Become a Creator today!Start creating today - Share your story with the world!
Start for free
00:00:00
00:00:01
Can AI Fit into Clinical Decision-Making? | $350m Regard Chief Medical Officer Dr. Francisco Alvarez
20 Plays5 days ago
In this episode, we sit down with Dr. Francisco Alvarez, a doctor and Chief Medical Officer at Regard, to explore a fresh perspective on how AI is actually being used inside real hospital workflows.
We start by breaking down what it really feels like to practice medicine today: navigating fragmented systems and overwhelming amounts of patient data. We then dive into how Regard is building AI tools to surface insights, streamline documentation, and reduce the cognitive burden on clinicians. We also discuss why physicians aren’t looking to be replaced, but instead want better tools to support their decision-making. Finally, we explore the deeper structural challenges in healthcare, from interoperability to misaligned incentives, and what it will take for AI to truly improve patient care.
Recommended
Transcript
Introduction to Healthcare Innovation Podcast
00:00:00
Speaker
Welcome to the Healthcare Theory Podcast. I'm your host, Nikhil Reddy, and every week we interview the entrepreneurs and thought leaders behind the future of healthcare care to see what's gone wrong with our system and how we can fix it.
Bringing AI into Hospitals with Dr. Alvarez
00:00:15
Speaker
Today's guest is Dr. Francisco Alvarez, a physician and the chief medical officer at Regard, an AI-powered platform working to bring AI into hospitals to improve the clinician standard of care.
00:00:26
Speaker
And this is a really interesting one. Dr. Alvarez sits at a really unique point in healthcare. He's both a practicing clinician and also someone building a system that sits inside these hospitals.
00:00:37
Speaker
And in this episode, we go beyond the typical AI and healthcare care conversation, which you often hear almost as a buzzword these days, and talk about what it actually feels like to practice medicine today.
00:00:49
Speaker
There's a lot of fragmentation, so much data to deal with, and there's also so much tension between administrators and clinicians, which means that the biggest problems and barriers in healthcare aren't really technical, but more structural, which I think this provides a really fresh perspective on how can we embed AI in clinicians and hospital workflows better.
Transition from Clinician to CMO
00:01:11
Speaker
So hi, Dr. Alvarez. Thank you so much for coming on and welcome to The Healthcare Theory. Thanks, Mikhail. Really excited to be here. Yeah, of course. and before we get into regard, I want to get into your background.
00:01:22
Speaker
You've been both a physician, then eventually working in clinical leadership, and now a working at a startup, which is kind of an interesting path. I'd love to hear, was it a more gradual transition? Did you see yourself always working in entrepreneurship? Or what made you want to pursue both sides of healthcare?
00:01:41
Speaker
Absolutely. um I've had a bit of ah of a meandering path ah through healthcare, care as I think many people in in similar roles to myself have in the past as well. um I'm a a physician by ah by background and and training. I did my ah my medical training in internal medicine and started my clinical career working as a
Improving Patient Care with Technology
00:02:03
Speaker
hospitalist.
00:02:03
Speaker
um early on i i was very interested in the policy side of things right like how does um how do uh policy and payment systems interface with the clinical care that we uh we deliver every day i spent some time ah working as a fellow at uh center for medicare for medicare and medicaid services ems um i also spent some time working with the uh the baltimore uh sorry the maryland department of health um as well as uh working with the WHO for a little bit of time. And really what ah what I came away from both with my ah clinical experience working with patients day-to- day to day, I was taking a ah broader view was ah a how much potential there was to really change ah the care that we're delivering to patients, the quality of care, ah the ah
00:02:59
Speaker
the the accuracy in which we were able to make diagnoses with technologies that are available today. Right. These aren't pie in the sky ideas um or or new technologies that are down the road, but it's leveraging technologies that just haven't been ah fully implemented across the health care system for various reasons that sort of you can get into um and um and using that to ah to deliver better care and and um and kind of fulfill our promise as physicians. So that kind of started me down the path of exploring that interface between ah health, health care technology and the health care system. Started working with a variety of startups um working in the in the Bay Area and ah eventually i got connected with regard.
