Latest / What's Your Problem? / Going to the Doctor Sucks. Can AI Make it Better?

Transcript

Going to the Doctor Sucks. Can AI Make it Better? · What's Your Problem?

• 0:15
  Speaker 1: Pushkin.
• 0:20
  Speaker 2: Going to the doctor sucks, and not just because you're sick,
• 0:24
  Speaker 2: but because it's hard to make an appointment. Then you
• 0:27
  Speaker 2: finally get an appointment, and you go in and you
• 0:29
  Speaker 2: fill out all these forms. Then you wait for whatever
• 0:32
  Speaker 2: an hour. Then you go in and see the doctor
• 0:35
  Speaker 2: and she asks you all the questions you just answered
• 0:38
  Speaker 2: on those forms.
• 0:39
  Speaker 1: And by the way, it's not great for the doctor either.
• 0:42
  Speaker 2: She has to see a million patients and do all
• 0:44
  Speaker 2: this paperwork and doesn't have time to keep up with
• 0:46
  Speaker 2: what's happening in the medical literature with all the new
• 0:49
  Speaker 2: studies and recommendations.
• 0:51
  Speaker 1: And everybody knows this is true.
• 0:53
  Speaker 2: Nobody's figured out how to fix it yet, but it
• 0:55
  Speaker 2: is a problem that people are.
• 0:57
  Speaker 1: Trying to solve. I'm Jacob Goldstein and this.
• 1:05
  Speaker 2: Is What's Your Problem, the show where I talk to
• 1:07
  Speaker 2: people who are trying to make technol logical progress. My
• 1:10
  Speaker 2: guest today is a lawn block. Alan is a serial entrepreneur.
• 1:15
  Speaker 3: He was the co CEO of Wix, the site for
• 1:17
  Speaker 3: building websites. He founded a company called Room that buys
• 1:21
  Speaker 3: and sells used cars, and then in twenty sixteen Alan
• 1:25
  Speaker 3: founded a company called k Health.
• 1:27
  Speaker 1: Today, thousands of patients a month are treated through k Health.
• 1:31
  Speaker 2: The company has this AI based patient interface and they
• 1:34
  Speaker 2: employ around one hundred.
• 1:35
  Speaker 3: And fifty doctors through an affiliate, and k help has
• 1:39
  Speaker 3: big growth plans. They just raised another fifty million.
• 1:42
  Speaker 1: Dollars to keep growing.
• 1:44
  Speaker 2: The problem Alan and k help are trying to solve
• 1:47
  Speaker 2: is this, can you use data and technology to make
• 1:50
  Speaker 2: going to the doctor better? Better for patients and for doctors.
• 1:56
  Speaker 4: We built essentially a machine that can engage with you.
• 2:00
  Speaker 4: It's a predictive mention that is trying to figure out
• 2:03
  Speaker 4: the next best question to ask Jacob. Jacob comes in
• 2:07
  Speaker 4: and he's got a headache, and and and you know,
• 2:12
  Speaker 4: if we don't know anything about you, we asked for
• 2:14
  Speaker 4: your age and your gender, and we have a series
• 2:17
  Speaker 4: of conversations given what we know, you know, you know,
• 2:21
  Speaker 4: what is the next best question to us to get
• 2:23
  Speaker 4: to the differential diagnosis?
• 2:25
  Speaker 3: I will say, uh, this is in fact some some
• 2:30
  Speaker 3: part of your product as it exists now is is
• 2:32
  Speaker 3: what you're describing. And I know this because I tried
• 2:36
  Speaker 3: it out before this interview, and as it happens, true story,
• 2:40
  Speaker 3: I have COVID right now. Despite my chipper demeanor, I
• 2:44
  Speaker 3: feel like asks. And so yesterday I loved Indo K
• 2:49
  Speaker 3: Health And the one thing I didn't tell it, I
• 2:52
  Speaker 3: didn't I have had.
• 2:53
  Speaker 2: A positive COVID test. But I thought it would be
• 2:55
  Speaker 2: too easy to tell it that I had a positive
• 2:56
  Speaker 2: COVID test. But I said I have. I'm feeling tired?
• 3:01
  Speaker 2: Was my ask you?
• 3:02
  Speaker 3: You log in and ask you first, like you know
• 3:04
  Speaker 3: you're presenting complain, asked like why are you here today?
• 3:07
  Speaker 3: What seems to be the trouble? And I started with tired.
• 3:09
  Speaker 3: And it has a specific list of possible answers. Right,
• 3:13
  Speaker 3: it's not like it's not like chat GPT, it's not
• 3:16
  Speaker 3: free form, right, it's very it's clearly categorized, and it
• 3:20
  Speaker 3: puts your answer into keating. And so once I said
• 3:22
  Speaker 3: I was tired, it asked me a lot of other
• 3:24
  Speaker 3: things that said, when did you start a fatigue? Right,
• 3:27
  Speaker 3: it called it fatigue, and then it's how long have
• 3:29
  Speaker 3: you been feeling fatigue? Lots of questions about fatigue any,
• 3:32
  Speaker 3: did your fatigue start after eating, after sun exposure?
