Announcer:
0:00
Welcome to MedEvidence!, where we help you navigate the ruth behind medical research with unbiased, evidence-proven. facts Hosted by cardiologist and top medical researcher, Dr. Michael Koren.
Dr. Michael Koren:
0:11
Hello, I'm Dr Michael Koren, the executive editor of MedEvidence!, and I'm really excited about our current podcast because I meet a lot of kindred spirits, but I'm going to introduce you to Mike Hafran, who's an MD and an MBA and runs a clinical research site in South Florida, and he's had a diverse of experiences in our field far more than me, quite frankly, as mostly I've been a clinical cardiologist but Mike has done it all. So Mike's going to talk to us a little bit about his experiences and, I think, most importantly, how we're thinking about involving unusual and non-traditional pathways into medicine into the clinical trial world. So, Mike, welcome to MedEvidence! And why don't you tell us a little bit about yourself for the audience
Dr. Mike Hafran:
1:01
Absolutely and thank you for such a generous introduction,
Dr. Michael Koren:
1:07
Well-deserved
Dr. Mike Hafran:
1:08
I want to say that it's similar to you. I am a physician and perhaps my journey has been slightly different from more of that traditional path, although it definitely started out in a very traditional manner. I grew up in the Midwest and did my schooling at University of Wisconsin-Madison and then from there I went on to the East Coast to do residency and other work and training at Columbia Presbyterian At that time. You know it's kind of funny how life can throw challenges at us and one decides like, how am I going to face this challenge? My first challenge was, after residency and joining, working with Columbia Presbyterian,
Dr. Mike Hafran:
2:02
9-11 occurred. I ended up diverting my career path a little bit and I joined the US Air Force and worked as a physician and a physician liaison for four years. And that was a very interesting experience because very early in my career path I was introduced to A: learning how to wear multiple hats and some of them were forced upon me but I gladly accepted them. But also B: being exposed to so many different perspectives, and that's not only within the US Air Force but and all of the different people there, but also, since I was stationed over in Europe, being exposed to the way that Italians approach medicine the way some of our allies, such as the Germans, also approach medicine.
Dr. Michael Koren:
3:02
Yeah, that's so neat. That's so neat, that's so neat. We could spend hours and hours talking about that alone, but we'll talk about this concept of non-traditional clinical trials and how you take something like artificial intelligence and study whether or not it works for healthcare. And do we need to do that? Do we need to apply the same methods that we use to develop new drugs to things like artificial intelligence? So I know that you've written in this area and you've thought about it, so why don't you share your feelings a little bit with the audience on how you think we should move forward with these new concepts?
Dr. Mike Hafran:
3:40
Sure, and I want to just this is kind of like a little bit of ending the last question but also moving into the second question that you just brought up about AI. Subsequently, after leaving the Air Force, I did actually I pursued my MBA at NYU, and there were two things that motivated me.
Dr. Michael Koren:
4:07
It's New York University. For people that don't know the jargon.
Dr. Mike Hafran:
4:11
Sorry about that you know lots of acronyms in medicine.
Dr. Mike Hafran:
4:16
But I was motivated by two things. Because now I had been working kind of more on the hospital side for a couple of years thereafter. My first motivation was I wanted to understand kind of the financial language and definitions and jargon, because so much of medicine is fortunately or unfortunately, you know, driven by money and financial outcomes as well as clinical outcomes. I think they really worked hand in glove.
Dr. Michael Koren:
4:50
Oh, no doubt.
Dr. Mike Hafran:
4:50
But the second thing that really motivated me to pursue my MBA was to get a better understanding of technologies. The technology not only in you know, like hey, how is a ceramic or a titanium artificial joint better than you know a product? Why? Not only from a mechanical but also from a financial side, but also artificial intelligence.
Dr. Mike Hafran:
5:25
But even back then that was just starting, it was more. I truly I was just interested in how do we utilize and bend technology in a manner that can really inform better clinical decisions? Um, and then you know back then that when I was studying that was the realm of more of the smart watches were really making an impact, right, um. So that's kind of those were my motivations to get my MBA and to hopefully leverage and utilize that information to help make medicine, better, help make people healthier.
Dr. Mike Hafran:
6:10
So sorry, very long answer to get to that second question, but right now and you just brought this up there are a lot of newer aspects of medicine that, wow, I never even really considered. AI in and of itself is really changing the game or disrupting the game. But even just the technologies we had you know, fitbit, the Apple Watch now the Apple Watch can take a reasonable ECG which is absolutely amazing or monitor one's rhythm. So how do we, as clinicians, A: just integrate all of this information and then, B: how do we make it meaningful and useful and how do we continue to push the frontier so that there are real results, and those are super important questions.
