Dr. Michelle Mello, professor of law and health policy at Stanford University, joins The Future of Medicine for a conversation about one of the most frustrating and consequential parts of American healthcare: prior authorization. For many patients and clinicians, prior authorization is the process behind the letter saying an insurance company will not cover a medication, procedure, or service unless additional criteria are met. As Dr. Mello explains, this system is already difficult, time-consuming, and often painful for patients and care teams. Now, artificial intelligence is beginning to reshape how these decisions are made.
In this episode, Dr. Mello and host Dr. Euan Ashley discuss how insurers and healthcare systems are using AI in prior authorization, billing, appeals, and utilization management. AI may be able to speed up routine approvals and reduce administrative burden, but it also raises serious questions about transparency, bias, accountability, and whether patients are being protected from wrongful denials.
The conversation explores what Dr. Mello describes as an “arms race” of AI systems, with insurers using technology to evaluate requests and health systems using their own tools to identify denials most likely to be overturned and draft stronger appeal letters. She also explains why appeals matter: a large share of insurance denials can be reversed, but many health systems do not have enough staff to appeal every case.
Together, Dr. Mello and Dr. Ashley consider what responsible AI governance should look like in healthcare, why simply banning AI may not solve the underlying problems, and how AI could be used in ways that support patients, clinicians, and fairer decision-making. They also discuss the need for greater transparency from insurers, the role of private governance, and why academic medical centers have a responsibility to help shape the safe and ethical use of these tools.
The Intro
“I got one of those letters last week for a drug, and they said, ‘You've not satisfied the criteria to have this more expensive drug. We're going to send you back to your doctor to find something cheaper.’”
Dr. Michelle Mello is a professor of law and of health policy at Stanford University. She researches issues at the intersection of law, ethics, and health policy.
"Because a huge percentage of denials are actually overturned if you appeal them — about half — but we don't have enough people."
In our conversation, we discuss the arms race of AI systems deployed by insurers and healthcare systems, and the role AI now plays in approving or rejecting pre-authorizations.
"We don't really have any visibility into how insurers are prompting them, but a risk is that there would be some leading of the model to frame things in such a way that it would seem less likely to be approved.”
We also talk about the need for transparency and governance in AI innovation, and who is responsible for putting guardrails in place.
“We've moved from a world where the government is interested in putting up safety guardrails to a world where it's all about accelerating innovation.”
Welcome to Stanford Department of Medicine's inside look at the future of medicine.
The Transcript
Dr. Euan Ashley: Well, Michelle, welcome to the Future of Medicine.
Dr. Michelle Mello: Thank you so much for having me.
Dr. Ashley: We were so lucky to hear your incredible grand rounds presentation to our department, and I'm really looking forward to talking through some of those issues and some other things. We always like to start, though, by finding out a little about the background of our guests. So I usually ask people to tell us a little of your life story and how you ended up here at Stanford working on this really fascinating and important topic.
Dr. Mello: Sure. Well, I've always been interested in the sociology of medicine. I think that's where things started with me, trying — I would say as early as college. But certainly, as more and more of my colleagues became physicians and I started to understand some of the things that made their lives miserable, it made me want to think about the role of policy in changing some of those things. And as somebody who was on a road to law school at the time as well, I zeroed in on medical liability as the first issue that I became really interested in trying to have an impact on.
Dr. Ashley: But give us a sense of your background. Where did you grow up?
Dr. Mello: I grew up in California's Central Valley, in a farm town. I was an undergraduate here at Stanford, which changed my life. And after that, I spent a couple of years in the UK.
Dr. Ashley: Common among the people you grew up with?
Dr. Mello: No. In my high school, only a handful went to four-year colleges. So it was a life-changing moment for sure.
Dr. Ashley: And what drew you to Stanford?
Dr. Mello: What didn't draw me to Stanford? In truth, when I was 11 years old, we spent six weeks here because my brother was in the hospital, and I was left to roam the grounds. And with that sort of hubris that only a middle schooler has, I thought, "Oh, this is the place for me."
