Pod Insights a16z Podcast a16z Podcast | Putting AI in Medicine, in Practice About this episode There’s been a lot of talk about technology -- and AI, deep learning, and machine learning specifically -- finally reaching the healthcare sector. But AI in medicine isn’t actually new; it’s actually been there since the 1960s. And yet we didn’t see it effect a true change, or even become a real part our doctor’s offices -- let alone routine healthcare services. So: what's different now? And what does AI in medicine look like, practically speaking, whether it's ensuring the best data, versioning software for healthcare, or other aspects?
In this episode of the a16z Podcast, Brandon Ballinger, CEO of Cardiogram; Mintu Turakhia, cardiologist at Stanford and Director of the Center for Digital Health; and general partner and head of a16z bio fund Vijay Pande in conversation with Hanne Tidnam discuss where will we start to see AI in healthcare first -- diagnosis, treatment, or system management -- to what it will take for it to succeed. Will we perhaps see a "levels" of AI framework for doctors as we have for autonomous cars?
Transcript Speaker names marked ? are inferred from speaking time and the show’s known hosts — they are not voice-verified.
0:00 Hanne Tidnam? Hi and welcome to the A16Z Podcast. I'm Hannah and today we're talking about AI in medicine, but we want to talk about it in a really practical way. What it means to use it in practice and in a medical practice, what it means to build medical tools with it, but also what creates the conditions for AI to really succeed in medicine and how we design for those conditions, both from the medical side and from the software side. The guests joining for this conversation in the order in which you will hear their voices are Minto Tarakia, a cardiologist at Stanford and director of the Center for Digital Health; Brandon Ballinger, CEO and founder of Cardiogram, a company that uses heart rate data to predict and prevent heart disease;
0:41 Hanne Tidnam? and Vijay Pandey, a general partner here at A16Z and head of our biofund. So let's maybe just do a quick breakdown of what we're actually talking about when we talk about introducing AI to medicine. What does that actually mean? How will we actually start to see AI intervene in medicine and in hospitals and in waiting rooms?
0:59 Speaker 01 AI is not new to medicine. Automated systems in healthcare have been described since the 1960s and they went through various iterations of expert systems and neural networks and called many different things.
1:11 Hanne Tidnam? In what way would those show up in the 60s and 70s?
1:14 Speaker 01 So at that time, there was no high resolutions. There weren't too many sensors and it was about a synthetic brain that could take what a patient describes as the inputs and what a doctor finds on the exam as the inputs.
1:29 Hanne Tidnam? Using verbal descriptions?
1:32 Speaker 01 Yeah, basically words. People created what are called ontologies and classification structures, but you put in the 10 things you felt and a computer would spit out the top 10 diagnoses in order of probability. And even back then, they were outperforming sort of average physicians. So this is not a new concept.
1:50 Hanne Tidnam? So basically doing what hypochondriacs do with Google today, verbally.
1:55 Speaker 01 Right, so Google is in some ways an AI expression of that, where it's actually used ongoing inputs and classification to do that over time, much more robust neural networks, so to speak.
2:07 Speaker 02 So an interesting case study is the Mison system, which is from 1978, I believe. And so this was an expert system trained at Stanford. It would take inputs that were just typed in manually and then it would essentially try to predict what a pathologist would show. And it was put to the test against five pathologists and it beat all five of them.
2:27 Hanne Tidnam? And it was already outperforming?
2:29 Speaker 02 It was already outperforming doctors, but when you go to the hospital, they don't use mycen or anything similar. And I think this illustrates that sometimes the challenge isn't just the technical aspects or the accuracy, it's the deployment path.
2:42 Hanne Tidnam? Right.
2:42 Speaker 02 And so some of the issues around there are, okay, is there a convenient way to deploy this to actual physicians? Who takes the risk? What's the financial model for reimbursement? And so if you look at the way the financial incentives work, there are some things that are backwards, right? For example, if you think about kind of a hospital from the CFO's perspective, misdiagnosis actually earns them more money.
3:08 Hanne Tidnam? What?
3:08 Speaker 02 Because when you misdiagnose, you do follow-up tests, right? And those, and our billing system is fee-for-service. So every little test that's done is billed for.
3:17 Hanne Tidnam? But nobody wants to be giving out wrong diagnoses.
3:20 Speaker 02 Right.