Dr. Michael Koren:
7:04
And what happens a lot is people talk about the promise of all the technologies, but they never talk about the downside or the potential downside. And what we do in clinical trials is we're constantly weighing the risks versus the benefits and we look to see the benefits, of course, but we have to really look most importantly at the net benefit, which is the benefit minus side effects and other things that may not be so great, and I'm not sure we do that for technology yet, and we probably should so. For every accomplishment of artificial intelligence, I think there has been a side effect of artificial stupidity. So I don't know if you've had that, but when you call a service provider that's using an AI algorithm, how many times have you wasted 10, 15 minutes and not gotten to the answer that you're looking for? Because you have to go through this AI algorithm and, at the end of the day, that algorithm doesn't know what you're trying to get at, or we were talking about this this morning at our meeting. There's all this excitement around autonomous driving and robo-taxis. Well, if somebody has a heart attack in a robo-taxi, what happens? How are we going to deal with that? So these are things that are super interesting to discuss, and that's the reason I think we need to use the sensibilities that we developed in developing new drugs and applying those ideas to artificial intelligence and to other non-traditional areas.
Dr. Michael Koren:
8:36
Another thing that comes up is people are injecting biological agents into people. For example, prp, platelet-rich plasma would be an example of that, and there's relatively little hardcore randomized clinical research on that. But because it's outside of the insurance realm, it's still being promoted fairly aggressively. You can make similar arguments about medical marijuana. It seems to get a pass in terms of being able to show it's effective. I think it probably is. There have been pharmaceuticals that have looked to use cannabinoids for medical reasons. But if you're smoking marijuana, shouldn't there be some clinical trials to see how it works rather than just sort of the sense oh well, it worked for the last three people, it's going to work for you
Dr. Mike Hafran:
9:25
Absolutely, absolutely and actually you're bringing up a really good point, and a key point is what is the driver behind the I'll just say the medicine or the medical product?
Dr. Mike Hafran:
9:40
Is it, for instance, technology that's pushing something forward, or is it, as you kind of mentioned with the marijuana? Is it perhaps more like special interest groups? Or maybe there's also just a pure cultural or social standard or norm that's been accepted but not thoroughly vetted? And the reason why I think something like this is important and I hope this feeds into and pushes our conversation forward I ended up becoming a Chief Medical Officer of a med-tech firm firm and our whole, and this was in the remote patient monitoring and remote therapeutic monitoring space. So those are devices that individuals, patients, have that may monitor some sort of objective physiologic data point, such as blood pressure, heart rate, ecg.
Dr. Mike Hafran:
10:46
The challenge was and this was kind of like the perhaps the secret sauce behind our company and the success we experienced. The technology is out there, people are utilizing it, a lot of data is flowing into physician offices, but is it valid data and what do I do with all this data? So our company helped manage A: vet the information coming in, because a machine is just a machine. And I'll tell you one thing that is really interesting was that the machinery could not tell if the patient was actually taking their blood pressure or if their niece or nephew was actually taking their own blood pressure Interesting. So docs would get highly volatile information. They're like, oh my goodness, we got to like call 911 on this patient, but our system worked to intercede and really it threw an individual, a licensed nurse, practitioner or a real RN, to call in and intercede and just confirm.
Dr. Mike Hafran:
11:59
So perhaps technology is great, especially in medicine, but I think, to your point, we'll never exclude the human factor
Dr. Michael Koren:
12:09
Yeah, well just a few examples that actually that I have experienced, literally, in recent months is an AI trial that we did here in Jacksonville that actually looked at a technology to estimate blood pressure based on your facial changes during the course of a period of time here in the office, and this technology is correlating the blood pressure obtained through a cuff or through some sort of monitoring device with what your face is showing us, and the concept being that, if we ultimately do more telemedicine, these algorithms can actually measure people's blood pressure based on what their face looks like.
Dr. Michael Koren:
12:49
So that's fascinating. We completed a trial here. I haven't seen the results yet, but the vendor involved is already discussing doing further trials, actually both here and perhaps in New York at one of the institutions in New York. So we're involved in that as we speak. So that's something that's kind of neat. I know that there are companies that are looking at psychedelics and trying to understand if psilocybin mushrooms, for example, can be used for depression or for schizophrenia and other things, and that's certainly interesting. It goes back to the old days of Timothy Leary on the LSD experiments from the CIA, but they weren't randomized clinical trials, they were observational studies and I think we would learn much more if we did a randomized clinical trial.