• 3:36
  Speaker 2: Are you more or less fatigued at the time of day?
• 3:37
  Speaker 3: And then it asked me other symptoms that I said,
• 3:40
  Speaker 3: I had a fever and It asked me lots of
• 3:42
  Speaker 3: specific questions including have you been in close contact with
• 3:45
  Speaker 3: anyone who has COVID? Lots of other symptoms, you know,
• 3:48
  Speaker 3: loss of smell, no malaise, yes, lots of general questions,
• 3:53
  Speaker 3: and then it said the results I'm about to show
• 3:55
  Speaker 3: you are not a diagnosis or medical advice, which I
• 3:59
  Speaker 3: get that. And then and then it gave me an
• 4:01
  Speaker 3: interesting It gave me a probabilistic answer, right, which was
• 4:06
  Speaker 3: quite interesting to me. The sort of final screen, and
• 4:08
  Speaker 3: I got a screen of I'll just read it to
• 4:09
  Speaker 3: you now.
• 4:10
  Speaker 2: It says what your symptoms may mean, and then it says,
• 4:13
  Speaker 2: here's how eleven thousand, eight hundred and eighty one people
• 4:17
  Speaker 2: with cases like yours were diagnosed sixty percent COVID, nineteen
• 4:21
  Speaker 2: forty percent upper respiratory infection. And then it said think
• 4:25
  Speaker 2: you needed doctor, says it get better from home today,
• 4:27
  Speaker 2: chat with a clinician now, and that's we can talk
• 4:29
  Speaker 2: about that part later, but let's just talk about I'm
• 4:32
  Speaker 2: really interested in that, the specific quantitative nature the eleven thousand,
• 4:37
  Speaker 2: eight hundred and eighty one people with cases like yours,
• 4:40
  Speaker 2: and then the probabilistic outcome. That's really interesting to me
• 4:44
  Speaker 2: because it's sort of like what would.
• 4:46
  Speaker 3: Happen if you go to the doctor, but also sort
• 4:48
  Speaker 3: of different. So tell me about that.
• 4:51
  Speaker 4: So first of all, that makes me ready happy and
• 4:53
  Speaker 4: we should have started there. I'm so happy you used it.
• 4:56
  Speaker 4: And by the way, coming in with the fatigue case
• 4:59
  Speaker 4: is not trivial. You go to your primary care or
• 5:02
  Speaker 4: good care and say I have fatigue. Okay, there's many
• 5:06
  Speaker 4: many things that can lead to fatigue. It can be temporary,
• 5:08
  Speaker 4: can be long term term. It could it could it
• 5:11
  Speaker 4: can be related to a serious health condition, or it
• 5:13
  Speaker 4: can be something transitory. So that's great. You came there
• 5:18
  Speaker 4: and you acted all down. Apologies, but you didn't say
• 5:21
  Speaker 4: I have COVID door I think I have COVID dours effected.
• 5:24
  Speaker 4: You gave it was a broad based thing. Fatigue can
• 5:26
  Speaker 4: also be related to cancer. It can be related to
• 5:29
  Speaker 4: thyroid issues, insomnia. So so what what we're trying to
• 5:33
  Speaker 4: do just to tile the pieces together. It builds a
• 5:37
  Speaker 4: dynamic cluster of people like you, of this, of this,
• 5:42
  Speaker 4: of these people that have the same gender, age, medical
• 5:45
  Speaker 4: history and symptoms that are relevant to what you're having,
• 5:48
  Speaker 4: and it's trying to figure out what it has it
• 5:51
  Speaker 4: gives you a statistical output because it's a probabilistic nature
• 5:55
  Speaker 4: of what you might have. And it's and as as
• 6:00
  Speaker 4: as we say very clearly we don't diagnose you, but
• 6:02
  Speaker 4: compare it to doctor Google or any other system out there,
• 6:06
  Speaker 4: you won't get that rigger. And every one of the
• 6:08
  Speaker 4: questions we ask in our clinical AI is a is
• 6:13
  Speaker 4: a very specific question. Is in medicine, this is not
• 6:16
  Speaker 4: a hallucinating let me just try and guess what you have.
• 6:20
  Speaker 4: It's not chat ept, it's not trying to predict some
• 6:25
  Speaker 4: set of sentences. It's trying to predict your diagnosis. We
• 6:28
  Speaker 4: generate a diagnosis in treatment, right, but because you know
• 6:33
  Speaker 4: this is not regulated, we tell people this is what
• 6:36
  Speaker 4: you might have, and if you want to, you can
• 6:38
  Speaker 4: press about it and be in front of a doctor
• 6:39
  Speaker 4: twenty for seven and deal if not.