3:20 Hanne Tidnam? So where's the incentive? The incentive is just in the system, the money that results from it.
3:24 Speaker 02 No one wants to give an incorrect diagnosis. On the other hand, there's no budget to invest in better diagnosis.
3:30 Hanne Tidnam? And making sure it doesn't happen.
3:31 Speaker 02 And so that's, I think that's been part of the problem.
3:34 Hanne Tidnam? Yeah.
3:35 Speaker 02 And so things like fee-for-value are interesting because now you're paying people for, say, an accurate diagnosis or for a reduction in hospitalizations, depending on the exact system. And so I think that's the case where actually accuracy is rewarded with greater payment, which sets up the incentives so that AI can actually, can actually win in this circumstance.
3:55 Speaker 01 Where I think AI has come back at us with a force is it came to healthcare as a hammer looking for a nail. What we're trying to figure out is where you can implement it easily and safely with not too much friction and with not a lot of physicians going crazy, and where it's going to be very, very hard. And that, I think, is the challenge in terms of both building, developing these technologies, commercializing them and seeing how they scale. And so the use cases really vary across that spectrum.
4:24 Speaker 02 Yeah, I think about there's being a couple of different cases where AI can intervene. One is to substitute what doctors do for already, and so people use the example of radiology as an example.
4:36 Speaker 02 The other area that I think is maybe more interesting is that AI can complement what doctors can't do already. So it would be possible for a doctor to say, read an ECG and tell you whether you're in an abnormal heart rhythm. No doctor right now can read your Fitbit data and tell you whether you have a condition like sleep apnea. I mean, if you look at your own data, you can kind of see restful sleeve has real structured REM cycles. So you can see some patterns there. That said, the gold standard that a sleep doctor would use is a sleep study where they wire you up with six different sensors and tell you to sleep naturally. There's a big difference here between kind of the very noisy consumer sensors that may be less interpretable and what a doctor is used to seeing.
5:17 Vijay Pande? Or it could be that the data is on the device, but the analysis can't be done yet. Maybe the analysis needs a gold standard data set to compare to. There are a lot of missing parts beyond just gathering the data from the patient in the first place.
5:30 Speaker 01 I think there's some inherent challenges in the nature of the beast. Healthcare is unpredictable. It's stochastic. You can predict a cumulative probability, like a probability of getting condition X or diagnosis X over a time horizon of five or 10 years, but we are nowhere near saying, you know, you're going to have a heart attack in the next three days. Prediction is very, very, very difficult. And so where prediction might have a place is where you're getting high fidelity data, whether it's from a wearable or a sensor. It's so dense that a human can't possibly do it, like a doctor's not going to look at it. And two, it's relatively noisy, inaccurate, poor classifiers, missing, periods where you don't actually have this continuous data that you really want for prediction.
6:15 Hanne Tidnam? Right.
6:16 Speaker 01 In fact, the biggest predictor of someone getting ill with a lot of wearable studies is missing data because they were too sick to wear the sensor.
6:23 Hanne Tidnam? Oh, that's so, so the very absence of the data is a big indicator.
6:26 Speaker 01 Yes, exactly.
6:27 Hanne Tidnam? That's so interesting. I don't feel well enough to put on my what have you, and that means something's not right.
6:34 Speaker 01 Possibly, or you're on vacation. And so that's the problem, that's the other challenge of AI is context. And so what are some of the more simple problems where you have clean data structures, you have less noise, you have Very clear training for these algorithms. And I think that's where we've seen AI really pick up in imaging-like studies. It's a closed-loop diagnosis. You know, there is a nodule on an X-ray that is, you know, cancer based on a biopsy proven later in the training dataset, or there isn't. In the case of an EKG, well, we already have expert systems that can give us a provisional diagnosis on an EKG. They're not really learning. And so that's a great problem because most arrhythmias don't need context. You can look at it and make a diagnosis.
7:19 Hanne Tidnam? We don't need them to learn, so that's why it's good to use right away to apply this technology to immediately.
7:24 Speaker 01 You don't need everything. You don't need to mine the EMR to get all this other stuff. You can look at the image and say, is it probably, does it have a diagnosis or does it not? And so, imaging of the retina, imaging of skin lesions, X-rays, MRIs, echocardiograms, EKGs, that's where we're really seeing AI pick up.