Leveraging EMR Data for Better Care
00:03:51
Speaker
which is a company that I'm working with today, and was really inspired by their vision for um ah leveraging the the data that we have in the in the EMR to and leveraging various you know multimodal technologies. And certainly this has been uh kind of spurred forward this is pre sorry let me back up here this was before the llm explosion before the kind of current wave in in ai but already back then ah we were thinking about how can we leverage these technologies to uh better sift through these uh this data and use this data more intelligently at the point of care um and um
00:04:33
Speaker
A, that you know certainly improves documentation, but it also improves quality the the care that we provide to patients. So I started working with them, and I currently serve as Chief Medical Information Officer at Regard while continuing my clinical practice on this side.
00:04:47
Speaker
I really appreciate that background. I think there's, of course, a kind of a meandering journey there a little bit, which is really exciting. You start with the policy side, obviously the healthcare clinicians have stays throughout, then you go to technology. And in some ways we're seeing a little bit of overlap between the two, but I want to start off with your role as
Challenges in Hospital Operations and Data Management
00:05:03
Speaker
a clinician. I think that when think about AI, for example, there's quite clear applications in other industries, but in healthcare, there's so many of them and they kind of muddied together. It's a little bit confusing, but I think it's helpful to ground into like, what are the actual workflows you're doing as a clinician? Can you walk me through like, what is the day to day for a doctor like yourself? And what were the pain points throughout that problem or process where you thought maybe this could be automated a little bit? Maybe that's from a policy or technology side needs some improvement.
00:05:32
Speaker
ah it's It's a huge question, actually. And I can i can only speak for for myself as a hospitalist. Certainly, I've had a little bit of insight into workflows for other um for other clinicians and other specialties. But every specialty is going to be a little bit different. um But certainly, ah kind of the day-to-day and...
00:05:52
Speaker
my job as ah as a hospitalist, ah you can break it down into a few different workflows. And depending on the the structure of the health system, the particular hospital, the particular group, you know physician group that you're working with, some of these might be combined and into different workflows in different ways, depending on scheduling and and staffing. ah But the the the first is ED triage and admission. And to the extent that hospitals participate in ED triage depends on the health system. But in some of my past roles, we actually had a very close working relationship between the hospital medicine division
00:06:30
Speaker
and the ED triage and the emergency department. And so it was a very collaborative atmosphere where we worked with them to ah to kind of triage patients patients within the hospital and then doing the admissions, right And so that's everything from ah understanding why the patient is is here, evaluating the patient, talking with them, writing that H&P, doing that med rec. Medication reconciliation is such an essential part of what we do um and is often ah you know overlooked. And you know this is an area where better information systems would go a long way. Right. There's no reason why we have to ah depend on
00:07:10
Speaker
an individual patient when they are super sick and and not necessarily able to provide the best history or or the best account of what um of their entire medical history to force them to remember exactly what medications to take when um or you know more often than not. myself or or my colleagues going through taking these deep dives in the chart to try to piece together ah what medications they are on, what medications they were on, what medications they should be on. ah Because all this often, unless ah unless it's a very unified health system that um that keeps its records in one place and keeps up-to-date medication lists, oftentimes it's ah this information is fragmented across different EMRs, different health systems, et cetera.
00:07:57
Speaker
um And so that's, you know, that's certainly one ah one application. but Just kind of a little bit of an anecdote here. I always remember was to think back to my medical training where one of one of the senior residents somewhat kind of, I think, berating us about something that we did, something or another that we did ah would say, ah you know, ah Surgeons do surgery. That's that's what they do.
00:08:23
Speaker
We're medicine doctors. It is literally our job to know what medications the patient is on, what medications they should be on. So that's an integral part of what we do day to day, whether we recognize it or not. And it's somewhat of a blindside, I think, in the current system.
