WEBVTT 00:00:01.399 --> 00:00:03.940 So I want to introduce you to a very specific 00:00:03.940 --> 00:00:06.700 patient today. OK, let's hear it. Well, his chart 00:00:06.700 --> 00:00:11.019 is, it's complicated, to say the least. He has 00:00:11.019 --> 00:00:13.480 been examined by literally hundreds of thousands 00:00:13.480 --> 00:00:15.660 of physicians over the years. That is a lot of 00:00:15.660 --> 00:00:18.260 copes. Right. And he currently suffers from over 00:00:18.260 --> 00:00:21.420 30 disdained cardiac diseases, often at the exact 00:00:21.420 --> 00:00:24.100 same time. Wait, so like aortic stenosis and 00:00:24.100 --> 00:00:26.899 mitral regurgitation simultaneously? Exactly. 00:00:27.070 --> 00:00:29.769 And despite being a resident at the University 00:00:29.769 --> 00:00:32.429 of Miami since, I think, 1968, he hasn't aged 00:00:32.429 --> 00:00:34.530 a single day. Well, that is a medical history 00:00:34.530 --> 00:00:37.170 that would stump anybody. I assume this patient 00:00:37.170 --> 00:00:39.530 is remarkably compliant, though. He never complains 00:00:39.530 --> 00:00:41.829 about a cold stethoscope. The ideal patient. 00:00:42.090 --> 00:00:44.149 Because obviously, we aren't talking about a 00:00:44.149 --> 00:00:45.829 human being here. No, we're talking about Harvey. 00:00:45.890 --> 00:00:48.409 Harvey, a cardiopulmonary patient simulator. 00:00:48.609 --> 00:00:51.770 An absolute legend in medical education. Exactly. 00:00:52.759 --> 00:00:55.240 you know, calling Harvey a simulator almost feels 00:00:55.240 --> 00:00:57.859 like we're underselling the engineering. Because 00:00:57.859 --> 00:00:59.740 when I dug into the background on this for our 00:00:59.740 --> 00:01:02.259 deep dive today, I was expecting, you know, a 00:01:02.259 --> 00:01:04.120 standard mannequin with a speaker in its chest. 00:01:04.900 --> 00:01:06.620 Just plain sounds. Right, a doll, basically. 00:01:06.859 --> 00:01:09.620 But we are talking about a machine that is much 00:01:09.620 --> 00:01:14.920 closer to a 1960s sci -fi robot. It's this incredible 00:01:14.920 --> 00:01:17.780 piece of hardware designed to mimic the human 00:01:17.780 --> 00:01:20.680 cardiovascular system with a level of accuracy 00:01:20.680 --> 00:01:24.540 that is honestly a little unsettling. It is uncanny, 00:01:24.599 --> 00:01:26.719 and you're completely right to focus on the engineering 00:01:26.719 --> 00:01:29.200 aspect of it, but I think the context is just 00:01:29.200 --> 00:01:31.379 as important here. Because Harvey represents 00:01:31.379 --> 00:01:33.799 this massive pivot in the history of medical 00:01:33.799 --> 00:01:36.420 ethics. Also. Well, it's the moment we collectively 00:01:36.420 --> 00:01:38.900 decided that practicing on sick, vulnerable people 00:01:38.900 --> 00:01:40.840 wasn't the only way to learn how to be a doctor. 00:01:41.019 --> 00:01:42.680 Right, the whole transition away from the old 00:01:42.680 --> 00:01:44.980 see one, do one, teach one model. Exactly. Moving 00:01:44.980 --> 00:01:47.239 toward practice on the robot until you stop killing 00:01:47.239 --> 00:01:49.640 it. So here's our mission for this deep dive. 00:01:50.060 --> 00:01:53.359 We are going to unpack how a plastic mannequin 00:01:53.359 --> 00:01:55.900 essentially revolutionized cardiology training. 00:01:56.379 --> 00:01:58.040 It really did. We're going to get into the guess 00:01:58.040 --> 00:02:01.620 of this thing. The telephone relays the cams, 00:02:02.040 --> 00:02:04.459 the servo motors. And we really need to figure 00:02:04.459 --> 00:02:08.539 out why, in an era where we have AI and VR and 00:02:08.539 --> 00:02:11.680 all this digital tech medical students are still 00:02:11.680 --> 00:02:15.319 required to put a physical stethoscope on a physical 00:02:15.319 --> 00:02:17.860 plastic chest. It really highlights that tension 00:02:17.860 --> 00:02:20.379 between the digital world and the tactile world. 00:02:20.979 --> 00:02:23.120 Medicine is a contact sport after all. I like 00:02:23.120 --> 00:02:26.479 that. So let's rewind a bit. Back to 1968. Miami. 00:02:27.159 --> 00:02:29.180 There's this guy, Dr. Michael S. Gordon. He's 00:02:29.180 --> 00:02:31.439 the architect behind all of this. He is. What 00:02:31.439 --> 00:02:34.199 drives a cardiologist to suddenly become a roboticist? 