Glenn Loomis: 14:19

It's funny, you know, the data would say that patients actually prefer talking to bots and agents over humans, unless you tell them it's a bot or an agent, in which case then they say that they prefer it less.

Dr Glenn Loomis: 14:38

So it's a. You know, human psyche is an interesting thing. We often are able to fool ourselves into thinking one thing or another, depending on sort of what the setup is, but in general, there's a lot of data that says, for example, that patients will tell a bot much more than they'll tell a psychiatrist, for example, about their mental health, because they feel less judgment, you know, et cetera. And so we're trying to make sure that this is as close of a encounter as they would have with a physician. We're trying to make it as close to that experience as possible while at the same time preserving sort of that no judgment and sort of ease of interaction for the patient, so that they feel like they can say anything, because that's the important thing, right? I mean, as a provider, we need all of the information if we're going to make great decisions, right? And so if things are hidden or forgotten, or just people don't want to say them because they feel bad, that really impedes our ability to make great decisions for the patient and with the patient.

Dr Andrew Greenland: 15:56

And on the other end of the spectrum, the clinician experience. Is all the clinician experience so far positive? Or are people skeptical? Are they worried about their jobs being replaced at some point? Where do the clinicians kind of sit with this?

Dr Glenn Loomis: 16:08

Yeah, I think the clinicians are skeptical at first until they sort of sit and see how it is. It's interesting in our application they can actually see the summary that's generated. They can actually see all of the back and forth that the patient and the agent has. So they can see the whole transcript if they want to read it and they can go back and forth between the two so they can see oh, this is where this came from. And, yeah, that patient did say that we also are in the process of instituting that the patient will sign off on the summary before it goes to the provider, and I think that's also gonna help help providers really feel a sense of of calm that yes, this is what the patient was trying to convey to me, because they signed off on it at the end.

Dr Glenn Loomis: 17:00

Um, but there is a skepticism there. So far, most, most physicians don't feel like it's going to replace them, um, but I think as they learn more and more about what it will be able to do, what ai is going to be able to do, there's going to be some of that, because it can do a lot of the functions that we have normally done. Personally, I think at least in the US we have such a, and actually around the world, and especially in sort of developing countries, there's such a paucity of physician access that I don't I'm not too worried about anybody losing their job right now at least you know, I think for the foreseeable future there's, there's plenty of need to go around.

Dr Andrew Greenland: 17:51

Cool, I think I might know the answer to this, but I'll ask anyway. I mean, do you think AI is being overhyped at the moment or underused, particularly in healthcare?

Dr Glenn Loomis: 18:01

I think the answer to that is yes. I think there is an overhype going on a little bit, that you know. Everybody is saying that you know, well, my thing has AI, you know, you know. So, like every everybody is trying to jump on the AI hype train. When you really get under the hood, there's very few sort of truly native AI applications, and so, yes, there's overhype. I think in medicine, though, or in health care, it's the opposite. I think we, as we always are, we have been slow to sort of get on the train, other than a few small things, and so I think, in healthcare, we actually are behind where a number of other industries are, and I think you're going to see that sort of explode over the next two years. So, yeah, yes to overhype, but yes to underhype, I guess. Okay.

Dr Andrew Greenland: 19:01

Following on from that one, what do you think is the biggest misconception that you hear about? Ai in patient care?

Dr Glenn Loomis: 19:08

Oh, that's a great question. I think that probably the biggest misconception is that there's a lot of that. It makes a lot of errors, that it makes up stuff. Yes, llms do make up things, but you can control for that right. So we have, for example, if we're looking at facts, we have an agent that checks the agent right and says you know, does that guideline actually exist and did they actually look at it correctly? So, like you can really reduce the hallucinations that the LLMs have? And even so, it's still at a very small rate anyway.

Dr Glenn Loomis: 19:59

And so if you look at the amount of errors that happen in healthcare by humans, the question is going to come down to one of and I don't know the answer to this, I'm just going to state it the question is going to be which is worse a human that didn't use AI and therefore didn't use the right guideline and you didn't get the right care, or an AI that most of the time gets it right but on this particular occasion, made an error and you got the wrong care? Like which is worse? Right? Well, I'd argue they're both bad, but I can show you many papers that would say that, on average, the AI gets it right more often than the human, and so like is it most important that we get it right most often or is it most important that we never allow AI to make an error? Don't know the answer to that. I'm sure the courts are going to weigh in on that. I'm sure you know the regulators are going to weigh in on that. But for me, I would tell you that I can see already that we are going to be able to use AI to give physician the tools so they don't forget things. They don't make an error because they just forgot about that guideline or that diagnosis that wasn't at the top of their list, it was down two or three, and they just forgot to even consider it, or they forgot to work it up, or they forgot to add that one lab test right, and that is where I think AI is going to make a huge difference from a quality perspective.