• 6:41
  Speaker 2: In fact, it says it is not a diagnosis. You're
• 6:44
  Speaker 2: saying it's a diagnosis.
• 6:47
  Speaker 4: Trying to make Jesus to be clear, this is not
• 6:49
  Speaker 4: a diagnosis. Is trying to mimic a diagnosis. What we're
• 6:53
  Speaker 4: trying to generate is trying to mimic a really high
• 6:57
  Speaker 4: quality physician by having this medical conversation, this medical chat
• 7:01
  Speaker 4: and having showing the differential because maybe if you didn't
• 7:06
  Speaker 4: have a positive laptist, maybe it's maybe it's an opera
• 7:10
  Speaker 4: respiratory infection, maybe it's something else going on here. And
• 7:14
  Speaker 4: our goal was not to say, Okay, this is what
• 7:17
  Speaker 4: you have and it's impossible that it's something else, because
• 7:20
  Speaker 4: that's not how medicine works. The goal is to consider
• 7:22
  Speaker 4: the diagnosis. But that's a tricky part of medicine. There
• 7:26
  Speaker 4: is a science and art here about how to do it,
• 7:28
  Speaker 4: and we're trying to mimic that capability. But in it
• 7:32
  Speaker 4: I think it's quite magical that you can engage with
• 7:35
  Speaker 4: a machine, have an understanding what you might have and
• 7:40
  Speaker 4: which is not some kind of guess, work on doctor Google,
• 7:43
  Speaker 4: and then press a button and was in a matter
• 7:44
  Speaker 4: of minutes and be in front of a doctor and
• 7:47
  Speaker 4: get medical advice and get prescription, get anything else, anything
• 7:51
  Speaker 4: else you need twenty four to seven.
• 7:53
  Speaker 2: So to build this thing, this machine that's trying to
• 7:57
  Speaker 2: come up with, you know, something like a differential diagnosis,
• 8:00
  Speaker 2: obviously you need a ton of data. And I know
• 8:03
  Speaker 2: there's this key moment early in the life of the
• 8:06
  Speaker 2: company when you got access to this like big beautiful
• 8:10
  Speaker 2: data set. Tell me about that, Okay, So.
• 8:15
  Speaker 4: People pointed me all over the world to try and
• 8:17
  Speaker 4: find data. At the time people spoke to me about Estonia, Denmark.
• 8:22
  Speaker 4: Any conversation I try to have with a with a
• 8:26
  Speaker 4: large hospital group in America at the time was completely stonewalled.
• 8:29
  Speaker 4: I can get anywhere because of hip et cetera. And
• 8:33
  Speaker 4: we ended up licensing a data set from an HMO
• 8:36
  Speaker 4: in Israel called Maccabi. And I'll explained specifically what we
• 8:39
  Speaker 4: have there. At the time, we didn't realize just how
• 8:43
  Speaker 4: what do you call it? Big and beautiful? The data
• 8:44
  Speaker 4: set is Maccabi is structured similar to Kaiser Permanenta. So
• 8:50
  Speaker 4: it's an insurance company that also has its own captive providers,
• 8:54
  Speaker 4: but they also have their own drugstore, their own lab,
• 8:57
  Speaker 4: a single EMR, their own hospital.
• 8:59
  Speaker 2: If providers means doctors who work for the company. Just
• 9:02
  Speaker 2: to be right, okay. So it's a complete integrated health system, right.
• 9:06
  Speaker 3: People who are in this system, they get their prescriptions there,
• 9:09
  Speaker 3: they go see the doctor there, they go to the
• 9:11
  Speaker 3: hospital there. So this Macabi, this Israel, is going to
• 9:16
  Speaker 3: have all of the data for the patients there.
• 9:18
  Speaker 4: And then one more big step. People don't switch HMOs
• 9:23
  Speaker 4: in Israel. So we essentially got access to two and
• 9:26
  Speaker 4: a quarter million people for a data set over twenty years,
• 9:30
  Speaker 4: and we essentially were a fly on the wall for
• 9:32
  Speaker 4: all these people. Did they go to a doctor, did
• 9:34
  Speaker 4: they get a prescription? We know also if they picked
• 9:36
  Speaker 4: it up, we know, if they had labs, we know.
• 9:39
  Speaker 4: We could put together a time series of all of that,
• 9:41
  Speaker 4: and of course postilizations as procedures, surgery's outcomes. So just
• 9:45
  Speaker 4: to give you a sense, we got access to about
• 9:48
  Speaker 4: four hundred million medical charts, over a billion labs, half
• 9:52
  Speaker 4: a billion prescriptions, two million hospitalizations.
• 9:57
  Speaker 2: And I presume it's anonymized.