Dr. Ashley: Right, of course. Well, you were right. There's plenty of space to roam indeed around here. That's true. And not so far — I mean, it's far from home in many ways, but not so far in some other ways geographically.
Dr. Mello: Well, it's a very different California than the one I grew up in, that's for sure.
Dr. Ashley: Right. And you bring that, I think, to all the work that you do — that sense of the fact that the bubble we live in here isn't the real world.
Dr. Mello: Well, yes. And that's the other thing I maybe would mention about my upbringing: I came from a family of very modest means, for whom healthcare access was a struggle financially. And then I spent time in the UK after I was an undergraduate here, and thought, "Okay, these guys have figured out a different way of providing access."
Dr. Ashley: So that was a Rhodes scholarship? A Marshall?
Dr. Mello: A Marshall scholarship, right. So one of the equivalent scholarships to Oxford. Yes.
Dr. Ashley: A place I had the joy of spending a few years myself. But yes — it's not just that I traveled to the UK a little bit. You were on this very prestigious scholarship.
Dr. Mello: I did a master's degree there in comparative social policy. Which really got me thinking about how there's more than one way to think about questions of what we owe to one another in a Western democracy, and how we provide those things.
Dr. Ashley: What are your memories of Oxford? Did you enjoy your time there?
Dr. Mello: It was fantastic. I guess the strongest memory is just having time for deep reading. As you know, the system there at the graduate level is to go off and read these 200 books and come back in a couple of years and sit this exam —
Dr. Ashley: Wearing formal clothes.
Dr. Mello: Wearing gowns.
Dr. Ashley: Yes. The amount I sweated during my PhD —
Dr. Mello: In a way that prepares you for today's graduation ceremonies, which now involve a lot of sweating.
Dr. Ashley: Yes, a lot. The weather is a little different there. You're sweating because you have so many clothes on. Here you're trying to take — wondering how few clothes you could actually wear.
Dr. Mello: That's right.
Dr. Ashley: This is one of the ways the Scottish kilt is helpful, actually, for California graduations. Turns out — that's quite helpful. Although they are made of thick wool. So you win on one side, because it's airy, and you lose on the other, because it's thick wool. So I wouldn't necessarily advise it. I'm usually the only person wearing a kilt at Stanford's graduation. But anyway — Oxford: an amazing town, and an amazing place to spend a bit of time. But also, with this academic focus, it sounds like you had some great mentors there.
Dr. Mello: Yeah, really did.
Dr. Ashley: What happened after that?
Dr. Mello: I came back to the States. I started a PhD program in health policy down at Chapel Hill, and then a couple of years later, I went to law school.
Dr. Ashley: Right. At Yale?
Dr. Mello: Yes.
Dr. Ashley: Yeah, I did read your bio.
Dr. Mello: You did.
Dr. Ashley: For a girl from a small town in California, these are pretty high-end institutions.
Dr. Mello: It's been a very fortunate trajectory, to say the least.
Dr. Ashley: Well, I think a little skill and talent come along the way as well. And then you went back to Stanford, right?
Dr. Mello: Well, by way of Harvard. I spent almost 15 years —
Dr. Ashley: How do you spell that again? H-A —
Dr. Mello: And then my dearest colleague there left and came here, and made the case very persuasively that this is really the place — this is where you should be.
Dr. Ashley: Okay. Well, we're very happy to have you. And of course, what an amazing career, an amazing trajectory. One of the reasons why you were one of our most requested speakers for our grand rounds series — as I was mentioning earlier, very well received, and lots of thoughtful conversation. This intersection right now, in particular, of AI and policy and healthcare and access is one of the real sweet spots where you've done really important work. To dive into that, we have to define a couple of things — phrases that are well known to some of us and maybe a bit less well known to some of the people listening. Let's start with pre-authorization. What is it? Why do we care? And let's see what we might be able to do about it.