00:02:34.439 --> 00:02:37.219 Pure necessity, honestly. Dr. Gordon looked around 00:02:37.219 --> 00:02:39.039 at the state of medical training at the time 00:02:39.039 --> 00:02:41.580 and just saw this glaring inefficiency. What 00:02:41.580 --> 00:02:43.830 was the main issue? Well, back then, if you wanted 00:02:43.830 --> 00:02:46.710 to teach a medical student what a very specific 00:02:46.710 --> 00:02:49.229 rare heart murmur sounded like, you literally 00:02:49.229 --> 00:02:51.370 just had to wait for a patient with that exact 00:02:51.370 --> 00:02:53.110 condition to walk through the hospital door. 00:02:53.169 --> 00:02:55.710 Yeah, so it was just the luck of the draw. Precisely. 00:02:56.009 --> 00:02:59.310 The luck of the draw curriculum. If you're a 00:02:59.310 --> 00:03:01.729 third -year med student and you desperately need 00:03:01.729 --> 00:03:04.629 to learn about angina pectoris, But nobody in 00:03:04.629 --> 00:03:06.729 the hospital happens to be having that specific 00:03:06.729 --> 00:03:08.469 chest pain that week. You just miss out on learning 00:03:08.469 --> 00:03:11.090 it. You just don't learn it. Or you read about 00:03:11.090 --> 00:03:13.789 it in a book, which is useless for a sound. That 00:03:13.789 --> 00:03:16.949 was the reality. Dr. Gordon wanted to fix that 00:03:16.949 --> 00:03:20.110 by creating what we now call a standardized patient. 00:03:20.590 --> 00:03:23.610 So before we get to how he actually built it, 00:03:23.689 --> 00:03:26.360 let's clear up the name. Harvey. I mean, it sounds 00:03:26.360 --> 00:03:28.280 like a polite accountant. It doesn't sound like 00:03:28.280 --> 00:03:31.099 a cutting edge medical robot. Why name it Harvey? 00:03:31.360 --> 00:03:33.699 It's actually a tribute. Dr. Gordon named the 00:03:33.699 --> 00:03:36.360 simulator after his own mentor, W. Proctor Harvey. 00:03:36.479 --> 00:03:38.919 Oh, OK. Yeah, he was a luminary physician at 00:03:38.919 --> 00:03:42.080 Georgetown University. It was a nod to the man 00:03:42.080 --> 00:03:45.000 who essentially taught Gordon the art of auscultation, 00:03:45.360 --> 00:03:47.939 the art of listening to the body. That is a high 00:03:47.939 --> 00:03:49.919 pressure tribute, though. Like, hey, I built 00:03:49.919 --> 00:03:53.120 a robot in your image. I really hope it doesn't 00:03:53.120 --> 00:03:54.919 break. Right. It's quite the compliment. But 00:03:54.919 --> 00:03:56.840 surely there were other simulators back then. 00:03:57.360 --> 00:03:59.360 Because, I mean, the sources mention Rosuchian. 00:03:59.840 --> 00:04:02.539 And I feel like everyone knows that face. Oh, 00:04:02.580 --> 00:04:05.759 yeah. The classic CPR face. But think about what 00:04:05.759 --> 00:04:08.800 Rosuchian actually is structurally. Just a hollow 00:04:08.800 --> 00:04:11.539 shell, right? Basically, she is a plumbing simulator. 00:04:11.860 --> 00:04:13.879 You blow air into the mouth, the chest rises. 00:04:14.280 --> 00:04:16.300 You push down on the chest, the air goes out. 00:04:16.360 --> 00:04:19.360 Right. So it's purely mechanics. Not diagnostics. 00:04:19.779 --> 00:04:21.899 Exactly. Rosacean doesn't have a disease. She 00:04:21.899 --> 00:04:23.980 doesn't have an internal pathology that you need 00:04:23.980 --> 00:04:26.000 to figure out. She's a balloon, essentially. 00:04:27.079 --> 00:04:30.519 And Dr. Gordon wasn't trying to teach basic CPR. 00:04:30.639 --> 00:04:32.980 He was trying to teach advanced internal medicine. 00:04:33.060 --> 00:04:36.120 Which requires way more nuance. Way more. He 00:04:36.120 --> 00:04:38.600 needed a machine that could mimic the incredibly 00:04:38.600 --> 00:04:41.980 subtle, terrifyingly quiet swish of a failing 00:04:41.980 --> 00:04:44.259 heart valve. There was also another one mentioned 00:04:44.259 --> 00:04:47.430 in the research, SIM -1. which was computer -driven. 00:04:47.569 --> 00:04:50.310 Yes, SIM -1 was way ahead of its time. But it 00:04:50.310 --> 00:04:53.329 was incredibly complex and extremely rare. Dr. 00:04:53.470 --> 00:04:55.129 Gordon needed something that could be replicated 00:04:55.129 --> 00:04:57.370 and mass -produced. He needed a practical teaching 00:04:57.370 --> 00:05:00.269 tool, not just an expensive proof of concept. 