• 9:59
  Speaker 4: We built an anonymizer.
• 10:02
  Speaker 3: So you have this very large, very robust, very complete,
• 10:06
  Speaker 3: longitudinal data set.
• 10:08
  Speaker 2: Great, what do you do with it?
• 10:10
  Speaker 4: We wanted to focus on these things are both what
• 10:13
  Speaker 4: you go to your doctor, your primary care physician, and
• 10:16
  Speaker 4: to urgent care. My head hurts, my stomach hurts, I've
• 10:19
  Speaker 4: got a rash, I've got an upper rest with our infection,
• 10:23
  Speaker 4: I've got a fever. All these different things.
• 10:26
  Speaker 2: And some kind of problem that just started. It's not
• 10:29
  Speaker 2: a chronic disease.
• 10:30
  Speaker 3: It's not a physical it's something that like, oh, my
• 10:32
  Speaker 3: head just started hurting, I better go to the doctor.
• 10:35
  Speaker 4: It's that, yes, And now it could be my head
• 10:38
  Speaker 4: started hurting and I've got a history of migrains a synoscientists,
• 10:41
  Speaker 4: and I think I know what it is, or it
• 10:43
  Speaker 4: might be something completely out of the blue.
• 10:45
  Speaker 2: What's one thing you tried at k Health over the
• 10:49
  Speaker 2: course of the company that didn't work.
• 10:52
  Speaker 4: We tried in the beginning as we were building through
• 10:55
  Speaker 4: as we're learning healthcare, we're trying to send people. When
• 10:59
  Speaker 4: we built the information, we said, okay, here's the information
• 11:02
  Speaker 4: and we'll connect you to physical doctors in person in
• 11:05
  Speaker 4: Manhattan and Brooklyn because we're basing the arc and then
• 11:09
  Speaker 4: work at all.
• 11:10
  Speaker 2: Why not? So why not?
• 11:12
  Speaker 4: Because people wanted immediate solutions of their problems right now,
• 11:15
  Speaker 4: that's a scientist or UTI or migraine. And you know
• 11:19
  Speaker 4: the whole beauty was it twenty four to seven service?
• 11:21
  Speaker 2: So was that when you realized that you had to
• 11:23
  Speaker 2: hire doctors.
• 11:25
  Speaker 4: We needed to hire doctors and build our own solution,
• 11:28
  Speaker 4: own the resolution.
• 11:29
  Speaker 2: Which is a huge that's a huge leap right there.
• 11:31
  Speaker 2: You're going from being a software company to being like
• 11:35
  Speaker 2: a much more complicated, much more expensive to build company
• 11:39
  Speaker 2: at that moment.
• 11:40
  Speaker 4: Right, Yeah, but you we didn't go in here, timid.
• 11:44
  Speaker 4: We went in here. We want to make real changes.
• 11:47
  Speaker 4: We don't want to be some yet another cute digital
• 11:50
  Speaker 4: health company that does cute things and sells usself. We
• 11:53
  Speaker 4: want to build a big, independent company, and we want
• 11:55
  Speaker 4: to build a company that will make a difference in
• 11:59
  Speaker 4: care delivery in people's lives, starting was primary care. So yeah,
• 12:03
  Speaker 4: you have to you have to pick your battles.
• 12:05
  Speaker 2: Yeah.
• 12:06
  Speaker 3: Well, so let me ask since that came up, I mean, so,
• 12:08
  Speaker 3: I know you're just raised another fifty million dollars, which congratulations.
• 12:13
  Speaker 3: Hard to do at this moment from what I've heard
• 12:15
  Speaker 3: for a lot of companies, and you've raised what three
• 12:18
  Speaker 3: hundred million or so so far? Is there some Well,
• 12:23
  Speaker 3: do you have a sense of when you'll be profitable?
• 12:26
  Speaker 4: I think we'll be profitable in the second half of
• 12:28
  Speaker 4: next year.
• 12:30
  Speaker 2: Great, So let's let's talk more more fully about how
• 12:36
  Speaker 2: your business works at this point, right, I have I
• 12:40
  Speaker 2: tried out this sort of consumer facing first step, I
• 12:43
  Speaker 2: didn't click through to talk with a clinician, but what like,
• 12:48
  Speaker 2: how does the business work now?
• 12:51
  Speaker 4: So our business is wet. We have both a direct
• 12:55
  Speaker 4: to consumer relationships and we work with payers, and we
• 13:01
  Speaker 4: work with large health system providers.
• 13:04
  Speaker 2: And payers means insurance companies. Payers is what people are
• 13:07
  Speaker 2: in healthcare call insurance companies.
• 13:09
  Speaker 4: So we essentially offer our services and we get paid
• 13:13
  Speaker 4: as a provider to take care of patients. So if
• 13:16
  Speaker 4: it's direct to consumer, we have an offer for consumers
• 13:19
  Speaker 4: and you can just sign up and pay out of pocket.