Dr Glenn Loomis: 21:34

But that fundamental question does exist Is it right to be most often right or is it most important to never be wrong?

Dr Andrew Greenland: 21:45

Yeah, it's a good way of putting it, and I suppose the question then becomes who takes responsibility for the error made by AI?

Dr Glenn Loomis: 21:52

Yes, and I think at the moment it's still our license as the provider right, it's still our license on the line. So you know, you always have to be diligent and make sure that you're you know that you're watching over it and that you're you know we still have humans in the loop, we still have physicians in the loop for the reason right. We need that check and balance and I think if we use that sort of pairing correctly, where you've got AI sort of surfacing, you know good differentials, good treatment plans, but maybe with the occasional error, and then you've got physicians checking that very quickly, you can do it very quickly. It doesn't take you very long to read that stuff and saying, oh well, I'll pick that one because this doesn't make any sense today, you know. Then I think we get the best of both worlds right when we get it right most of the time and we don't have those errors. But I think eventually you'll see AI going direct to patients, as you know, in a sort of basic primary care way at the very least.

Dr Glenn Loomis: 23:01

But I think that's a few years off. You know regulations, never. It takes them a long time to catch up to where the, where the tools actually are. So you know, I I don't like to predict, you know that they I don't remember who said it, but somebody once said you know, we always overestimate what technology is going to do in the short run and underestimate what it's going to do in the long run. And I think that's where we are on AI, for sure.

Dr Andrew Greenland: 23:26

Cool. So what are you seeing in terms of clinical operational outcomes so far with the tools you're implementing? Obviously, you must be looking at the outcomes at the other side. We know these things can do wonderful things. What are you seeing in terms of outcomes?

Dr Glenn Loomis: 23:39

Yeah, I think in terms of outcomes, what we're seeing is that it can decrease the time that providers spend with patients without making patients feel bad about that, that actually the patients feel as good or better. Still preliminary in our data haven't published anything yet, or really, but that's really. That's what we believe at the moment is it will reduce the time spent significantly and, at the same time, allow patients to actually feel even more heard than they feel now. We haven't gathered a lot of quality outcome data yet because we're not. That's sort of in our next release is this is the clinical decision support piece of this, and so that part. I can only tell you what's in the literature. I can't tell you what we're seeing yet.

Dr Andrew Greenland: 24:29

Okay, and what metrics do your clients use to track the success of this?

Dr Glenn Loomis: 24:36

Yeah, so obviously we have to look at financial metrics, because you know, otherwise I'll be out of business. We are looking at patient satisfaction metrics and seeing you know whether patients like this or don't like this. We're looking at physician satisfaction metrics as well, and then we're looking at sort of time spent metrics. In the US measuring productivity metrics like number of visits or number of relative value units that we use here. Those are the big ones. We also look just at sort of how much time does it take for the AI to do a visit with a patient? Time does it take for the AI to do a visit with a patient? How many back and forth does it take on average for that conversation to happen? Again, there's some data out of Stanford on that from before and we're trying to validate that that actually holds true. Yeah, those are the sort of big things that we're looking at. We have a list of metrics that's, you know, as long as my arm, but those are the sort of the big ones that I care about the most.

Dr Andrew Greenland: 25:44

I mean, are you seeing anything around patient retention or staff efficiency at all?

Dr Glenn Loomis: 25:49

I have not looked at that yet because we're not quite far enough down the road, but that will definitely be something that we will look at. Yeah, good points.

Dr Andrew Greenland: 25:59

And what's next for Query Health? Have you got any upcoming releases or strategic goals in your company?

Dr Glenn Loomis: 26:05

We do. So our next release for the fall will be adding the sort of scribe portion for the in-person part of the visit. We also are adding the clinical decision support in terms of differential diagnosis although we actually have that now, but adding to that the treatment planning portion, so allowing the physician to say yes to this diagnosis and then it spits out here's a proposed treatment plan. And then our winter release we'll actually be looking to begin to allow patients to talk directly to this, to ask it questions like what you know, what does this term mean? Or my doctor says I have congestive heart failure. What's congestive heart failure? And have it explained? So again, trying to take sort of the best of webmd and you know, uh, and and highly tailor that to a patient and their actual medical record and their actual experience, rather than just, uh, sort of at a very high level, um, so yeah, that's kind of where we're headed over the next uh, six to 12 months.