00:08:41
Speaker
But anyway, moving past the admission, kind of putting in the initial orders, getting them set, making sure that they kind of stabilize and have and have a vision for what that hospitalization is going to look like.
00:08:55
Speaker
Then we transition into the day to day rounding. Right. this is a bit of a different workflow. um Oftentimes you have a list of patients that you see every day. Usually you know there'll be some patients that you've discharged. And so you'll have some patients that are new admissions that you're taking on, either that you did yourself or that someone did they're passing on to you. um And you start off in the day pre-rounding. And so that's essentially reviewing all the data that's come in, any consults, any labs, um you know information that has happened since the last time you basically checked in on the patient. um making your plan, putting in any you know following up with any ah any consult services, you then go and see your patients. Right. Examine them, ah see what's changed, see any changes that you need to make to the plan. um
00:09:44
Speaker
And then afterwards, you kind of go back to your finish your documentation, finish any orders, update, follow up with consultants. ah Some places, if you are if you're especially efficient, will do um will do kind of a ah second set of rounding. But here again, there's ah tons of ah of opportunities for optimization. This kind of manual data review, i think every clinician needs to be able to, needs to review all the data, right? Like they need to know what's happening with their patients. um But it can be a little bit of a ah find the needle in the haystack or make sure you know making sure that you saw
00:10:23
Speaker
XYZ medication that was changed or or or or some order that might have been put in by a consultant, etc. It can be a little bit of ah of a wild goose hunt and that that can be optimized. Similarly,
00:10:36
Speaker
Any physician will tell you, at least any internets will tell you how much time they spend um writing notes right and in the and the documentation part. And obviously that's a huge um focus of the industry right now by the healthcare technology industry and the ai and healthcare care industry. ah But lots of lots of improvement that can be made both in and in in improving the efficiency as well as the accuracy of of the clinical documentation that we produce day to day. And then lastly, ah discharging the patient. Usually this happens ideally earlier in the day. So usually this is part of your morning workflow by identifying which patients need to be, which patients are ready to leave the hospital. Usually you've been um checking in during like a multidisciplinary rounds where you're checking in with the case management, social work day to day to see
00:11:29
Speaker
you know, what ah what needs to happen to get the patient out of the hospital. One thing that we have in hospital medicine is this both starts on day one, right? Like the first day that the patient's in the hospital. ah Your job is to already be thinking what ah what barriers there might be for them being able to leave the hospital and start addressing them immediately. um And so make sure, you know, everything like do they have if they if they're on home oxygen, do they have home oxygen? Do they need um any kind of ah ah durable medical equipment delivered? they're if they need ah if they can't go straight home after the hospital, do we have ah what's called a placement for them? Do they they i go to subacute rehab facility or or long term?
00:12:10
Speaker
ah nursing facility, things like that, you know, are things that hopefully you've worked ah ah worked out over the the course of the hospitalization, but then kind of that final push to actually get them out of the hospital um it often requires a lot of work to make sure that they don't bounce back, right? You're kind of ah sending them out on the right foot so that they can kind of continue their healthcare care journey.
AI: Threat or Opportunity for Doctors?
00:12:33
Speaker
Wow. There's, I'm starting to realize that question was a lot bigger than I probably had planned for. That makes a lot of sense. I can imagine that when you're a doctor, it's not just one task getting through throughout the day, like you would be if you're in finance or marketing.
00:12:45
Speaker
There's so many different things that are nuanced. Each task we have has different variables. um So just that sounds like very chaotic, which I imagine a lot of doctors like yourself had had been going through. But it seems like there's a couple of strands that are there throughout. i mean, the data side is really interesting.
00:13:00
Speaker
We have this talk in the industry about you want more data, more data, and more data. But then what do you do with all that data when you're just a doctor who has 10 minutes of their time on their hand? Are you going to really it just adds a lot of noise, really. And I think that's a little confusing. And also documentation part, something that really is a little bit can be commoditized and maybe should be to some extent.