• 13:22
  Speaker 4: We announced that a.
• 13:23
  Speaker 2: Few weeks ago that we're.
• 13:25
  Speaker 4: Working with Cedar Sinai. We're working with other health systems.
• 13:28
  Speaker 2: Theesar sign up a big hospital system in Los Angeles.
• 13:31
  Speaker 4: Yeah, the largest hospital system in California. We see thousands
• 13:35
  Speaker 4: of patients every day, which would be the size of
• 13:39
  Speaker 4: a large health system with multiple hospitals and multiple art
• 13:43
  Speaker 4: patient centers.
• 13:44
  Speaker 2: So large, and how often do you end up referring
• 13:48
  Speaker 2: people to see a doctor in the physical world, or
• 13:53
  Speaker 2: get a test in the physical world, to do something
• 13:55
  Speaker 2: in real life, to do something in meat space.
• 13:58
  Speaker 4: So we refer people to in person physician about eight
• 14:04
  Speaker 4: percent of the time. That could be to the er,
• 14:07
  Speaker 4: it could be to specialty care or to an person
• 14:09
  Speaker 4: primary care. Okay, so I was going to say that,
• 14:13
  Speaker 4: Just to be clear, there is many times a physical interaction.
• 14:18
  Speaker 4: For example, our doctors will send a patient to do ELEB. Yeah,
• 14:22
  Speaker 4: but ninety two percent of the time we can resolve
• 14:24
  Speaker 4: it without a bricks and marker doctor reviewing that lab.
• 14:29
  Speaker 4: A portion of the time we need to we need
• 14:32
  Speaker 4: to refer the patient to an in person visit in
• 14:34
  Speaker 4: primary care, or to specialty care or to er. But again,
• 14:40
  Speaker 4: the vast majority of stuff we can do online. So
• 14:42
  Speaker 4: it completely changes the paradigm on access because if it's
• 14:47
  Speaker 4: ten pm, what are your choices?
• 14:51
  Speaker 3: In a minute, Alan talks about the problems that k
• 14:53
  Speaker 3: Health is still trying to solve. It's interesting to think
• 15:06
  Speaker 3: about to compare k help Health versus a traditional doctor.
• 15:11
  Speaker 3: Say good, you know, seventieth percentile a good but not
• 15:14
  Speaker 3: amazing traditional doctor.
• 15:17
  Speaker 2: And I'm curious. I would imagine there are some settings
• 15:21
  Speaker 2: where KA health is better and some settings where a
• 15:23
  Speaker 2: traditional doctor might be better. Some things that are just
• 15:25
  Speaker 2: easier to detect in person, say, and I'm curious, like,
• 15:31
  Speaker 2: are there particular kinds of cases that worry you? Are
• 15:35
  Speaker 2: there particular kinds of patients who you think KA health
• 15:39
  Speaker 2: may not be well well suited for.
• 15:42
  Speaker 4: I think the same things that doctors will struggle with
• 15:47
  Speaker 4: online are things that they will struggle with in person.
• 15:52
  Speaker 4: Patients that are not compliant with their medications or with
• 15:56
  Speaker 4: their care protocols. Patients that have are just very fragile,
• 16:01
  Speaker 4: they're very old, they have you know, they have pneumonia,
• 16:08
  Speaker 4: but they also to have multiple other background diseases. Remember
• 16:12
  Speaker 4: that when you go into a primary care doctor or
• 16:14
  Speaker 4: clinic and you walk out you're ass it's very difficult
• 16:18
  Speaker 4: for the doctor to interact with you and follow up
• 16:20
  Speaker 4: with you. So there's multiple advantages actually to build a
• 16:24
  Speaker 4: system online. People can come back very easily to our platform.
• 16:28
  Speaker 3: I mean, I've asked, is there somewhere where KA health
• 16:31
  Speaker 3: is worse? And you've given me somewhere where KA health
• 16:33
  Speaker 3: is better. But I am curious within the universe where
• 16:38
  Speaker 3: you were treating people, are there places where for whatever
• 16:42
  Speaker 3: then maybe there's a smell, maybe there's some you know,
• 16:45
  Speaker 3: truly right, like there are things that can be perceived
• 16:48
  Speaker 3: in the physical world.
• 16:51
  Speaker 2: It's a genuine question.
• 16:52
  Speaker 4: No, No, I understand the question. I think the areas
• 16:55
  Speaker 4: that we're less strong and now is areas that we've
• 16:58
  Speaker 4: seen for your patients. It's stuff a little bit different
• 17:01
  Speaker 4: from what your intuition is. Your intuition was similar to mine,
• 17:04
  Speaker 4: you know, six seven years ago. But it's stuff around
• 17:07
  Speaker 4: or simpedic huh, back pain, extremity, medicine, you know. Uh,
• 17:14
  Speaker 4: the trauma. We don't do trauma, you know, so so
• 17:18
  Speaker 4: stuff around.