Dr. Michael Koren:
13:32
Or we're talking about cannabis, and we all know from our teenage years that smoking marijuana and eating pot brownies will give you a different feeling. And the question is for medicine, is one better than the other? So we can easily do a double-blind trial that looks at inhaled cannabinoids versus ingested cannabinoids and see if one is better for anxiety or for sleep, whatever the case may be. So there's a lot of really important medical questions that can only be solved in a randomized clinical trial and I'm personally excited to be part of that exploration.
Dr. Mike Hafran:
14:08
Yeah, absolutely.
Dr. Mike Hafran:
14:09
And what you brought up, randomized clinical.
Dr. Mike Hafran:
14:15
That to me is so key because, ultimately, if I look at things really, you know what, honestly, I consider that there are 7 billion clinical trials going on right now.
Dr. Mike Hafran:
14:29
We are every human being is engaging in all sorts of micro and macro decisions that is impacting their health. Our key as clinical research physicians is to take decisions, individuals that share certain sets of parameters, characteristics or attributes, and kind of collate them together so that the investigational product, the intervention, whatever it is, whether it is a medicine that's inhaled or injected, or even the utilization of a particular device, you know, such as a vape device, or even utilizing a particular technology. By minimizing or controlling a lot of the volatility of variables, we can, as a group, reach an estimated or a best guess scenario of like how is this one intervention affecting a group of people with set attributes, as set as can be? You know, so I and that is where the impact to me really comes out that you know, wow, it's like we're involved in this. We can set some parameters and we can yield, we can come out with a general conclusion that seems to be valid and shows that there is a difference or there is not a difference, and who is this going to now help?
Dr. Michael Koren:
16:15
Right, right, yeah, great points. Just to give people sort of very specific examples of the type of things that we can test, and I believe that you can test all these things at your location in South Florida. Correct me if I'm wrong, but, for example, people have been giving different infusions. You see, for example, on the retail market, vitamin B12 infusions or hyperoxygenation chambers or a lot of other things that are sold for cash and may or may not work. Maybe they work, maybe they don't, and we would challenge sponsors of these things to hire a site like your site in South Florida and show that it either works, it doesn't work, this is how well it works, or these are what the side effects are, and I think, ultimately, that's how we move medicine forward.
Dr. Michael Koren:
17:12
Otherwise, it's just a fad that lasts for a period of time and then maybe goes away. But there are many examples of that with infusions of different substances that are people talking about. I've heard things about taking embryonic stem cells for doing different things. You can go to places, for example, in Panama, and get a doctor to inject embryonic stem cells in any part of your body that you want for a fee, but does it work? We don't know.
Dr. Mike Hafran:
17:38
And Does it work?
Dr. Mike Hafran:
17:40
And even to a greater concern is like is it even, is it safe? Yeah, absolutely, you know, are you doing something that is putting maybe not yourself now, but in the near future or further out in the future at significant risk?
Dr. Michael Koren:
17:57
Right, and so what I would say is sites like your site in South Florida or our sites here in North Florida are poised to be able to run these studies, and I believe that sponsors who truly care about the public health impact of these interventions would actually invest in the research to see exactly how well things work.
Dr. Mike Hafran:
18:18
Absolutely, absolutely. And you know, not only should it be and I think this is such a great conversation, and I'm really grateful for the time to spend with you because you're bringing up very interesting and uh forward-looking uh areas where we should continue as as a society, as the United States, invest in researching. Yes, you know, right now, cardiometabolic, um yeah, development of drugs and other things is very important because of our very high rates of um cardiac death, um heart attacks and other strokes. That definitely needs to be further explored. But, um, as where you were mentioning before just artificial intelligence, the artificial intelligence is becoming more and more integrated into the EMRs sorry acronyms the Electronic Medical Records that a lot of organizations are using throughout the United States, probably throughout the world.
Dr. Michael Koren:
19:32
Oh, no doubt.
Dr. Mike Hafran:
19:34
Yeah, and as well as what I had mentioned, you know, with devices, more and more. Not only are there more and more devices, you know, this isn't any longer the internet of things, it's more like the internet of the human body. Now, how many devices can one wear? And I'm not saying they're all bad, not at all,
Dr. Michael Koren:
19:59
No, no.