7:46 Speaker 02 I sort of divide the problems into inputs and outputs. And we talked a little bit about some of the inputs that have newly become available, like EMR and imaging data. I think it's also interesting to think about what the outputs of an AI algorithm would be. And these examples are kind of self-contained, well-defined outputs that fit into the existing medical system. But I think it's also interesting to imagine what could happen if you were to reinvent the entire medical system with this assumption that we have a lot of data, intelligence is artificial and therefore cheap, so we can do continuous monitoring. So one of the things I think about is, what are the gaps of people who do not have access to EKGs, right?
8:28 Speaker 02 Most people, like, I've actually never had an EKG done, aside from the ones I do myself. So, and most people, actually, in the US, you get your first EKG when you turn 65 as part of your Medicare checkup.
8:39 Hanne Tidnam? Oh, wow. I didn't realize, it's so late.
8:41 Speaker 02 You know, people like my parents. So like, my dad's an excavator, so he digs foundations for houses, and he hasn't seen a doctor in 20 years. And if he leaves a job site, the entire job site was shut down. So it's hard for some people, I think, to go into the doctor's office between the hours of nine to five. And if you look at that in aggregate, about half of people in the US have a primary care physician at all. Which seems astonishingly low, but that's that's kind of the fact. There's kind of a gap, right? About a third of people with diabetes don't realize they have it. About a fifth of people with hypertension. For AFib, it's 30 or 40%. For sleep apnea, it's like 80%.
9:17 Vijay Pande? And I think it's one thing just finding out, but not being able to do anything about it. But the actionable aspect, I think, really is a huge game changer. It means that you can have both better outcomes for patients and in principle, lower costs for pairs.
9:28 Hanne Tidnam? Right, and these are areas where there are clear ways of addressing these specific conditions.
9:36 Speaker 01 I will take a little bit of a different view here, which is that I don't know if AI, artificial intelligence, is needed for earlier and better detection and treatment. To me, that may be a data collection issue.
9:50 Hanne Tidnam? How is that different from what we're saying about finding it early? How can that not be good?
9:55 Speaker 01 Because that may have to do with getting sensors out of hospitals and getting them to patients. And that's not inherently an AI problem. It could be a last mile AI problem so that if you want to scale the ability to get this stuff. Okay, so let's say we get to a point where our bathroom tiles have built in EKG sensors and scales, and the data is just collected while we brush our teeth. It's the sensing technology that may detect things discreetly, like in arrhythmia. You may not necessarily need intelligence, but who's going to look at the data? So that's a scaling issue.
10:30 Vijay Pande? Well, but the idea is, yeah, so AI could look at the data. And the other thing is that if you're using this as screening, you want to make the accuracy as high as possible to avoid false positives. And AI would have a very natural role there, too.
10:41 Hanne Tidnam? But it's interesting that you're saying it's not necessarily about the analysis. It's about where the data comes from and when.
10:46 Speaker 01 I think there are two different problems. There may be a point that it truly outperforms the cognitive abilities of physicians. And we have seen that with imaging so far. And some of the most promising aspects of the imaging studies and the EKG studies are that the confusion matrices, the way humans misclassify things, is recapitulated by the convolutional neural network.
11:12 Hanne Tidnam? Can you actually break that down for a second? So what are those confusion matrices?
11:16 Speaker 01 So a confusion matrix is a way to graph the errors and which directions they go. And so for rhythms on an EKG, and a rhythm that's truly atrial fibrillation could get classified as normal sinus rhythm, or atrial tachycardia, or supraventricular tachycardia. The names are not important. What's important is that the algorithms are making the same type of mistakes that humans are doing. It's not that it's making a mistake that's more necessarily more lethal, and just nonsensical, so to speak. It recapitulates humans. And to me, that's the core thesis of AI in medicine, because if you can show that you're
11:58 Speaker 01 recapitulating human error, you're not going to make it perfect, but that tells you that if in check and with control, you can allow this to scale safely, since it's liable to do what humans do. And so now you're automating tasks that, you know, I'm a cardiologist, I'm an electrophysiologist, but I don't enjoy reading 400 ECGs when it's my week to read them.
12:20 Hanne Tidnam? So you're saying it doesn't have to be better. It just has to be making the same kinds of mistakes to feel that you can trust.
12:27 Speaker 01 Right.
12:27 Hanne Tidnam? The decision making.