Dr. Mello: Yes. So people may not know the term pre-authorization, or prior authorization, but chances are they've been through it. If you've ever gotten a letter from your insurance company explaining that they're not going to cover a service that you haven't received yet but your doctor thinks you should have, that's prior authorization. I got one of those letters last week for a drug, and they said, "You've not satisfied the criteria to have this more expensive drug. We're going to send you back to your doctor to find something cheaper." So that's a process that we all probably have been through, and that you probably spend a fair amount of time on as a clinician as well. And, as I'm sure we'll talk about, everybody hates it.
Dr. Ashley: We definitely hate it.
Dr. Mello: And so the question now is: how can AI make that better, and how might it make it worse?
Dr. Ashley: And obviously it seems obvious — insurance companies are trying to save money. They have their own business, and their business is to take money from people and then try to make the books add up while they help pay for healthcare. But there's an immediate tension within that, where they're trying to save money and make more margin, because in most cases they do have either shareholders or owners that they're looking to speak to. But it's probably the single most frustrating element of what is a very frustrating spectrum of non-medical elements that doctors and other healthcare providers have to deal with. And we literally spend — sometimes it's us, generally the doctors; often it's our staff — who spend literally hours waiting on the phone, first of all just to get put through, and then put on hold, and then to find essentially someone with another screen on the other side who's giving them fairly standard answers, and then they give their standard answers. This is the way it used to work; you're going to tell us in a minute how it started to work now. And then they maybe come back and say, "Okay, well, the outcome is a doctor needs to talk to another doctor." So then a highly specialized doctor usually will talk to a non-specialized, often retired general doctor who's, I assume, incentivized to get to the point of not paying for something, whereas we're simply there trying to provide the best care for our patients. Nothing is ever completely black and white, and everybody's trying to get through the day, but that tension — the system is built in a way that's built for failure.
Dr. Mello: And, to be fair to insurers, they do need to keep premiums down. We all want that; we all pay that cost. And this is pretty much the best idea they've had so far. In other work that I've done with Professor Alyce Adams here, and one of our medical students, Steven Gong, we looked at the history of insurers' attempts at cost control, and every other thing that we've tried — for example, making networks narrower, limiting which physicians and which hospitals patients can see — are also terrible for people. You also constrain access, and prices just keep going up and up and up. And that's not really insurers' fault either. So we've got to do something, particularly in a fee-for-service system where physicians who are trying to do things in patients' best interests face no personal financial constraints on doing more. They get financially rewarded for doing more. So it's a system designed to press costs up relentlessly. And again, this is the best idea we've had, in the sense that it does seem to have an effect on costs, and it's maybe less painful than some of the alternatives.
Dr. Ashley: The best of the worst, in a way. But there's no healthcare system, regardless of how it's organized, that can escape the idea of spiraling costs. You put your finger right on the money, literally, there — that our system is designed on both sides to push up costs, because there is an incentive on the doctor's side. It's one that we're much more immune to on the academic side.
Dr. Mello: Because you're salaried instead — but the hospital isn't, of course.
Dr. Ashley: No, right. The system in general benefits the more things that are done. And it is true that some things pay more and some things pay less, and if — I'm not an economist, but any system like that —
Dr. Mello: You're a cardiologist, you're a chair of cardiology, so you know very much about how these specialties have incentives to do more when it's in patients' best interest, and you can do it with good conscience.
Dr. Ashley: And the fact is that, again, because nothing is black and white — if the situation arises every day where it would be completely reasonable to do this procedure, it might also be reasonable not to do the procedure. And if it's completely balanced and one is incentivized by a payment and the other is not, then on balance, probably the payment will happen more often. So I don't want to suggest that we're not all part of the system, but it is also true that that process could not be less efficient, and a huge amount of our healthcare costs come from the burden of overhead and administration that could easily be lower. We know it could be lower, because lots of other healthcare systems around the world have much lower administration costs.Yes. So what is the science and the work that you've done telling us about how modern-day tools might help that process be a little more efficient, and what sort of arms race have we entered?