• 17:18
  Speaker 2: That trauma trauma like did I break my wrist when
• 17:22
  Speaker 2: I fell that?
• 17:23
  Speaker 4: Or you've got a pain in your in your in
• 17:25
  Speaker 4: your foot for a few days. Maybe it's planter fasciaetus
• 17:30
  Speaker 4: if you've ever had it, it's pretty painful and appears
• 17:34
  Speaker 4: out of the blue. But you know, maybe maybe you
• 17:37
  Speaker 4: broke a bone or maybe you've got you know, you know.
• 17:41
  Speaker 3: Is the reason is the reason you don't do that,
• 17:43
  Speaker 3: because the doctor needs to touch the patient in short
• 17:45
  Speaker 3: to make that diagnosi.
• 17:47
  Speaker 4: I think we can build systems for orthopedics as well. Uh,
• 17:51
  Speaker 4: it's just not it hasn't been an error focus about
• 17:53
  Speaker 4: for us, and and over time, I think the whole
• 17:57
  Speaker 4: online versus in person is going to be mute. It's
• 18:00
  Speaker 4: kind of like going to the ATM to get out
• 18:02
  Speaker 4: cash versus venmo. Both are utilities of of getting cash
• 18:07
  Speaker 4: or transferring cash. You will use both interchangeably in different
• 18:12
  Speaker 4: reasons for different settings. I think that's how medicine's gonna
• 18:15
  Speaker 4: go to. If you can resolve your problem right now
• 18:19
  Speaker 4: from home in a high quality, high confidence environment, and
• 18:22
  Speaker 4: I can follow up with you wherever you are. Maybe
• 18:24
  Speaker 4: you need a prescription, maybe not. It's typically better than
• 18:29
  Speaker 4: going in in person. Sometimes you need to go in
• 18:31
  Speaker 4: person for various things. Medicine is really tricky. It's really
• 18:34
  Speaker 4: tough to I'm still at all of how doctors make
• 18:38
  Speaker 4: decisions because they don't always have all the facts, but
• 18:42
  Speaker 4: yet they need to act. They need to decide should
• 18:43
  Speaker 4: I send the patient to the R do I tell
• 18:45
  Speaker 4: them to rest at home? I would just jacb I'm
• 18:48
  Speaker 4: jumping into something to me is ready burning and important
• 18:51
  Speaker 4: for us. We built k to give people access to
• 18:55
  Speaker 4: high quality medicine. I think we have uneven quality medicine. Yes,
• 19:00
  Speaker 4: there's Mayo Clinic and Cleveland Clinic and top academic institutions
• 19:05
  Speaker 4: that act as lost results or solving complex care Cedar
• 19:08
  Speaker 4: sin I other institutions, But most of the time people
• 19:11
  Speaker 4: need to manage their diabetes and hyppertension, or even better
• 19:14
  Speaker 4: avoid getting diabetes talked to or high pretension. Many people
• 19:18
  Speaker 4: who have insurance and have money do not manage their
• 19:21
  Speaker 4: diabetes and hard conditions. Well, it's not one or two percent,
• 19:25
  Speaker 4: it's double digit percent. Why Because it's tough, it's it's expensive,
• 19:30
  Speaker 4: it's messy that you know who wants to deal with it?
• 19:33
  Speaker 4: Can we build care delivery systems that are dramatically better,
• 19:38
  Speaker 4: that are much more data driven and are tailored to
• 19:40
  Speaker 4: you admin that take into account prevention. Everybody talks about
• 19:45
  Speaker 4: prevention that are preventative systems are similar to for the
• 19:49
  Speaker 4: most part, I'm not talking about oncology here to the
• 19:52
  Speaker 4: seventies and the sixties. That's pretty awful.
• 19:54
  Speaker 1: So let's this is interesting.
• 19:56
  Speaker 2: Let's talk about it at a at a slightly more
• 19:58
  Speaker 2: specific level, Like you're making a compelling kind of abstract,
• 20:02
  Speaker 2: big picture society wide point. Tell me something specific.
• 20:07
  Speaker 3: About what you're doing to do a better verse of
• 20:10
  Speaker 3: helping patients manage their diabetes or high blood pressure.
• 20:14
  Speaker 4: I'll give you one specific example. We took maoclinic data
• 20:18
  Speaker 4: and we're able to look at mayoclinic patients in an
• 20:22
  Speaker 4: anonymized manner over long periods of time, and look at
• 20:25
  Speaker 4: high pertension patients and figure out the best medication or
• 20:29
  Speaker 4: combination of medicines to stabilize high blood pressure given their
• 20:34
  Speaker 4: current blood pressure, gender, age, comorbidity, and ethnicity, and be
• 20:39
  Speaker 4: able to tell a doctor in that same statistical split
• 20:43
  Speaker 4: here is potentially the best way to give this medicine,
• 20:47
  Speaker 4: and then math medicine.