12:28 Speaker 01 And you dip your toe in the water by having it be assistive. And then at some point, we as a society will decide if it can go fully auto, right? Fully autonomous without a doctor in the loop. That's a societal issue. That's not a technical hurdle at this point.
12:45 Hanne Tidnam? Right.
12:45 Vijay Pande? Well, and you can imagine, just as for, let's say, self-driving cars, you have different levels of autonomy. It's not nothing versus everything. You know, you can imagine level one, level two, level four, level five in self-driving cars. And so, and I think that would be the most natural way, because we wouldn't want to go from nothing to everything.
13:01 Speaker 01 Exactly. And just like a self-driving car, we as a society have to define... Who's taking that risk on, right? You can't really sue a convolutional neural network, but you might be able to make a claim against the physician, the group, the hospital that implements it. And how does that shake out?
13:20 Hanne Tidnam? To figure out literally like how to ensure against these kinds of errors.
13:23 Speaker 02 I think the way you think about some of these error cases kind of depends on whether the AI is substituting for part of what a doctor does today, or if the AI is doing something that's truly novel. I think in the novel cases, you might not actually care whether it makes mistakes that would be similar to human.
13:38 Hanne Tidnam? Oh, that's an interesting point because it's doing something that we couldn't achieve. What kinds of novel cases like that can you imagine?
13:45 Speaker 02 Wearables are an interesting case because they'll generate about 2 trillion data points this year. So there's no cardiologist or team of cardiologists who could even possibly look at those. That's a case where you can actually invert maybe the way the medical system works today. Rather than being reactive to symptoms, you can actually be proactive. And the AI can be essentially something that tells you to go to the doctor rather than something that the doctor uses when you're already there.
14:12 Vijay Pande? So to take radiology as an example, where you could have one levels where it's as good as a common doctor. Another level where it's as good as the consensus of doctors.
14:22 Hanne Tidnam? Right.
14:22 Vijay Pande? Another level is that it's not just using the labels of a radiologist would say on the image. It's using a higher level gold standard, like it's predicting what the biopsy would say. And so now you're doing something that-
14:33 Hanne Tidnam? Which would be into your kind of novel plus-
14:36 Vijay Pande? Yeah, something no human being could do. And so, and it can do that because it in principle could fuse the data from the image and maybe blood work and other things that are easier to get than, and much less risk inducing than removing, you know, tissue in a biopsy.
14:51 Hanne Tidnam? So pulling those multiple streams of information into one, sort of synthesizing them is another area that-
14:57 Vijay Pande? It's very difficult for a human being to do and very natural for a computer.
15:00 Speaker 01 It is very natural, but I think we need a couple of things to get there. We need really dense, high quality data to train. And the more data you put in a model. I mean, so machine learning by definition is statistical overfitting. And sometimes-
15:16 Vijay Pande? Well, actually, I- I think that's wrong. That machine learning done poorly, it's like saying driving is driving a car off a cliff. Like, you know, poor driving is driving is poor driving, but machine learning tries to avoid statistical overfitting.
15:29 Speaker 01 It it does. My point is, is that you, one of the unknowns with any model, it doesn't matter if it's machine learning or regression or a risk score, is calibration. And as you start including fuzzy and noisy data elements in there, first of all, often the the the validation data sets don't perform as well as the training data set, no matter what math you use.
15:51 Hanne Tidnam? And why is that?
15:52 Vijay Pande? Well, that's that's a sign of overfitting, and usually it's because there wasn't sufficient regularization during the the training process.
15:57 Speaker 01 So overfitting is a concept in statistics to to effectively indicate that your model has been so highly tuned and specified to the data you see in front of it, that it may not apply to other cases.
16:13 Hanne Tidnam? It can't generalize.
16:14 Speaker 01 So if you had to use a model and to identify a bunch of kids in a classroom and pick the kid who's the fastest, a overfitted model might say it's the redheaded kid wearing Nikes.
16:28 Hanne Tidnam? Okay.
16:29 Speaker 01 Okay. And because in that class, that was the case.
16:31 Hanne Tidnam? That was the one child.
16:32 Speaker 01 But but that has no plausible biological or other plausibility.
16:36 Hanne Tidnam? Extrapolate, it doesn't generalize, yeah.
16:37 Speaker 01 And so if you take that to a place where the prevalence of Nike shoes or redheads is low, you might miss the fastest person.