00:05:00.410 --> 00:05:02.509 And that leads us right to the engineering, which 00:05:02.509 --> 00:05:05.250 I know is the wildest part of this story. Because 00:05:05.250 --> 00:05:08.509 the specs on that 1968 prototype... Oh, they're 00:05:08.509 --> 00:05:11.410 fascinating. ...wild is an understatement. When 00:05:11.410 --> 00:05:14.129 I read telephone relays in the source material... 00:05:14.189 --> 00:05:16.430 I actually had to double check my notes. I thought 00:05:16.430 --> 00:05:18.310 it was reading a history of the telegraph by 00:05:18.310 --> 00:05:21.629 mistake. It sounds so primitive now. We are talking 00:05:21.629 --> 00:05:24.470 about the literal physical components used to 00:05:24.470 --> 00:05:27.350 wrote phone calls in the 1950s. Yes, but that's 00:05:27.350 --> 00:05:29.250 what was available to them. You really have to 00:05:29.250 --> 00:05:32.209 put yourself in the mindset of a 1968 engineer. 00:05:32.769 --> 00:05:35.370 You don't have microprocessors. You don't have 00:05:35.370 --> 00:05:38.329 cheap stepper motors or Arduinos. Right, you 00:05:38.329 --> 00:05:40.509 just have raw electricity and mechanical switches. 00:05:40.769 --> 00:05:43.529 Exactly. So how does a telephone relay create 00:05:43.529 --> 00:05:46.290 a heartbeat? because a relay is binary right. 00:05:46.449 --> 00:05:48.870 It's on or off. Click clack. But a heartbeat 00:05:48.870 --> 00:05:51.529 is fluid, it's a wave. That is the genius of 00:05:51.529 --> 00:05:54.589 this whole MacGyver phase of the prototype. Dr. 00:05:54.769 --> 00:05:57.089 Gordon and his engineering team chained all these 00:05:57.089 --> 00:05:59.310 relays together. Like in a sequence? Yes, imagine 00:05:59.310 --> 00:06:01.750 a series of dominoes falling. By firing these 00:06:01.750 --> 00:06:04.569 binary relays in a very rapid, highly specific 00:06:04.569 --> 00:06:07.389 sequence, they could actually create a mechanical 00:06:07.389 --> 00:06:10.050 wave. So one relay pushes a small lever, then 00:06:10.050 --> 00:06:12.170 a fraction of a second later the next one fires, 00:06:12.209 --> 00:06:14.550 then the next, and it creates this rising and 00:06:14.550 --> 00:06:17.149 falling physical motion. Which mimics the actual 00:06:17.149 --> 00:06:19.839 pulse. Right. It mimics the precordial impulse, 00:06:20.100 --> 00:06:22.600 the literal physical kick of the heart muscle 00:06:22.600 --> 00:06:25.300 against the inside of the chest wall. That is 00:06:25.300 --> 00:06:27.540 just brute force engineering. It's almost like 00:06:27.540 --> 00:06:29.860 stop motion animation, but built with solenoids. 00:06:30.060 --> 00:06:32.660 It really is. And that's just the movement. For 00:06:32.660 --> 00:06:35.639 the audio, they used a four -track tape recorder. 00:06:35.860 --> 00:06:39.540 A four -track? Yeah. So inside this highly sophisticated, 00:06:39.800 --> 00:06:42.959 groundbreaking medical robot, there is essentially 00:06:42.959 --> 00:06:46.620 a garage band. demo tape playing Mitchell Stenosis 00:06:46.620 --> 00:06:49.040 in D minor. Hey, it was the high fidelity audio 00:06:49.040 --> 00:06:51.519 of the day. That's true. But think about the 00:06:51.519 --> 00:06:53.240 synchronization challenge there. Oh, it was a 00:06:53.240 --> 00:06:55.720 nightmare. Because you have a magnetic tape playing 00:06:55.720 --> 00:06:58.480 an audio loop of a sound, and then you have a 00:06:58.480 --> 00:07:00.920 massive bank of telephone relays clacking away 00:07:00.920 --> 00:07:03.220 to create the physical movement. Right. Getting 00:07:03.220 --> 00:07:05.420 those two entirely different systems to line 00:07:05.420 --> 00:07:08.319 up so the physical pulse matches the auditory 00:07:08.319 --> 00:07:11.160 thump exactly. If the tape drags even a little 00:07:11.160 --> 00:07:13.860 bit, your plastic patient suddenly has a severe 00:07:13.860 --> 00:07:16.279 arrhythmia. Exactly, and the students get completely 00:07:16.279 --> 00:07:18.620 confused. Which is exactly why that complexity 00:07:18.620 --> 00:07:21.199 led them to phase two. They moved away from those 00:07:21.199 --> 00:07:23.000 telephone relays pretty quickly when they started 00:07:23.000 --> 00:07:24.759 building the commercial models. What did they 00:07:24.759 --> 00:07:27.759 switch to? They adopted a system of cams and 00:07:27.759 --> 00:07:30.660 