• 20:49
  Speaker 3: And high blood pressure is presumably a good place to
• 20:51
  Speaker 3: do that, because there are lots and lots of good, cheap,
• 20:56
  Speaker 3: old high blood pressure medicines, but which ones to use
• 20:59
  Speaker 3: or which combination to use for which.
• 21:02
  Speaker 2: Patient is hard to figure out. Presumably that's why that's
• 21:05
  Speaker 2: a good sort of target problem.
• 21:06
  Speaker 4: I would venture to say most doctors do not in
• 21:09
  Speaker 4: the in their head that algorithm of gender, age, medical history, ethnicity, culpabidity,
• 21:15
  Speaker 4: and you know, the current blood pressure. So this is
• 21:17
  Speaker 4: really helpful to our doctors. Eight percent of the time
• 21:20
  Speaker 4: they change the patient's regimen or change change their mind
• 21:23
  Speaker 4: around what they want to do. And it's based on
• 21:25
  Speaker 4: male clinical mail clinics standard of care. So that's that's
• 21:30
  Speaker 4: really powerful. Every day that you stabilize in better, you know,
• 21:33
  Speaker 4: to reduce the risk of strokes and heart attacks.
• 21:37
  Speaker 3: What's something you haven't figured out yet that you're still
• 21:41
  Speaker 3: working on. What's what's the sort of frontier problem, a
• 21:44
  Speaker 3: frontier problem that you're working on.
• 21:46
  Speaker 4: Hmm. The stuff that we're working on integrating right now
• 21:52
  Speaker 4: is tying what we're doing to a non hallucinating large
• 21:57
  Speaker 4: language model. So how do you take what we're doing
• 22:01
  Speaker 4: and we we help the patient understand what they might have,
• 22:06
  Speaker 4: We put them in front of a doctor and they
• 22:08
  Speaker 4: get diagnosed and tree. But let's assume the patient is
• 22:11
  Speaker 4: a diabetic patient who also wants to lose ten pounds
• 22:14
  Speaker 4: and doesn't like fish and wants to do a certain diet,
• 22:17
  Speaker 4: and it needs to be tailored to what we do.
• 22:19
  Speaker 4: How do I enable the patient in a free form,
• 22:23
  Speaker 4: free text engage with the system that's non hallucinating, that
• 22:27
  Speaker 4: is also tied directly to what we're doing. I think
• 22:29
  Speaker 4: that's a direction we go.
• 22:31
  Speaker 3: So plainly, what you have right now is very structured.
• 22:33
  Speaker 3: I can't just come in and like say a million
• 22:35
  Speaker 3: things about myself. I can have a presenting complaint and
• 22:37
  Speaker 3: then your system will ask me some questions, and there's
• 22:40
  Speaker 3: a sort of set universe of categories of conditions whatever.
• 22:46
  Speaker 2: And plainly, large language models like chet GPT, which is
• 22:49
  Speaker 2: very different from your system, offer a much more wide
• 22:51
  Speaker 2: open playing field. A wide open means of interaction between
• 22:56
  Speaker 2: a human and machine, with the notable downside, the profound
• 22:59
  Speaker 2: downside in this case, that they are not reliable sources
• 23:02
  Speaker 2: of information.
• 23:03
  Speaker 3: Right, So what if I understand you correctly, what you
• 23:05
  Speaker 3: want is the reliability of the system you have built.
• 23:08
  Speaker 2: With the freedom of a large language model like chat chip.
• 23:11
  Speaker 4: Yeah. So fundamentally, nobody wants to go to a doctor
• 23:14
  Speaker 4: that hallucinates ten percent of the time and you don't
• 23:16
  Speaker 4: know when the doctor's almost think that's kind of meaningless.
• 23:18
  Speaker 4: It's going back to the same doctor Google WebMD problem.
• 23:22
  Speaker 4: We care about the clinical context and that is really important.
• 23:26
  Speaker 4: I think that's where I's going, and I think LLLM
• 23:30
  Speaker 4: could be a huge ally here. But fundamentally, when you're
• 23:33
  Speaker 4: asking people specific questions, when you're asking questions that are
• 23:37
  Speaker 4: medical and clinical by nature, there's no guesswork here. It
• 23:40
  Speaker 4: might need to be modified and change over time, but
• 23:42
  Speaker 4: you're asking very specific questions. So for me, a big
• 23:45
  Speaker 4: part of what we're answer is integrating that.
• 23:50
  Speaker 2: We'll be back in a minute with the lightning round.
• 24:04
  Speaker 2: You were the co founder and CEO of Room, a
• 24:07
  Speaker 2: used car retail.