16:45 Hanne Tidnam? Not helpful, yeah.
16:45 Speaker 01 And so these are some of the issues, the underlying shifts in population, the in natural language processing that's embedded in AI, the lexicon that people use, how doctors and clinicians write what it is that they're saying with their patient is different from not even specialty to specialty, but hospital to hospital, sort of mini subcultures.
17:05 Speaker 02 It's going to be different at Stanford than it was at UCSF, which is going to be different at Cleveland Clinic. I think that's actually a nice thing about wearable data is that fit bits are the same all over the world. This label problem, though, is is interesting because, you know, in our context, each label represents a, you know, human life at risk, right? It's a person who came into the hospital with an arrhythmia. And so you're not going to get a million labels the way you might for a computer vision application. I mean, it'd be kind of unconscionable to ask for a million labels in this case. So I think one of the interesting challenges is training deep learning base models, which tend to be very data hungry, with limited labeled data.
17:42 Speaker 01 The kitchen sink approach of taking every single data element, even if you're looking at an image, can lead to these problems of overfitting. And what Brandon and Vijay are both alluding to is you limit the labels to really high-quality labeling.
17:59 Hanne Tidnam? Right.
17:59 Speaker 01 And see if you can go there. And so don't complicate your models unnecessarily.
18:04 Vijay Pande? And don't build models that are overly complicated for the amount of data you have.
18:08 Speaker 01 Right.
18:09 Vijay Pande? Because if you have the case where you're doing so much better on the training set than the test set, that's proof that you're, you know, that you're overfitting and you're doing the ML wrong.
18:17 Speaker 02 Modern ML practitioners have a whole set of techniques to deal with overfitting. So I think that problem is very solvable with well-trained practitioners. One thing you alluded to, which is the interpretability aspect. So let's say you train on a population that's very high in diabetes, but then you're testing on a different population, which has a higher or lower prevalence. That is kind of interesting. So identifying shifts in the underlying data.
18:42 Hanne Tidnam? What would that mean?
18:44 Speaker 02 So let's say we train on people in San Francisco and, you know, everyone runs to work and eats quinoa all day. But then we go to a different part of the country where, you know, maybe obesity is higher. Or you could be somewhere in the stroke belt where the rate of stroke is higher. It may be that the statistics you trained on don't match the statistics that you're now testing on. That's fundamentally a data quality problem. If you collect data from all over the world, you can address this. But it's something you have to be careful of.
19:14 Hanne Tidnam? A while for that to happen as we start gathering the data in different ways. How does that actually even happen? How are these streams of data funneled in, examined, and fed into a useful system?
19:27 Speaker 02 So it used to be the way you'd run a clinical trial is that you would have, you know, one hospital. You'd recruit patients from that hospital. That'd be it. If you got a couple of hundred patients, that might actually be quite difficult to attain. I think with ResearchKit, HealthKit, Google Fit, all of these things, you can now get 40,000 or 50,000 people into a study from all over the world, which is great, except the challenge that the first five ResearchKit apps had is that they got... Forty thousand people and then they lost ninety percent of them in the first ninety days.
19:57 Hanne Tidnam? So everybody just drops out?
19:59 Speaker 02 Everyone just drops out because the apps, the initial versions of the apps weren't engaging.
20:04 Hanne Tidnam? Oh.
20:04 Speaker 02 So this is, it adds an interesting new dimension. As a medical researcher, you might not think about building an engaging, well-designed app, but actually you have to bring mobile design in as now a discipline that you're good at.
20:15 Hanne Tidnam? So there has to be some incentive to continue to engage.
20:19 Speaker 02 Yeah, exactly. And I mean, you need to measure cohorts the same way Instagram or Facebook or Snapchat does. So I think the teams that are starting to succeed here tend to be very interdisciplinary. They bring in the clinical aspect because you need that to choose the right problem to solve, but also design the study well so that you have convincing evidence. You need the AI aspect, but you also often need mobile design if it's a mobile study. You may need domain expertise in other areas if your data is coming from somewhere else.
20:48 Hanne Tidnam? And then it all has to be like gamified and fun to do.
20:50 Speaker 02 Yeah, yeah. Or I mean, gamification is sort of extrinsic motivation, but you can also give people intrinsic motivation, right? Giving them insights into their own health, for instance, is a pretty good way to hook people.