levers. Mechanical engineering, basically. Cams 00:07:30.660 --> 00:07:33.319 feel a bit more industrial, like something you'd 00:07:33.319 --> 00:07:35.750 find in a car engine. Because it's the exact 00:07:35.750 --> 00:07:39.029 same principle as a camshaft in an internal combustion 00:07:39.029 --> 00:07:41.930 engine, you have a central rotating shaft with 00:07:41.930 --> 00:07:44.389 these egg -shaped lobes on it. Those are the 00:07:44.389 --> 00:07:48.410 cams. Okay. And as that shaft spins, those lobes 00:07:48.410 --> 00:07:51.550 push against levers. So the physical shape of 00:07:51.550 --> 00:07:54.069 the cam dictates the movement pattern on the 00:07:54.069 --> 00:07:56.329 chest. So if you want a really sharp, sudden 00:07:56.329 --> 00:07:59.050 pulse, You physically cut the steep cam. Exactly. 00:07:59.490 --> 00:08:01.629 And if you want a slow rolling pulse, you cut 00:08:01.629 --> 00:08:04.509 a smooth, rounded cam. That's incredible. So 00:08:04.509 --> 00:08:07.389 the specific disease was literally hard -coded 00:08:07.389 --> 00:08:09.550 into the metal shape of the component. It was 00:08:09.550 --> 00:08:12.449 a fully mechanical program. No software at all. 00:08:12.610 --> 00:08:15.970 But here is my absolute favorite detail about 00:08:15.970 --> 00:08:18.209 this phase two version of Harvey. What's that? 00:08:18.329 --> 00:08:20.490 They completely overhauled the mechanics, but 00:08:20.490 --> 00:08:22.350 they kept the four -track tape for the sound. 00:08:23.069 --> 00:08:25.750 The audio source remained magnetic tape for years. 00:08:25.930 --> 00:08:27.750 And until they literally couldn't use it anymore, 00:08:27.970 --> 00:08:30.410 right? Right. The Ampro Corporation story from 00:08:30.410 --> 00:08:34.139 the sources. This detail killed me. It is the 00:08:34.139 --> 00:08:36.559 perfect example of how fragile innovation can 00:08:36.559 --> 00:08:38.759 be sometimes. So what happened was the Ampro 00:08:38.759 --> 00:08:41.860 Corporation, which manufactured the very specific 00:08:41.860 --> 00:08:44.559 four track tape decks that Harvey relied on, 00:08:44.679 --> 00:08:46.960 just went out of business. Right. The whole supply 00:08:46.960 --> 00:08:49.159 chain just broke overnight. They couldn't source 00:08:49.159 --> 00:08:52.200 the one vital audio component anymore. Completely 00:08:52.200 --> 00:08:56.000 forced Dr. Gordon's team to innovate again. They 00:08:56.000 --> 00:08:57.960 didn't switch to digital sound because they wanted 00:08:57.960 --> 00:08:59.860 to be cutting edge. They switched because they 00:08:59.860 --> 00:09:01.700 literally couldn't buy the tape. decks anymore. 00:09:02.059 --> 00:09:04.559 Necessity is the mother of invention, but in 00:09:04.559 --> 00:09:06.899 this case, bankruptcy is the mother of upgrades. 00:09:07.340 --> 00:09:09.299 That's a great way to put it. And that forced 00:09:09.299 --> 00:09:11.500 upgrade pushed them right into the modern era, 00:09:11.700 --> 00:09:14.440 phase three, which is the Harvey that most medical 00:09:14.440 --> 00:09:16.600 students actually encounter today. So we move 00:09:16.600 --> 00:09:19.940 from metal cams and magnetic tape to servo motors 00:09:19.940 --> 00:09:23.840 and digital sound cards. Yes. And servo motors 00:09:23.840 --> 00:09:27.419 represent a huge leap. Now we are talking actual 00:09:27.419 --> 00:09:30.620 robotics, precision control. Because you get 00:09:30.759 --> 00:09:33.259 infinite adjustability, right? Exactly. With 00:09:33.259 --> 00:09:36.000 a metal cam, you are strictly limited to the 00:09:36.000 --> 00:09:38.840 shape of that specific piece of metal. But with 00:09:38.840 --> 00:09:41.659 a servo motor, it's just software code. You just 00:09:41.659 --> 00:09:44.320 tell the motor to move to position X at speed 00:09:44.320 --> 00:09:47.879 Y. So it allowed them to simulate nuances that 00:09:47.879 --> 00:09:50.399 were just physically impossible with the older 00:09:50.399 --> 00:09:53.639 mechanical systems? Yes. And crucially, it dropped 00:09:53.639 --> 00:09:56.360 the price tag significantly. I saw the numbers 00:09:56.360 --> 00:09:58.750 in the report and they were stark. The original 00:09:58.750 --> 00:10:01.090 Harvey cost around $100 ,000 back in the day. 00:10:01.330 --> 00:10:03.970 Which is a massive capital