Dr. Mello: Well, it's important to recognize that the AI market has responded to all kinds of things that could be more efficient in healthcare. I think we're going to talk mostly today about insurance processes, but something that's been in the news very recently has been tools that help providers bill — which is probably another thing you hate to do, right? Find those codes and enter them.
Dr. Ashley: Seventeen clicks, I think, to draw up one bill, right?
Dr. Mello: So, on the one hand, the federal government has been trying to identify tools that can detect fraud, waste, and abuse in billing — upcoding. On the other hand, there's a separate market of tools sold to billing departments and hospitals that would like to be sure that you are billing for everything you possibly can. In my group, we call these tools, facetiously, "no code left behind" — you want to just make sure you've done as thorough a job as you can claiming reimbursement. So there's all kinds of controversy about those tools, because Blue Cross Blue Shield has just issued a report saying that they are associated with much, much higher costs. On the other side of the coin is the pre-authorization and other processes that insurers have, to say no sometimes to services that require clearance. And there too, what we see in the marketplace is companies that are marketing both to insurers and to providers. And maybe we can talk separately about each of those kinds of tools. What's gotten the most media attention is tools that are for insurers — for prior authorization, but also for concurrent review. Like, if you've got a patient you've sent to a rehab hospital, and you'd like her to stay for another two weeks, and the insurer isn't so sure about that, they might ask you to put in a medical-necessity justification, and we've got tools now for the insurers that can review those. And then there's retrospective utilization management, where you've already received some service and now the insurer has to decide whether to pay it or not — tools for that as well. And all of those are designed, purportedly at least, to sort requests into clinically appropriate and clinically inappropriate, or perhaps clinically inappropriate.
Dr. Ashley: And just to be clear — this is an AI system that looks at the medical record of an individual patient, at their chart, the notes from the doctor every day, their labs?
Dr. Mello: No. What insurers will quickly remind you of is that they don't have access to our medical records. And that puts them at a huge disadvantage in evaluating these requests, particularly if, as you noted, the request was maybe prepared by someone who's not a clinician, maybe didn't do a great job of finding the right information, left some fields blank. They can't just fix that, right? So instead they have to go back and forth with your office to figure —
Dr. Ashley: They made a request on a sort of survey, almost, like a questionnaire.
Dr. Mello: Yeah, they just get a form with this person's answers to certain questions. But the AI can help identify things that are missing or blank, and resolve those questions quickly. It can also automatically approve things that are clearly in bounds — and that's most of the time. We complain about the 10 or 15% that get denied, but most of the time the answer is yes.And so rather than have a human process all this paperwork, let's just have an AI look through there and say —
Dr. Ashley: Faster and cheaper.
Dr. Mello: Yeah. And then, for harder cases, it can prepare a preliminary determination that the request isn't approvable, and then a human medical professional would review that request at some level of depth — which is debatable.Right. So that's all on the insurer side. And then on the provider side, because clinicians like you do not enjoy preparing these forms, you delegate it down. And again, some of the people who are preparing those requests struggle to do that well. There are now tools that will pull information out of the electronic health record, populate those forms for the submitter to look over. And then there are a range of tools that deal with the no's. So when you get a denial back, we have a tool here at Stanford that can look through all the denials that we've got. Now, we don't have enough people to process appeals of all of those denials, right? And that's a shame, because a huge percentage of denials are actually overturned if you appeal them — about half — but we don't have enough people. So we can use this tool to find the denials that are most likely to be reversed if we were to appeal them, based on history. And then we have another tool that can draft the appeal letter and find the missing information in the record.
Dr. Ashley: What you're describing, though, is now an arms race of AI systems —
Dr. Mello: Yes.
Dr. Ashley: — where they're being deployed by the insurance company, hopefully in many cases for the straightforward ones, to get through paperwork quickly and approve them; in other cases, to begin the process of denying them. And then, on the healthcare-system side, another AI that could potentially then be talking directly to the insurer's AI without a human in the loop.