21:03 Hanne Tidnam? What's the system's incentives? I mean, of course, doctors want it if it makes them more accurate or to scale better. Patients want it if you can predict whether or not you're going to have a problem. How do we incentivize the system as a whole?
21:15 Speaker 01 I believe fundamentally it is going to come down to cost and scale and what willingness does a healthcare entity, whoever that may be, whether it's employer-based programs, insurer-based programs, accountable care organizations, are they going to be willing to take on risk to see the rewards of cost and scale? And so the early adopters will be ones who have taken on a little more risk.
21:37 Vijay Pande? Yeah, I think, you know, it is a challenge where the hope is in terms of value and in terms of better outcomes, but one has to prove it out and hospitals will want to see.
21:46 Speaker 01 The regulatory risk thing is being largely addressed by this new Office of Digital Health and the FDA. And they're really doing, seem much more forward-thinking about it, but there are going to be challenges that we have to solve. And I'll give you one just to get the group's input here is, should you be versioning? AI? Or do you just let it learn on the fly? And so normally, when we have firmware, hardware, software updates in regulated products, FDA-approved, there's static. They don't learn on the fly. If you keep them static, you're sort of losing the benefit of learning as you go. On the other hand, bad data could heavily bias the system and cause harm, right? So if you start learning from bad inputs that come into the system for whatever reason,
22:31 Speaker 01 you could intentionally or unintentionally, you know, cause harm. And so how do we deal with versioning and deep learning?
22:37 Vijay Pande? I mean, to just freeze the parameterizations of versioning from a computer science point of view is trivial. There's the deeper statistical question, which you just version, you could version every day, every week, every month.
22:48 Hanne Tidnam? It's when and how often?
22:49 Vijay Pande? What you want to do is, to the point that we're talking about earlier, you want to bring in new validation sets, things that it has never seen before. Because you don't want to just test each version on the same validation set, because now you're intrinsically sort of overfitting into it. And so what you always want to be doing, holding out bits of data such that you can test each version separately, because I want to make sure that they have very strict confidence that this is doing no harm and this is helping.
23:13 Hanne Tidnam? Right. It's like, we're actually introducing this whole new data set of a different kind of thing, and that's when you make new considerations.
23:20 Vijay Pande? It's data's coming in all the time, and so, you know, you just version on what came in today, version, and that's it. I mean, so it's pretty straightforward, and as you're training it.
23:28 Speaker 02 This is the way speech recognition works on Android phones, is that obviously data is coming in continuously, every time someone says, "Okay, Google," or "Hey, Siri," it's coming into either Google or Apple. But you train a model in batch, and then you test it very carefully, and then you deploy it.
23:44 Hanne Tidnam? Right.
23:45 Speaker 02 And the versions are indeed tagged by the date of the training data.
23:47 Hanne Tidnam? It's already embedded in the system. Who are the decision makers that are kind of green-lighting when like, "Okay, we're going to try this new algorithm, we're going to start applying this to these, you know, radiology images." What are the decision points?
24:00 Speaker 01 So with EKGs, the early companies used expert systems to just ease the pain points of me having to write out and code out every single diagnoses.
24:11 Hanne Tidnam? The super-low-hanging fruit of...
24:12 Speaker 01 Yeah. Can you improve the accuracy of physicians with this? Can you increase their volume and bandwidth? Can you actually use it to see which physicians are maybe going off course? Right? What if you start having a couple physicians whose error rates are going up? Right now with quality, the QI process isn't really based on random sampling. There's actually no standardized metrics for QI in any of this. When people read EKGs and sign them off, they just sign them. There's nothing telling anyone that this guy has a high error rate. And so that is a great use case of this, where you're not making diagnoses, but you're helping anchor and show that, well, if you believe this algorithm is good
24:56 Speaker 01 and broadly generalizable across data, you're sort of restating the calibration problem now. It's not that the algorithm has gotten necessarily worse, because in fact, in seven of the eight doctors, it's right on par with them. But in this other doctor, it could be that that doctor, if that doctor is not agreeing with the algorithm, which is agreeing with the other seven, that doctor is actually not agreeing with the other seven. And so now you have an opportunity to train and relearn. Those are the use cases to go off. And train and relearn the person? The person. Address their reading errors, coding errors, see what's going on. And that qualitative look, I think, is very, very valuable.
25:33 Hanne Tidnam? So what are the ways we're actually going to start seeing it in the clinical setting? You know, the tools that we might see our doctor actually use with us or not?