00:13:21
Speaker
But i was curious from your perspective, you said this before, a lot of people think in other industries, they're like, they don't want AI coming into their space. They don't want to be replaced by AI. And they're scared, right? Because I think a lot of employees don't have the best moat. I think they're a little insecure about what value they provide. But interestingly, you said that doctors want more ai And I mean, I was wondering if you still agree with that and to what extent you've heard that and um um What do you is explain what you know about that as a doctor? I mean, what role does AI play in your opinion and also just to the conversations with other doctors? What are you seeing today?
00:13:56
Speaker
Absolutely. I can't I certainly can't speak for every physician out there. um You know, I think the ah opinions will run the gamut. um You know, there's certainly people who are very tech forward and and people who are very much not so. But I do think ah my personal opinion is that ah physicians are in are one of the last to be replaced, right? if they If they're able to fully replace physicians, they've also already replaced the accountants and the lawyers and the bankers and and and all of these other um ah professions. you know There's so much that we do
00:14:32
Speaker
um in terms of taking care of patients, the healing process, um physical exam. ah You know, obviously I'm not a surgeon, but, just you know, surgical specialties ah that are not easily replicable by a computer or an AI system. So, you know, I think the over the, if you look at the history of the profession of medicine, um it has changed dramatically, right? Like the the work that we do, ah the data, you know, and I'm talking about this admitting and rounding and discharging workflows um looks very different than, you know, what a physician
00:15:12
Speaker
100 200 years ago might have done but what what is true is that at every stage the practice of medicine has benefited whenever we have taken the latest technology taken the latest tools and developed it to advance our field and i think that that's not that there's no reason to think that this is any different in my opinion i think we need to really as a profession think about ah this is an incredible opportunity and how can we do this to enhance the profession and to I think really as ah really key to this is influence on to make sure that these technologies are used for the ah for the betterment of the patient, right? As opposed to
00:15:53
Speaker
other stakeholders or or other processes. I think that's kind of as ah as a guardian of the healthcare system, I think that that's very important for for the medical profession. um But yeah, to to kind of go to where you were alluding, it's been really enlightening to me and and and honestly very refreshing when, um you know, with the work that I do at Regard and launching our software across dozens of health systems and going on site, meeting our users, shadowing them. And more often than not, the the response I get is not one of being afraid or being
00:16:28
Speaker
ah thinking that AI is going to take our company, they want to take our jobs. ah They want more. They're like, why why can't it do this for me? Why can't it do that for me? And sometimes it's almost fantastical where I feel like we as as doctors a little bit um downplay or or or underappreciate the amount of training and education that has gone into our own formation as physicians, right? um Where they want the AI to make decisions where as a physician, I'm like,
00:16:59
Speaker
you could probably make a more informed decision on that particular topic, given your clinical background, expertise, your clinical gestalt, compared to just asking a
Enhancing Clinical Decisions with Regard's Technology
00:17:09
Speaker
computer.
00:17:09
Speaker
A computer will probably give you an answer, but I'm not sure for certain questions whether it'll give you a better answer than than a physician can come up with. So obviously this depends on the particular situation, but so it's so fascinating to me how many of the of the doctors i I talked to wanted to do even more than we than I had even imagined would be possible.
00:17:29
Speaker
Yeah, that makes it so interesting. I think there's some sort of probably correlation between how much people want AI with how defensible they think their job is. For example, I don't think you're very worried about being replaced with chatGPT the next day. i imagine you're pretty um you're feeling good about that in regard. But imagine someone in finance, I think, where things, workflows can be a little bit more commoditized, standardized,
00:17:51
Speaker
is a little bit of more of that fear. But I think let's take the premise that healthcare, care especially doctors, they provide unique value that AI cannot provide. They're human. They have a lot of intuition, years of medical tool and training, and that's something that an LLM cannot really pick up on instantly.