Dr Andrew Greenland: 27:17

Nice, and are you looking to partner with anybody or serve more deeply over the next year in your development?

Dr Glenn Loomis: 27:25

We are. So we've had a couple of insurance companies reach out to us. We've had a couple of actually countries reach out to us about potentially taking this internationally, and it's interesting it may well be that our tool can leapfrog in certain countries sort of where we've all been in, the sort of developed world, if you will. Because you know, if you take some of the countries in Latin America or Africa, they don't really have good medical records or medical record keeping, but almost everybody has a phone. So if I can give them an agent that can take their history, put it all in the phone and they can always have that with them, then suddenly they can go from provider to provider or when they go into an emergency room and they can actually have their record with them in a way that they just don't now and so and we can do that very cheaply Actually looks like we can do that for less than $5 a year per patient, and so that starts to look pretty appealing in certain countries where everybody has a phone but nobody has a medical record.

Dr Glenn Loomis: 28:44

So we're looking at possible options there and seeing whether that's going to be something that happens. But I'm excited about the possibilities of serving others in a larger way, outside of just, you know, serving the physician community that I live to serve here in the US. But I have a heart for serving patients in places where they don't have access, and it may be that, you know, that's the first place where we'll actually be able to use the sort of direct patient primary care, because they just don't have access to anything. So perhaps access to an AI agent is better than access to nothing, and so we'll see. I'm excited about the possibilities, though I like to dream big.

Dr Andrew Greenland: 29:33

Amazing. What about wearables? Do you see wearables being integrated into the work that you do?

Dr Glenn Loomis: 29:39

Absolutely so. Our goal is to integrate Apple Health and Android into our fall release and certainly by the winter release, when we're going to integrate you know as many wearables as possible, because, again, there's a lot of data there that really could inform the provider's decision making. You know, we, our patients, are using these things. You know, I mean, I have an aura ring that I wear every day that captures yeah Right, it captures everything, right and so, but my providers don't know anything about that.

Dr Glenn Loomis: 30:22

You know that my sleep was bad this week, or you know that my you know heart rate is, you know, always fast or you know whatever Like. And so I think there's a way to capture that, summarize it and give it to the providers in a way that actually is useful, because right now, as a provider, I don't know what to do with that data, because they give me some data point here, data point there. But if I had all of it and I knew it was summarized in an appropriate way, I think I could start to use that for the good of my patients. Cool.

Dr Andrew Greenland: 30:59

Biggest challenge, biggest bottleneck in doing what you do.

Dr Glenn Loomis: 31:03

The biggest challenge by far has been LLMs are an amazing tool, all right, and they really are great at doing differential diagnosis and doing treatment planning and actually the scribing part isn't that hard because you've got both people talking.

Dr Glenn Loomis: 31:21

But teaching, building an agent that actually thinks like a physician, takes the history in the appropriate way, does it in a standardized way, doing it for all the different visit types that we all have, right, that has been an amazing challenge, but we have. We have done it and we've we've made it work and it's working well now, and so I think that's a a big moat for us, because it turned out to be way more difficult than we thought it was going to be. Number one and number two one of the things that we learned is engineers and coders are important in this process, but actually having a big group of clinicians to actually help us create the prompts, test the prompts, make sure this thing actually works the way we think it's going to work, and then continually recycle that that has been extremely important, because coders and engineers are never going to be able to write the prompts that you need. Uh, I think for any agent actually I think cysts, you know subject matter experts are the key to creating great agents great.

Dr Andrew Greenland: 32:33

Final question what advice would you give someone launching a healthcare startup today, based on your experience, have a really rich grandparent or something.

Dr Glenn Loomis: 32:43

I think that would be my biggest advice. Raising money is the hardest part, but I would just say, if you have something that you're passionate about and that you love, take a chance, do it, because it's important that we get technology into the hands of patients, into the hands of providers that's actually going to make a difference and not going to slow us down the way some technology has in the past, that we need technology that's actually gonna make our lives better and the lives of our patients better.

Dr Andrew Greenland: 33:23

Glenn, thank you so much for your time this afternoon. It's been a really illuminating conversation. Ai is so huge and topical at the moment. It's really interesting to hear your insights in it from a medical perspective and also from a tech perspective. So thank you so much for spending your time and answering and having a conversation this afternoon. Really appreciate it.

Dr Glenn Loomis: 33:40

Andrew, the pleasure has been mine and you're a great host and easy to talk to. Thank you.