expenditure for any 00:10:03.970 --> 00:10:06.769 medical school. And then the newer Servo and 00:10:06.769 --> 00:10:09.470 digital version dropped the cost to around $50 00:10:09.470 --> 00:10:12.210 ,000. Right. Which is still the price of a luxury 00:10:12.210 --> 00:10:14.929 sedan, but for a university budget, that's a 00:10:14.929 --> 00:10:16.450 discount that lets you buy two of them instead 00:10:16.450 --> 00:10:19.379 of one. It democratized the technology. way more 00:10:19.379 --> 00:10:21.580 schools could suddenly afford to build a standardized 00:10:21.580 --> 00:10:24.019 patient lab. So let's talk about the actual user 00:10:24.019 --> 00:10:26.440 experience now, because we've established that 00:10:26.440 --> 00:10:29.139 Harvey isn't just a basic doll. Far from it. 00:10:29.379 --> 00:10:32.179 When a med student walks into the lab, and up 00:10:32.179 --> 00:10:35.899 to the current version of Harvey, What are they 00:10:35.899 --> 00:10:37.700 actually interacting with? It's not just putting 00:10:37.700 --> 00:10:40.019 a stethoscope on and listening, right? Not at 00:10:40.019 --> 00:10:41.799 all. And this is really where the simulator tag 00:10:41.799 --> 00:10:44.620 falls short. It is a completely multi -sensory 00:10:44.620 --> 00:10:46.980 experience. Bring it down for me. OK, so you 00:10:46.980 --> 00:10:49.460 have the auditory component, obviously. You are 00:10:49.460 --> 00:10:52.960 listening to six entirely different breath sounds. 00:10:53.620 --> 00:10:57.460 And there are nine distinct cardiac auscultation 00:10:57.460 --> 00:11:00.200 areas on the chest. Nine places just to listen 00:11:00.200 --> 00:11:03.679 to the heart. Right. And he can simulate 30 different 00:11:03.679 --> 00:11:06.320 cardiac diseases. That is a crowded chart. It 00:11:06.320 --> 00:11:09.139 is. But the tactile feedback is really where 00:11:09.139 --> 00:11:12.299 Harvey shines, the touching. Palpation. Palpation, 00:11:12.580 --> 00:11:14.940 exactly. You actually place your bare hand on 00:11:14.940 --> 00:11:18.000 the plastic chest. And you can physically feel 00:11:18.000 --> 00:11:20.259 the apical impulse. You can feel the heave of 00:11:20.259 --> 00:11:22.320 the right ventricle against your palm. That's 00:11:22.320 --> 00:11:23.899 wild. And you don't just look at the chest either. 00:11:23.899 --> 00:11:25.919 You have to look at the neck. The neck is a really 00:11:25.919 --> 00:11:27.700 big deal in this, right? The source has highlighted 00:11:27.700 --> 00:11:31.019 it. It's critical. The jugular venous pulse in 00:11:31.019 --> 00:11:34.759 the neck gives a doctor a direct visual readout 00:11:34.759 --> 00:11:37.879 of the pressure inside the right atrium of the 00:11:37.879 --> 00:11:41.080 heart. Harvey physically simulates that delicate 00:11:41.080 --> 00:11:43.940 pulsation in the synthetic neck material. Which 00:11:43.940 --> 00:11:46.100 brings up the aortic stenosis update that was 00:11:46.100 --> 00:11:49.320 mentioned in the notes, the 1985 patch, effectively. 00:11:50.179 --> 00:11:52.059 I want to drill down on this because it seems 00:11:52.059 --> 00:11:54.320 like a perfect example of a hardware and software 00:11:54.320 --> 00:11:57.029 disconnect. It is the definitive example of that 00:11:57.029 --> 00:11:59.470 problem. So aortic stenosis. The heart valve 00:11:59.470 --> 00:12:02.110 narrows. The blood is forcefully pushed through 00:12:02.110 --> 00:12:04.809 a very tight opening, which creates this purulent, 00:12:05.169 --> 00:12:08.149 loud, rushing murmur. But that sound doesn't 00:12:08.149 --> 00:12:11.370 just stay isolated in the chest cavity. The acoustics 00:12:11.370 --> 00:12:13.309 of the body cause that sound to radiate right 00:12:13.309 --> 00:12:15.730 up the carotid arteries into the neck. So a good 00:12:15.730 --> 00:12:17.710 diagnostic doctor isn't just listening to the 00:12:17.710 --> 00:12:20.309 chest. They're actively listening to the pageant's 00:12:20.309 --> 00:12:22.850 neck. to confirm the diagnosis. Correct. If you 00:12:22.850 --> 00:12:24.850 hear the specific murmur in the chest and you 00:12:24.850 --> 00:12:26.649 also hear it in the neck, it strongly suggests 00:12:26.649 --> 00:12:30.210 aortic stenosis. Let me guess. They didn't transmit 00:12:30.210 --> 00:12:33.009 the sound to the neck. They did not. So you have 00:12:33.009 --> 00:12:35.269 all these medical students doing everything perfectly 00:12:35.269 --> 00:12:38.250 right. They listen to the chest. They hear the 00:12:38.250 --> 00:12:41.029 correct murmur. But then they check the neck 00:12:41.029 --> 00:12:43.529 like they were taught. They hear absolute silence 00:12:43.529 --> 00:12:46.870 and they confidently rule out aortic stenosis. 