00:18:06
Speaker
So I think we've kind of hinted at regard a little bit, but now i think it's a good time to mention, like, where does regard come into play? We have some demand for AI, and I think there's dozens or hundreds of companies now building AI for healthcare, each have a little bit of an angle, but You made a big step. You're a clinician.
00:18:22
Speaker
You had a pretty big role as a clinical leader. and you're taking and now you're joining a large company. So I imagine you there's something you believe in there, right? And so i'd love to hear what is regard? How would you explain to someone who's not familiar?
00:18:34
Speaker
And what makes it unique to you? And why do you think they're uniquely positioned to solve this AI utilization problem or implementation problem now? Yeah.
00:18:44
Speaker
We've talked so far a lot about the the data issue, right? We talked about how there's so much data in healthcare, there's so much data that a clinician needs to go through a day to day. The fact of the matter is health systems are sitting on troves of patient data, right? And as clinicians, we...
00:19:03
Speaker
ah we we're trained to be able to quickly sift through the data, to understand what's the most important thing, the most important data points that we need in order to make the next decision, the next issue to resolve whatever ah whatever we're trying to do to move forward to patient's plan of care.
00:19:21
Speaker
But we're not... We're not individually going through this entire ah the entire data that the health system has on that patient. So I think there's certainly a role for having a a system, an agent that is able to ah go through all that information and bring forward insights or.
00:19:44
Speaker
or data points at the point of care, things that might influence the clinical decisions that you're making day to day, things that might drive ah better outcomes or better quality of care for the patient. um And so that's really what regard is all about is really, in a way, kind of data mining the patient's electronic medical record ah to bring forward these things at the point of care. And the way that that kind of takes takes shape, at least today, is in the form of ah pre-writing a note for the clinicians, such as you know ah a progress note, an H&P, a discharge summary. and But that's
00:20:21
Speaker
just the beginning, right? Like we, you know, you're also able with regard, you're able to query the the medical record using natural language. And so this makes ah doing more complex review and analysis, when we call it chart biopsy ah much quicker. I think, you know, one of the issues with ah with data and healthcare care is that data is is messy, right? um There's Healthcare care data is especially messy. There's so so so many things are different from health system to health system. Within individual health system, the way that data might be organized, depending on whether a lab was performed in this lab or that lab, or whether a procedure was done you know in the hospital or in their ambulatory surgical ah ah surgical suite. All these in impact the way that data is recorded, stored. um And you know i often say,
00:21:16
Speaker
if people who are um people who are maybe not in the trenches of in healthcare, they get very excited about you know the latest models or the latest, you know what what can we do with the AI? But really 90% of the work is is preparing the data and and cleaning that and understanding that and standardizing the data. you know I often say, if you if you could just hook up a ah patient's medical record into an LLM and let it rip,
00:21:46
Speaker
probably your outputs are not going to be that great ah because there's just so much context around
Standardizing Healthcare Data
00:21:52
Speaker
a lot this data. And again, you know, ah ah messiness around the data. And so I think that that's where regard really shines in terms of having a deep understanding of the data and optimizing it to get to get high quality outputs for clinical care.
00:22:11
Speaker
Yeah, it seems like there's a couple of things there. I think there needs to be a clear orchestration layer between different the disparate systems that we have in health care. But it's almost like we haven't gotten there for what I think to be a couple reasons. One, I think everyone wants to own their own data.
00:22:24
Speaker
They don't like sharing the data to the kind of gatekeep it, which makes it hard to standardize. And another level, it's just incredibly hard to standardize and share that data, get it all a row together. It's almost like who has the incentive? People might have some incentive to do that, but for the amount of work it takes, like the cost benefit analysis wouldn't really be there. So it's a pretty difficult problem. But as regards continuing to build this out and pairre not only like use the data better to create better insights, what does that actually result in for the actual physician and clinician? How does that change their workflows?