25:41 Speaker 01 I think it's going to be these adjacencies around treatment with management. There are a lot of things that happen in the hospital that seem really primitive and arcane, and no one wants to do them. And I'll give you a simple one, which is OR scheduling.
25:53 Hanne Tidnam? So is it actually the way it looks like it is in Grey's Anatomy? Is it just a whiteboard?
25:57 Speaker 01 It's a whiteboard and somebody at the phone, the OR front desk.
26:00 Hanne Tidnam? That's unbelievable.
26:01 Speaker 01 There's a backend of scheduling that happens for the outpatient. But you have add-ons, you have emergencies, you have finite...
26:05 Hanne Tidnam? I mean, it seems like even an Excel sheet would be better than a whiteboard.
26:08 Speaker 01 The way OR scheduling works now is primitive, and it also involves gaming. It involves convincing staff X, Y, and Z to stay to get the case done or do it tomorrow.
26:20 Hanne Tidnam? So there's so much behind the scenes, like human negotiation?
26:23 Speaker 01 For example, when I do catheter ablations, we have many different moving parts, equipment, the support personnel. Of the equipment manufacturer, anesthesia, fellows, nurses, whatever. And everyone has little pieces of that scheduling. And it all comes together, but it comes together in the in the art of human negotiation. And very simple things like, well, this is your block time. And if you want to go outside your block time, you you know, you need to write a hallmark card to person X. And so very simple problem where there's huge returns in efficiency. If you can have AI do that, right? And and the AI inputs over time could be like, well, you can really truly know which physicians are quick and speedy.
27:08 Speaker 01 Which ones tend to go over there a lot of times? Which patient cases might be high risk? Which ones may need more backup? Which should be done during daytime hours?
27:17 Speaker 02 You could add their Fitbit data and then you could tell who's drowsy at any given moment.
27:21 Hanne Tidnam? Oh, that's fascinating. Yeah. Whether or not they want to do it.
27:25 Speaker 02 How stressed are they feeling?
27:27 Speaker 01 And so people stay at times that they're really needed. And that kind of elasticity can come with automation where we fail right now. And so this is a great place where you are not making diagnoses. There's nothing you're being committed to by from a kind of, you know, basic regulatory framework. You're just optimizing scheduling.
27:49 Hanne Tidnam? So who actually says, so say that that technology is available? And how do you actually get it in? You know, where's the sort of confluence of the regulation and the actual rollout? And how does it actually make its way into a hospital, into a waiting room?
28:01 Speaker 02 There's an alternative model I've seen, which is that startups acting as full stack healthcare providers. So Omada Health or Verta Health would be examples of this, where if you have prediabetes or diabetes, respectively, the physician can actually refer the patient to one of these services. They have on staff physicians. They're registered as providers with national provider IDs. They bill insurance, just like a doctor would. And they're essentially acting as a provider who addresses the whole condition end to end. I think that case actually simplifies decision making because you don't necessarily have to convince both. You know, Stanford and United Healthcare to adopt this thing. You can actually convince just a self insured employer that they want to include one of these startups as part of their health plan.
28:46 Speaker 02 And so I think that simplifies the decision making process. It ensures that the physicians and the AI folks are under the same roof. I think that's a model that we're going to see probably get the quickest adoption.
28:57 Hanne Tidnam? Right.
28:58 Speaker 02 At least in the payer world.
28:58 Vijay Pande? There are many models and which is the best model will depend on how you're helping and the indication and on the accuracy and what you're competing against and so on. So.
29:07 Speaker 02 This is a case where probably we'll see the healthcare industry maybe reconstitute itself by vertical with AI based diagnostics or therapeutics. Because if you think right now, providers are geographically structured. But with AI, every data point makes the system more accurate. Presumably in an AI based world, providers will be more oriented around a particular vertical. So you might have the best data network in radiology, the best data network in pathology.
29:32 Hanne Tidnam? Oh, that's interesting. Yeah. Thank you so much for joining us on the A16Z podcast.
29:37 Vijay Pande? Great, thank you.
29:37 Speaker 02 Thank you. Thanks for having us.
Previous Company Building in Crypto: A Conversation between Brian Armstrong and Chris Dixon Next What's Next for Education Startups Page updated October 26, 2026 · transcript and summary are machine-generated