00:12:47.169 --> 00:12:49.970 Exactly. The robot was actively teaching them 00:12:49.970 --> 00:12:53.029 the wrong diagnostic criteria. Wow. Or rather 00:12:53.029 --> 00:12:56.009 the physical limitation of the hardware was inducing 00:12:56.009 --> 00:12:59.370 a diagnostic error in the student. So the 1985 00:12:59.370 --> 00:13:02.309 update was a major overhaul specifically just 00:13:02.309 --> 00:13:04.529 to address this one flaw. It had to be. They 00:13:04.529 --> 00:13:06.809 had to deeply integrate the software and the 00:13:06.809 --> 00:13:09.110 audio output to ensure that when the instructor 00:13:09.110 --> 00:13:11.690 selected aortic stenosis on the control panel, 00:13:12.090 --> 00:13:14.470 the sound actually physically radiated to the 00:13:14.470 --> 00:13:16.409 next speakers. It's so fascinating that they 00:13:16.409 --> 00:13:18.950 literally had to patch reality. They had to ensure 00:13:18.950 --> 00:13:21.649 the simulation perfectly matched the real human 00:13:21.649 --> 00:13:24.090 physiology because otherwise you were actively 00:13:24.090 --> 00:13:26.710 training a generation of doctors to miss a critical 00:13:26.710 --> 00:13:29.330 diagnosis in real living patients. And now from 00:13:29.330 --> 00:13:31.509 what I understand, the visual component is fully 00:13:31.509 --> 00:13:33.409 integrated too. It's not just a blind physical 00:13:33.409 --> 00:13:35.600 test. anymore. Right, the supporting materials 00:13:35.600 --> 00:13:38.379 evolved right alongside the robot. How so? Well 00:13:38.379 --> 00:13:41.120 in the old days you had a literal slide carousel 00:13:41.120 --> 00:13:43.360 clicking away in the background, click -clack, 00:13:43.700 --> 00:13:46.820 showing a static x -ray or an EKG printout while 00:13:46.820 --> 00:13:49.139 you listen to the robot's chest. Very analog. 00:13:49.539 --> 00:13:53.259 Very. But starting around 1998 they moved to 00:13:53.259 --> 00:13:56.639 software. Now it's a seamless digital experience. 00:13:57.080 --> 00:13:59.320 You have a computer monitor right next to Harvey 00:13:59.320 --> 00:14:02.179 showing the patient's full history, their lab 00:14:02.179 --> 00:14:06.259 results, dynamic echo loops, all perfectly synchronized 00:14:06.259 --> 00:14:08.080 with what you are hearing and feeling on the 00:14:08.080 --> 00:14:09.940 mannequin. It's basically a flight simulator 00:14:09.940 --> 00:14:12.360 for cardiologists. That's the perfect analogy. 00:14:12.500 --> 00:14:15.500 Which brings us to the big why of this whole 00:14:15.500 --> 00:14:18.139 deep dive. We know how it works now. We know 00:14:18.139 --> 00:14:21.320 it mimics these complex diseases. But why is 00:14:21.320 --> 00:14:23.419 this actually better than the old school method? 00:14:23.639 --> 00:14:25.620 Why not just send the students straight to the 00:14:25.620 --> 00:14:27.940 hospital wards to find real sick people? Two 00:14:27.940 --> 00:14:30.220 fundamental reasons. Standardization and safety. 00:14:30.460 --> 00:14:32.820 Okay, let's break those down. We touched on standardization 00:14:32.820 --> 00:14:35.299 earlier. With Harvey, every single student in 00:14:35.299 --> 00:14:37.340 the class hears the exact same murmur. There 00:14:37.340 --> 00:14:40.779 is no ambiguity. Real human bodies vary wildly. 00:14:41.129 --> 00:14:43.389 But Harvey does not. So it makes testing fair. 00:14:43.590 --> 00:14:46.649 Exactly. But the safety aspect, and I mean specifically 00:14:46.649 --> 00:14:50.009 the psychological safety of the student, is the 00:14:50.009 --> 00:14:52.049 real game changer here. The source is called 00:14:52.049 --> 00:14:55.470 this cognitive load theory, right? Yes. Imagine 00:14:55.470 --> 00:14:59.169 you are a 23 -year -old medical student. You 00:14:59.169 --> 00:15:02.629 are standing over a real patient who is in active 00:15:02.629 --> 00:15:05.480 heart failure. High stakes. Very high stakes. 