00:22:55
Speaker
We went pretty detailed into like what your workflow was like and just the overall, generally what it's like to be a hospitalist. But when you have a solution work at regard, where does that change things? I think there's two... Two points to that I'd like to touch on.
00:23:07
Speaker
The first is you mentioned the difficulty of standardization of data or interoperability. The interoperability problem is one that I've been thinking of about a long time before I started at Regard. And it's difficult because it's it's often more of a social problem than a social or political problem than technical one. Right. You kind of alluded to, you know, ah misaligned incentives from health systems and payers, other actors in terms of. safeguarding their data for many for for many different reasons, you know some of them ah justified. And you know certainly there's been a lot that has ah that has been done on this front in the policy realm.
00:23:50
Speaker
ah When you think about the ACA, the HITECH Act, ah be ah as part of meaningful use, making sure that ah that EMRs have some basic interoperability standards um that has really opened up the um the gates for, I think, a lot of innovation and a bit of a side note here. I think it would be a shame if if these gates were to then just be closed off and become walled gardens. I think that'd be a shame for ah for innovation in health care. But, um, but even once you kind of, as we were discussing, even once you kind of opened the gates, it, you still have the problem with the data. I think what we've realized that regard is that you can't put the onus on the health system or the patient, um, or anyone else to pay or any other actor, uh, to, to clean that data for you. You, we kind of take it on ourselves to, uh, to process that data and and standardize it. And you know they ah they willingly give us the data because they want what REGARD is able to offer their clinicians. But if you expect that that the health systems are going to do the work um the dirty work of of data cleaning, you're you're sadly ah very mistaken.
00:25:06
Speaker
And then kind of flipping over to what that looks like on the clinician end, i think it's a similar philosophy where the other lesson is, you know, change management is hard. Right. I think everyone will tell you that um you you need to be ah embedded into workflows. Otherwise, clinicians are not going to use ah your software, your product, um you need to provide value to them. um Otherwise, they're not going to use if they can, um if they have any reason to not use it, they will not use it, right? They only use things that are provide value to them or the patient um in terms of, you know, workflow efficiencies or better care. Like these are the types of things that that that really make a difference for doctors and for patients. And so It's integrating seamlessly into the workflow ah almost you know in a way that you might not even be noticed. right like You obviously want them to know that they're using Regard and you know we we're being completely transparent. um
00:26:05
Speaker
But it wants to be so seamless that it they don't even have to think about, oh, I need to now um exit and go into this separate regard tab or regard window. That that will completely defeat the purpose, right? So the closer the closer integration into ah into existing software, into existing workflows, into existing processes, ah that is absolutely necessary to have a ah to have ah successful product.
Aligning Incentives with AI Integration
00:26:36
Speaker
Yeah. And I would be curious, I think if you want to i ideally have like regard being a silent and partner, a silent coworker working behind the scenes, that's kind of the goal there. But it's almost it's quite hard to, I guess, match or mirror the workflows of the clinician.
00:26:51
Speaker
And I think to some extent, when you're selling to health systems, As you know, if the buyer is not the user in this case. You have the buyer being the ROI or finance team or someone at a big health system.
00:27:02
Speaker
And there may be a little bit... Can we do a time match? Yeah. i i don't i think I don't want to be misleading in terms of like... I i think like talking about like the silent co-worker, like yeah my my my little ah comment around you know almost you know not being noticed at all. I think...
00:27:22
Speaker
I don't want to give the false impression that like, you know, regard is some like nefarious AI working in the, in the shadows and, you know, you know, your, your patient data isn't safe. So I, if possible, I would, if we could kind of strike that, but I think I like where you're going next with the same question. i just want to kind of, uh,
00:27:40
Speaker
Okay, okay, okay. I'll rephrase that. Yeah, I guess I met it i meant it in more optimistic light, like, oh, it's helping in the background. But I guess people will be like, oh, no, what is it doing? Yeah, okay. um Okay, I'll just... Okay, sure, sure.