• 24:08
  Speaker 4: Retail, and oh yeah, I.
• 24:10
  Speaker 3: Definitely have a couple of used car questions for you. One,
• 24:13
  Speaker 3: what's the most surprising thing you learned about the used
• 24:15
  Speaker 3: car business?
• 24:18
  Speaker 4: How predictable it is? Huh, how predictable car pricing at
• 24:22
  Speaker 4: a certain period. But I was lucky to work there
• 24:25
  Speaker 4: pre COVID, where you know, the supply and demand and
• 24:27
  Speaker 4: completely changed. That was a big surprise to me. It's
• 24:31
  Speaker 4: much more data driven and predictive than people realized.
• 24:35
  Speaker 3: Okay, what's one tip for somebody buying or selling a
• 24:40
  Speaker 3: used car?
• 24:41
  Speaker 4: So most people are buying a car also trading in.
• 24:43
  Speaker 4: They're selling a car, okay, And about eighty percent of
• 24:46
  Speaker 4: the time people are taking a loan okay. And most
• 24:49
  Speaker 4: people come in and they do reams of research and
• 24:52
  Speaker 4: coming with their armed guard to a car dealer, and
• 24:55
  Speaker 4: they focus just on the buy. In the meantime, they
• 24:57
  Speaker 4: get a shitty price on the trade and they get
• 25:00
  Speaker 4: a and they get a you know, a bad financing
• 25:05
  Speaker 4: deal on their car, and that tax on you know
• 25:08
  Speaker 4: many South of daughters, so the cost, but they got
• 25:10
  Speaker 4: a great deal on the bi in their minds.
• 25:14
  Speaker 3: I read that you bought a house on Ridyon Avenue,
• 25:17
  Speaker 3: which was interesting to me because I used to.
• 25:19
  Speaker 2: Live on Ridian Avenue in South Beach. Uh, And I
• 25:22
  Speaker 2: couldn't think, like, oh, what's that? We both have lived
• 25:25
  Speaker 2: on Bridan Avenue. Question, I don't know. Do you go
• 25:27
  Speaker 2: to that Publis? How's South Beach strading you?
• 25:31
  Speaker 4: I'm here right now. I spend most of the time, okay,
• 25:34
  Speaker 4: most of the time in New York, but I'm here.
• 25:36
  Speaker 4: I'm here right now.
• 25:38
  Speaker 2: It's a very hot time to be on Ridian Avenue.
• 25:41
  Speaker 4: From my experienced, I haven't here this time of the year.
• 25:44
  Speaker 4: I have a pool, and I'm okay, and I'm close
• 25:47
  Speaker 4: to the beach, and I love Miami and I love
• 25:49
  Speaker 4: Miami Beach. So it's a lot of fun. It's never
• 25:52
  Speaker 4: at an moment here.
• 25:53
  Speaker 2: Well, so you grew up in Israel, you lived in
• 25:55
  Speaker 2: New York.
• 25:56
  Speaker 3: Right, New York little known fact is on the beach,
• 25:59
  Speaker 3: and you have a house now in Miami Beach. You
• 26:01
  Speaker 3: live a lot of the time in Miami Beach. Of
• 26:04
  Speaker 3: all those places, where are the beach snacks? The best
• 26:07
  Speaker 3: best beach snacks?
• 26:09
  Speaker 4: Oh, Israel, by far, Israel is much more street food
• 26:13
  Speaker 4: and snacks.
• 26:14
  Speaker 2: Yeah, so, I mean, what is falafela beach snack? I
• 26:16
  Speaker 2: don't think of falafel as a beach snack. What's the
• 26:18
  Speaker 2: beach snack or is it falafel?
• 26:21
  Speaker 4: Yeah, well you can have falafel and almost anywhere in Israel,
• 26:24
  Speaker 4: so you could you can also get it on the beach.
• 26:28
  Speaker 4: But people people in Israel a eating all the time everywhere,
• 26:31
  Speaker 4: so you know, they also eat on the beach, and
• 26:33
  Speaker 4: it's very family oriented and the beaches are crowded and
• 26:36
  Speaker 4: people are happy.
• 26:43
  Speaker 3: Alon Block is the co founder and CEO of k Health.
• 26:48
  Speaker 3: Today's show was produced by Gabriel Hunter Chang and Edith Russlo.
• 26:52
  Speaker 3: It was edited by Sarah Nicks and engineered by Amanda
• 26:55
  Speaker 3: k Wah. You can email us at problem at Pushkin
• 26:59
  Speaker 3: dot fm. You can find me on Twitter at Jacob Goldstein.
• 27:03
  Speaker 3: I'm Jacob Goldstein and we'll be back next week with
• 27:05
  Speaker 3: another episode of What's Your Pop?
• 27:10
  Speaker 2: Don't Myself