Dr. Mello: Well, there will always be a human in the loop on the insurer side, because right now, as a matter of law, they can't say no without a human having passed their eyes over the thing and clicked submit. On the provider side, we may well get to a point where it's more automated, but again, you're legally liable for fraud if you submit something that's false. So there are incentives for humans to review — but the question is how thoroughly.
Dr. Ashley: Yeah. And I think that's inevitable. We talk about this a lot, even on this podcast — we've talked about it before. Our cognition is designed such that if you're reading very well written, very well punctuated, capitalized text — we were talking earlier about threes, with some of the language models — it's just very good at writing.
Dr. Mello: Yes.
Dr. Ashley: And we're kind of trained through our lives that when we're reading newspaper articles, we're not reading them as carefully — we're not interrogating the text for accuracy in general if it's beautifully written. If we read text as a teacher — as a lecturer, as a faculty member — we're often correcting, or trying to provide feedback on work that has been submitted to us, say by students; that's a different form of interrogation. And so those two cognitive states — one where your brain is quickly moving, even if you are reading at some level every word, your brain's quickly moving through something that's beautifully written, versus something that needs some work where you need to comment — they're very different. And I imagine, as you say, even though those humans are in the loop, it's the former: they're letting it go quite quickly, because it's written well, and the assumption, at some subconscious level, is that it's correct.
Dr. Mello: That's interesting. That seems very plausible. And it's also the case that humans in general are subject to framing bias and confirmation bias. So when we receive, let's say, a preliminary write-up of an AI's decision about your request, and it states its conclusion and gives a reason, we are now being asked to move off of a default as a human reviewer, right? Here's a default — the default is no. If I'm going to move off that, it's going to take more cognitive work, and I'm going to actually have to go into that file and find a reason to reverse it. So not just with AI, but in all kinds of things, we humans are prone to confirm previous things, especially when they happen to align with our priors about a particular kind of service or patient. And then there's framing bias too — AI is really good at teeing up a summary in the way that you tell it to, and we don't really have any visibility into how insurers are prompting them. A risk is that there would be some leading of the model to frame things in such a way that it would seem less likely to be approved.
Dr. Ashley: So a lot of this is important work in understanding how the system is working, and it's moving pretty quickly. Where do your conclusions lead in terms of what we might do about this, or what better system we could move toward?
Dr. Mello: Well, I think one of the things that I talk to policymakers about a fair amount is that we need to resist the urge to just grind everything to a halt. There have been states that have proposed outright bans on using AI in these processes, for example. That is a memory failure, because, again, we have to remember what the counterfactual is — which is that it's terrible right now, right? And we're also moving into a space where lawmakers have created more time pressure for insurers, because patients have complained about delays in getting these decisions. Now, many plans have to make this decision routinely in seven days, sometimes in just 48 or 72 hours. So there's all kinds of pressure pushing them to be more efficient — and again, because most of the time the answer is going to be yes, we want AI enablement before that.The question is, how do we do that while protecting patients from wrongful decisions? And it's really hard to come up with a policy prescription right now, because insurers, while vigorously protesting the assertion that they over-rely on AI, have given us very little information to help us understand: is AI making things better or worse? But maybe an interim measure would be to say: use it to make auto-approvals, but as soon as the AI's judgment is that it's not approvable, just take the AI out and give the human a fresh file with no curation, no framing. So that removes a lot of the workload, saves the human for the 15% that involve really hard calls, but doesn't prime them to agree with the model.
Dr. Ashley: Yeah, that makes a lot of sense. I think one of the things, though — and maybe it's even a little counter to that — is that what the models are really good at is the long tail of knowledge –
Dr. Mello: The things that humans aren't good at recalling.