00:27:52
Speaker
Yeah, I think that's really interesting. i I think the end goal is to have AI working like alongside the clinician and just like not being something the clinician has to like spend so much of their day worrying about. They don't want more documentation and more things to deal with.
00:28:05
Speaker
They want their job done and spend as much time with the clinician as possible or the patient as possible. And I think to some extent, mirroring or matching your workflows to a clinician is a little bit difficult because, as you mentioned, there's so many. They're hybridized. They're not They change by system to system. Each person has their own different, I'd say, almost approach to different tasks.
00:28:25
Speaker
So when you're doing that, I'd be curious. one One issue I see is that the buyer and the user are a little bit different. You have someone who is a finance executive or someone else in the health system who commands like the ah ROI and wants to see his control for the spending.
00:28:40
Speaker
And that's what you're selling to. But your user is a doctor who you're most familiar with. So you have that kind of maybe a disconnect or some little pathway there. How does that look like when you're trying to sell to a larger health system and um but build for ah a clinician? Is there like a little bit of a disconnect there? Or do you guys have some way to make sure that things are aligned and you can advertise everything and it's just there's some communication between the two. That's actually one of the largest challenges of working in the healthcare space that the the user is often not the buyer, right? The person who is making the decisions to ah integrate the software, purchase the software um is ah oftentimes not the one who's, you know, at the bedside taking care of patients. and so
00:29:24
Speaker
um And so it's a, sometimes there's quite a big divide, as you mentioned, you know, the the kind of the executive suite, the the admins might be focused on certain KPIs, ah certain, you know, ROI for the health system. that They might be looking at certain performance metrics that while certainly informed by the care that is ah um that is taking place at the patient's bedside can feel very alien to, you know, the average clinician who day to day is trying to ah trying to do their work.
00:29:59
Speaker
and their work to take care of patients. um But I think one of the, it's it's a challenge to navigate. I think one of the strengths of regard really trying to unify those two, trying to bring those together, right? Trying to say, hey, we can help health systems achieve more, achieve better, achieve their quality goals while also helping physicians be more efficient and take better care of patients, right? These two things that oftentimes from at least from certain viewpoints might have been opposed historically don't need don't need necessarily to be opposed. They can work in concert. I think that one of the role of regard and other other novel technologies used in healthcare care is to smoothen that path and align incentives.
00:30:49
Speaker
That makes sense. I think ideally, can remove a lot of the friction between different things because I think a lot of the misaligned incentives don't come from but people just like, it's it's just a little bit of a friction between the different individuals within healthcare. And ideally, we can reorient that a little
Future Role of AI in Healthcare
00:31:05
Speaker
bit.
00:31:05
Speaker
um But one question I was a little curious about is that like having AI within healthcare, I mean, what do you envision as being like a long-term feature there. I think that is little bit like a philosophical question about how much support should AI provide, like in terms of decision support or just, um you don't want to kind of de-skill these clinicians. And I think it's unlikely that'll happen if you're just kind of automating more recurrent and less intensive workloads. But from that lens, like where do you think AI should be in the next three to five years in terms of integrated within a clinician's pipeline? Like where should they be helping? Where should they not?
00:31:40
Speaker
and what do you think the future is from that lens? Oh, that's such a difficult question. You know, we're really, Yeah, I know, I know. Or yeah yeah asking the big questions. You know, this is, as you as we all know, the the field is moving so quickly. It's almost impossible to ah to see ah many years into the future. And, you know, not just in the healthcare care space, but in the general kind of AI, LLM world, you know, there's ah there's many different ah paths that people have talked about in terms of, you know, everything from, you know, what
00:32:15
Speaker
AI doomerism to people who talk about artificial general intelligence. you know I think from ah my own personal point of view, I probably take ah a middle path. i don't I think these are going to be incredibly helpful tools that are definitely going to change a lot of the workflows and a lot of things that we do, but i don't think ah you know I don't think it's going to fundamentally replace the role of of humans in many areas of society. um and and hopefully if implemented correctly, won't be um
00:32:46
Speaker
won't be negative. I think talking about healthcare in particular, the obvious ah way to leverage these tools early on which a lot of which a lot of companies have been, is in automating a lot of the back office processes, a lot of the administrative processes. And there's certainly huge opportunities there, right? That I think will create a lot of efficiencies.