00:15:05.899 --> 00:15:08.379 The patient is scared. Their family is probably 00:15:08.379 --> 00:15:10.620 in the room watching you. You are absolutely 00:15:10.620 --> 00:15:12.659 terrified you're going to hurt them or just look 00:15:12.659 --> 00:15:15.580 totally incompetent. Your brain is flooded with 00:15:15.580 --> 00:15:18.059 cortisol. It's really hard to learn the subtle 00:15:18.059 --> 00:15:21.279 nuances of a split S2 heart sound when you're 00:15:21.279 --> 00:15:23.820 sweating entirely through your scrubs. Exactly. 00:15:24.059 --> 00:15:26.620 You can't focus. The raw cognitive load of the 00:15:26.620 --> 00:15:28.759 situation just completely overpowers the educational 00:15:28.759 --> 00:15:30.899 moment. And Harvey just strips all of that away. 00:15:30.980 --> 00:15:33.299 Completely. The patient is plastic. He doesn't 00:15:33.299 --> 00:15:35.379 care if you take 10 minutes just to find the 00:15:35.379 --> 00:15:38.259 pulse. He doesn't judge you if you put the stethoscope 00:15:38.259 --> 00:15:40.120 in the completely wrong spot the first three 00:15:40.120 --> 00:15:43.360 times. You can fail safely. You can fail repeatedly. 00:15:43.799 --> 00:15:46.399 And that repetition is what builds the physical 00:15:46.399 --> 00:15:48.980 muscle memory and the acoustic memory. So when 00:15:48.980 --> 00:15:51.980 you finally do stand over that real terrified 00:15:51.980 --> 00:15:55.080 patient, you aren't using your brain power to 00:15:55.080 --> 00:15:56.799 think about how to hold your stethoscope. You're 00:15:56.799 --> 00:15:59.179 just listening. I get that. But I do have to 00:15:59.179 --> 00:16:00.980 play devil's advocate here for a second. Go for 00:16:00.980 --> 00:16:03.659 it. The sources cite these studies showing that 00:16:03.659 --> 00:16:06.279 students who trained on Harvey got significantly 00:16:06.279 --> 00:16:08.840 higher scores on their cardiology electives. 00:16:09.240 --> 00:16:12.000 Yes, the data is quite robust on that point. 00:16:12.240 --> 00:16:14.279 But are they actually better doctors or are they 00:16:14.279 --> 00:16:16.500 just better at taking the test? I mean, if the 00:16:16.500 --> 00:16:18.779 test is identify the specific sound and they 00:16:18.779 --> 00:16:20.639 spent 10 hours listening to that exact sound 00:16:20.639 --> 00:16:22.980 on a robot, of course they're going to ace the 00:16:22.980 --> 00:16:26.440 exam. Does that clinical scale actually translate 00:16:26.440 --> 00:16:29.710 to the messiness of a real human body? That is 00:16:29.710 --> 00:16:31.950 the eternal question of all simulation training. 00:16:32.049 --> 00:16:34.129 And you are totally right to be skeptical, because 00:16:34.129 --> 00:16:36.970 a robot usually presents the textbook case. Real 00:16:36.970 --> 00:16:39.490 bodies are noisy. They have thick chest walls. 00:16:39.590 --> 00:16:41.309 They might have severe lung disease. They mask 00:16:41.309 --> 00:16:43.470 the heart sounds. They shift around. They breathe 00:16:43.470 --> 00:16:46.049 heavily. Real patients don't have a convenient 00:16:46.049 --> 00:16:48.289 volume knob. They absolutely do not. However, 00:16:48.330 --> 00:16:50.409 the counter argument from educators is that you 00:16:50.409 --> 00:16:53.750 cannot possibly recognize the messy complex version 00:16:53.750 --> 00:16:56.350 if you don't have a rock solid understanding 00:16:56.350 --> 00:16:59.500 of the perfect baseline version first. Harvey 00:16:59.500 --> 00:17:02.299 builds that initial mental model. Once you know 00:17:02.299 --> 00:17:04.900 exactly what aortic regurgitation should sound 00:17:04.900 --> 00:17:07.200 like in a vacuum, you actually have a fighting 00:17:07.200 --> 00:17:09.259 chance of picking it out of the noise in a real 00:17:09.259 --> 00:17:12.200 sick patient. So Harvey is the baseline. He is 00:17:12.200 --> 00:17:14.519 the reference standard. And you have to consider 00:17:14.519 --> 00:17:17.740 who uses this. It's not just first year med students. 00:17:17.759 --> 00:17:20.240 It's used by senior residents, anesthesiologists, 00:17:20.440 --> 00:17:23.420 emergency physicians, and even military combat 00:17:23.420 --> 00:17:26.700 medics. Combat medics. Yes. And combat medics 00:17:26.700 --> 00:17:29.009 aren't training pass a multiple -choice test. 00:17:29.089 --> 00:17:31.369 They are training to keep human beings alive 00:17:31.369 --> 00:17:34.710 in total chaos if they value the simulator that 00:17:34.710 --> 00:17:37.910 really speaks volumes about its real -world translation. 