Dr. Ashley: I think that's where they excel. And ultimately, the humans who are often in those positions — there's a stereotype here, and it's not true all the time; these are caring people — but often it is the classic situation of a retired generalist of some kind who's doing a little extra work in their retirement. I can just speak to my own experience in doing those calls: it's very rare that I've ever spoken on a peer-to-peer call with someone who was remotely expert in what we were talking about. And that's a challenge. So in a way, it is good to have the human, because we don't want bias, and the models can be biased. But the models have this massive advantage, which is that they really do have enormous knowledge, especially of subspecialty care, that I don't think anyone would argue goes beyond — well, goes beyond what even the subspecialist might know, and certainly goes beyond what an insurance-employed, likely generalist at the end of the phone would know.
Dr. Mello: That's very interesting. So it raises a question of whether AI might play a different kind of role in helping a medical reviewer understand a case. So if they had access to OpenEvidence or something at the time of the human decision, they could make a better-informed decision with the aid of AI that is not designed for insurance processes at all — just designed to make everybody better as a clinician.
Dr. Ashley: Right. I think that's certainly one thought, where we're all thinking about how this could lift all the boats. And certainly, for the most part, most doctors most of the time are seeking to adhere to the guidelines that their professional bodies have extolled. There are very few who are heading out on their own, making rare decisions. And those guidelines are available, but they're like 72 pages, or 150 pages, and a generalist who's going to work through a call list isn't going to have time to read those 150 pages. But again, the models could certainly help parse that.
Dr. Mello: And if we think about those provider-facing tools, also, a lot of those are built to directly incorporate guidelines and recommendations into your request or your appeal, right? To help that medical reviewer in the insurance company.
Dr. Ashley: And in a way, I think that's probably what they're looking for. They're like, "Help me help you," right?So that seems like it could be a real benefit, in the end, to the patient — who is the one that we care about. Yeah. There are so many different elements. I think most people don't realize that this is actually probably the most penetrant area of AI — this is probably the area where AI has been most penetrant, at least in the American healthcare system. We talked here just a few weeks ago, for example, with Bob Wachter, who's the chair of medicine at UCSF and has written a recent book — it's probably right here — about AI and medicine. And he goes through lots of use cases, including some that have been really transformational for practicing doctors, like ambient scribes — really just the ability to bring two people back together again, where the electronic health record sort of pushed them apart. So there are lots of examples that are coming, but I feel that's the only one that's probably as penetrant as some of these.
Dr. Mello: Yeah, I think maybe even more so — although, again, it's hard to know, because insurers aren't very transparent. But about three-quarters of them report that they either are using now, or will be using, AI for some purpose within the next year.
Dr. Ashley: Yeah, it seems very likely. And I think there's another question there — as to how much transparency do you think we need, or is it fair to ask for? They're private companies at the end of the day, into how they're using those AIs for their processes.
Dr. Mello: I think it's very fair to ask for transparency, and in fact they've volunteered it. So last summer, in response in part to some stories about use of AI, a press release was issued by America's Health Insurance Plans, which represents all the big insurers, committing to a number of steps to make prior authorization better, including greater transparency. And I think it's also fair because, they may be private companies, but they're doing huge business with our government, and with all of us as taxpayers. So to answer basic questions about how they're making decisions is in no way out of bounds. Of course, we have to protect patients' confidential information, but simple data — like, what did you do before, what was your approval rate, denial rate, overturn rate for different categories of service, and then what does it look like when you use AI? Those are things that state insurance regulators are already starting to request.
Dr. Ashley: Yeah, that definitely makes sense. It seems such a rich area, and a rich time, for someone with your background and your interests. How do you choose what to work on next? I'm really interested in your thought process.
Dr. Mello: Well, whenever anyone asks me a question like that, I always feel anxious that I don't really have a plan, because I don't have any strategic vision. But I try to go where the need is, I think, as we all do. And these days in health policy, there is a great deal of need for information to try to — honestly, mostly — push back against most of the things that are being done federally right now. But in the AI space, I want to continue to try to work on strengthening private governance of AI, because we've gotten a very clear signal with the changeover in presidential administration that we've moved from a world where the government is interested in putting up safety guardrails to a world where it's all about accelerating innovation, really. So that leaves the private sector to try to create those guardrails and make sure that innovation is balanced with oversight. And we've done, I think, really landmark work here at Stanford Healthcare, trying to make sure that AI here is safe, and used responsibly, and is fair to our clinical workforce. So I'm trying to help other organizations figure out how to do that, including, I hope, the insurance industry — which likes to talk about how responsible it is in this area. But, you know, a company like UnitedHealthcare has over a thousand AI use cases running right now.