00:33:07
Speaker
What I would encourage us as an industry to do is to not settle just for that. I think if that's the only thing that AI is used for in healthcare, it would be um it would it would be a missed opportunity. I think we need to be thinking about how we can also leverage AI, as we talked about before, to ah improve the care that we're providing to patients, to to discover new ways that we can deliver care, new ways um that ah that we can combine you know treatments to ah to improve patient care. There's so much, I think, um
00:33:45
Speaker
hidden in a way in the data that we have left to discover. And so, um you know, i hope I hope we use this opportunity to really leverage that.
Integrating Policy, Technology, and Care
00:33:52
Speaker
um And again, i I think from a physician point of view, there's so much that physicians bring to the table beyond just a interpretation of data ah that I think physicians should really, ah physicians in all clinicians, you know, clinicians of all stripes should um should jump at the opportunity to use this additional tool to help us provide better care and do and do the the work of healing better.
00:34:21
Speaker
Yeah, I think that that makes a lot of sense. And ideally, it's it's hard to say what will ai where will AI be, but we can kind of have our own opinions on what we hope it'll do in healthcare. care And I think there's a lot of things that might hope it'll do, and the timeline for each one of them might be a little bit different. But before we end off here, one really quick question I'd love to hear. mean, you've looked at healthcare from both a policy angle clinical like delivery angle and a technology angle through AI.
00:34:45
Speaker
And I was curious, do you see these as all three different pathways to improving the overall patient journey and experience in healthcare? Do you think these things will come together? and you see friction between these two different angles? Because from what I can see, policy is trying to enable or reorient the delivery side and AI is trying to help that too. But It seems like some things kind of collide or not so mesh together as well. But from your perspective, what do you think is going to happen the next few years? Or what do you think is going on right now? They absolutely need to. You know, we can talk about I think there's different signals, you know, the but in order to.
00:35:20
Speaker
if we're going to align around ah leveraging ah AI to implement it in a safe and effective way that benefits patient care, that improves health care workers' lives ah and well-being, it needs to be a unified front. We can't be you know just you know just industry or just policy or just the health care systems. um I hope that all three of them, you know, i think each one of those can stymie efforts in their own way. Right. And so I think what it really is going to take is a lot of collaboration and um and stakeholders who, again, are keeping in mind that at the end of the day, what's most important is the patient is patient care um so that
00:36:07
Speaker
everything that we do in terms of the technologies that we develop, how we develop and test it, how we implement it on the industry side, the way that health systems incorporate this into their day-to-day workflows, and the way that policy chooses to regulate that, um reimbursement models, everything aligns around ah making sure that the care that we deliver is better, not just you know ah moving perverse incentives or moving numbers on a spreadsheet, right?
Optimism for AI's Impact on Healthcare
00:36:38
Speaker
Yeah. And hopefully, want tot stay optimistic about the future. i think it's going to be it's very difficult to do so. you have to take all three angles and not all of them are as easy to move as the other. But I think technology is the one I'm most excited about. It's it's newer, but also things are moving moving really, really quick lately. So I'm super excited to see how that goes and just where regard goes in the future. But I really appreciate the time today again, Francisco, just walking us through not only your journey, but just what's going on with the debates around AI and healthcare, I think.
00:37:06
Speaker
A lot of times people think AI will change it instantly, but it's good to be grounded in terms of what do clinicians need and where can AI help today? And we'll let you look at in the future. thanks again for coming on.
00:37:16
Speaker
appreciate your time. Absolutely a pleasure, Nikhil. Thanks lot.