00:17:38.210 --> 00:17:40.250 It's so interesting to look at the massive scope 00:17:40.250 --> 00:17:43.390 here. We started with Dr. Gordon in Miami in 00:17:43.390 --> 00:17:45.890 the 60s, probably looking at a pile of old telephone 00:17:45.890 --> 00:17:48.089 switches and wondering if he was completely crazy. 00:17:48.430 --> 00:17:51.049 And now this is the global standard. It is. In 00:17:51.049 --> 00:17:53.670 fact, Laredal, the exact same company that makes 00:17:53.670 --> 00:17:56.329 Recessien, they now distribute Harvey globally. 00:17:56.970 --> 00:17:59.650 massive integrated ecosystem. Which is incredible, 00:18:00.069 --> 00:18:01.950 but I won't end on a thought that's just been 00:18:01.950 --> 00:18:04.009 nagging me while reading all the source material 00:18:04.009 --> 00:18:06.670 about this. What's that? We are rapidly moving 00:18:06.670 --> 00:18:09.349 toward a world where the simulation is almost 00:18:09.349 --> 00:18:11.829 better than the reality. We have these high -fidelity 00:18:11.829 --> 00:18:15.789 mannequins, we have VR surgery, we have AI diagnostic 00:18:15.789 --> 00:18:18.930 tools now. We are getting very, very good at 00:18:18.930 --> 00:18:21.210 mimicking biology. But there's this uncanny valley 00:18:21.210 --> 00:18:23.740 here. And I don't just mean in how the robot 00:18:23.740 --> 00:18:26.599 physically looks, but in the training methodology 00:18:26.599 --> 00:18:28.960 itself. How do you mean? We are approaching a 00:18:28.960 --> 00:18:30.779 point where a medical student might go through 00:18:30.779 --> 00:18:33.660 their entire preclinical training, years and 00:18:33.660 --> 00:18:36.559 years of intense education, without ever touching 00:18:36.559 --> 00:18:39.279 a living, breathing human being who actually 00:18:39.279 --> 00:18:42.720 has a pathology. That is a very legitimate concern. 00:18:42.779 --> 00:18:45.799 We are, in a way, sanitizing the education process. 00:18:46.160 --> 00:18:49.019 Right. Because Harvey is clean. Harvey is polite. 00:18:49.259 --> 00:18:51.980 Harvey doesn't smell. Harvey doesn't cry when 00:18:51.980 --> 00:18:55.480 you give him a bad prognosis. Are we unintentionally 00:18:55.480 --> 00:18:58.460 training a generation of doctors who are technically 00:18:58.460 --> 00:19:01.980 flawless at diagnosing a machine, but emotionally 00:19:01.980 --> 00:19:05.240 completely unprepared for the human element of 00:19:05.240 --> 00:19:07.619 medicine? It's certainly possible. It's a risk. 00:19:07.680 --> 00:19:10.519 But I would argue the flip side. Which is? Because 00:19:10.519 --> 00:19:12.900 Harvey handles the raw technical training so 00:19:12.900 --> 00:19:16.240 efficiently and removes that initial panic, it 00:19:16.299 --> 00:19:18.920 potentially frees up the student's mental bandwidth 00:19:18.920 --> 00:19:21.519 to actually focus on the human connection when 00:19:21.519 --> 00:19:23.819 they finally do meet a real patient. Oh, I see. 00:19:23.980 --> 00:19:26.299 If you aren't internally panicking about just 00:19:26.299 --> 00:19:28.759 finding the murmur, maybe you finally have the 00:19:28.759 --> 00:19:30.440 emotional capacity to actually look the patient 00:19:30.440 --> 00:19:32.619 in the eye and hold their hand. That is definitely 00:19:32.619 --> 00:19:35.299 the optimistic view. The robot handles the raw 00:19:35.299 --> 00:19:37.640 data so the human can handle the actual care. 00:19:37.799 --> 00:19:39.619 That's the hope, anyway. Well, I know the next 00:19:39.619 --> 00:19:41.660 time I'm sitting in the doctor's office. I'm 00:19:41.660 --> 00:19:43.539 definitely going to wonder if they practiced 00:19:43.539 --> 00:19:46.160 on a servo -driven plastic man named after a 00:19:46.160 --> 00:19:48.460 mentor at Georgetown. Chances are very high that 00:19:48.460 --> 00:19:51.059 they did. Thanks for unpacking the mechanical 00:19:51.059 --> 00:19:53.339 heart of medical training with us today. Always 00:19:53.339 --> 00:19:56.180 a pleasure. We'll see you on the next Deep Dive. 00:19:56.559 --> 00:19:57.259 Stay curious.