Dr. Ashley: Wow.
Dr. Mello: Yeah. So you tell me how you governed that. We haven't figured that out at Stanford at that scale, but it needs to happen.
Dr. Ashley: Yeah. And I'm sort of fascinated as well, because of the unique situation that academic healthcare centers live in — right next to, and in Stanford literally owned by, the university, in a not-for-profit setting. Of course, the hospital, the healthcare system, has to operate as a business. But we do have, I think, a unique ability to potentially step back and lean on our academic base to provide, I hope, neutral advice to governments, to the regulators, as to how to move forward. I think it's really our responsibility. It's one of the reasons we're here.
Dr. Mello: We helped create these technologies, and so we should be leading in their responsible use.
Dr. Ashley: Yeah, I think that's important. And it does appear like there are plenty of questions. It doesn't seem like you're going to run out.
Dr. Mello: No, I haven't got the sense that AI in particular is going to put me out of work anytime soon.
Dr. Ashley: No, it doesn't seem like it. I always love to ask some of our guests about other things that they do. And I happen to know that you spend time in a boat on the water.
Dr. Mello: I do. And you know that because I row with your wife. That's correct. Yeah, I do. I've been a rower for about 30 years.
Dr. Ashley: I see. What first took you into rowing?
Dr. Mello: Well, as you know, in Oxford it's actually a spectator sport, unlike in the US. So I think I got drawn in there. But the reason I've stayed with it is that it's this rare opportunity to go deep inside yourself and really focus on one thing for an extended period of time, several times a week. In our age of digital distraction, I find myself lacking the discipline to do that when my phone is next to me. But it's an opportunity for self-improvement and focus that helps me in other parts of life.
Dr. Ashley: So help me with this: why do rowers get up so early in the morning?
Dr. Mello: Well, you name another time that 50 people with jobs can get together and be happy to — Retire to the early-morning practices.
Dr. Ashley: That's true. I have lived with someone whose alarm has regularly gone off at 4:45 in the morning. And I used to be a complete — I've changed these days; I try to get to bed on time — but I was always a night owl. I would be up late.
Dr. Mello: That's tough.
Dr. Ashley: Coding as a — So there were times when I would be going to bed at 1:30 or 2:00 and the alarm was going off for my wife to go off to rowing at 4:45 or 4:30.
Dr. Mello: You're a rowing widow.
Dr. Ashley: Yes, a little bit. I've been at many rowing events. I've been a spectator at rowing events. They're not so easy, actually, to be a spectator at, unless you literally have a bunch of people around. They're like baseball that way.It's partly about the social, because the boat goes by quite quickly.
Dr. Mello: It's just you and your beer after that.
Dr. Ashley: Correct. But an amazing sport from an exercise-physiology perspective. When I'm taken to parties with other rowers, I push them into that domain, since that's a shared interest. And certainly rowing is a sport of study, since it's one of the whole-body exercises — you're literally using every muscle in your body. So those athletes have among the highest VO2 maxes, and certainly the highest ventilation, of almost any athletes.
Dr. Mello: Interesting. We also do our part to support our orthopedic surgery department, to be sure.
Dr. Ashley: Injuries happen as well.
Dr. Mello: Yeah, I've heard a few stories.
Dr. Ashley: And then I see some of the photos of the still water in the early morning at dawn, and that's the only time I'm like, "Oh —"
Dr. Mello: I had a moment like that this morning, where we had to pause and just look at the sunrise. We're so lucky to live in California and get on the bay and just enjoy this natural beauty.
Dr. Ashley: Well, Michelle, thank you so much for joining us — and all the best for the future for all the work that you're doing.
Dr. Mello: Thank you.