Dr. Michael Koren:
20:01
I'm bullish on the technology. I just think that you have to prove it works.
Dr. Mike Hafran:
20:05
Exactly.
Dr. Michael Koren:
20:06
That it's intelligence, not stupidity.
Dr. Mike Hafran:
20:08
Exactly
Dr. Michael Koren:
20:10
I was at the American Heart Association meeting I know that you attended it as well and I saw a presentation I don't know if you went to this, but there was a fellow that was talking about using AI artificial intelligence algorithms to determine, based on an electrocardiogram, if somebody was having a heart attack, and making the point that, unless you have a very experienced cardiologist, that mistakes are made. Sometimes you don't recognize a heart attack, or sometimes you think it's a heart attack and it's something else. So they're hoping that AI can do a better job than the average emergency room physician or the average ambulance person. And the concept being is that AI can be everywhere. It can be in an ambulance, it could be in a remote ER, it could be in a remote physician's office. So that was all good.
Dr. Michael Koren:
20:52
But then I got to thinking well, if you're replacing cardiologists with this machine, what are the unintended consequences? And so is the machine that's looking at EKG is going to be able to say well, you're having chest pain and this EKG is not particularly remarkable, but did you check to see if it's a bleeding ulcer? Did you check to see if this is a pulmonary embolism? What else could this be, and so when you use technology that's trained to make a decision that's a relatively narrow decision you may lose part of human intelligence, which is to create a more broad-based understanding of the presentation.
Dr. Michael Koren:
21:36
So that would be an example
Dr. Mike Hafran:
21:38
Yeah, and actually I did not attend that particular meeting, but that's fascinating. I ended up going to view some of the pitches that startups were-.
Dr. Michael Koren:
21:51
That was interesting. Yeah, yeah
Dr. Mike Hafran:
21:53
That was a really interesting uh and it was all technology driven.
Dr. Mike Hafran:
21:58
But to your point-
Dr. Michael Koren:
21:59
Were they willing to test it in clinical trials?
Dr. Mike Hafran:
22:03
No, probably didn't want to do that not at this stage, which is, you know, and, um, I I definitely don't want to forget this thought because you just triggered this in me, so I will get to that, but I wanted to backtrack just a little bit and say that I agree with you 100% that, in terms of the ability and speed to process large amounts of data, AI and computers will always be- will always, you know always be far more efficient-
Dr. Michael Koren:
22:37
Yeah, for sure.
Dr. Mike Hafran:
22:38
Then, you know, our little, our flesh, driven um mini computers,
Dr. Michael Koren:
22:46
Right
Dr. Mike Hafran:
22:48
But to your point it's, it really is. You can't remove the human factor. You can't remove the cardiologist, you can't remove the physician, because the computer cannot develop context, um and that and sometimes that context and perspective. That is the total difference. You were just talking about how there's a drive for artificial or just remote, not remote-controlled cars, self-driven cars, autonomous driving, yeah, a Robotex, exactly.
Dr. Mike Hafran:
23:25
And you had brought up this scenario of, well, what happens if an individual is actually experiencing a heart attack or having some sort of health crisis in the confines of that car? You know what is the? What is the car's MO? It's to get you to the destination
Dr. Michael Koren:
23:44
Right.
Dr. Mike Hafran:
23:44
It is not going to. And again, these are fascinating things. Have they really been explored? You know what does happen if there is a health crisis in the car? Will the car have enough intelligence or understand that perspective to be like, hey, you know what? I need to stop and get an ambulance? We don't know, because the AI, to a certain degree, really isn't being vetted, it's being trusted and I certainly hope you know I don't come across as like a technological troglodyte, because I actually love technology.
Dr. Michael Koren:
24:21
Same Here.
Dr. Mike Hafran:
24:21
I think there's a lot of benefits. It's just, we need to be no more-
Dr. Michael Koren:
24:24
I've been accused of being a technology Neanderthal, because I ask questions but-.
Dr. Mike Hafran:
24:28
Exactly,
Dr. Michael Koren:
24:29
I'm like you. I'm very supportive.
Dr. Mike Hafran:
24:31
We need to be diligent.
Dr. Michael Koren:
24:31
Yeah, well, yeah, it's so interesting. So one of my theories is that human intelligence is fundamentally different than artificial intelligence because we have a limbic system and we have this. And if you're talking about Columbia University I think his name is Daniel Kahneman won the Nobel Prize for describing fast brain, slow brain.
Dr. Michael Koren:
24:55
I may be pronouncing that wrong, but I'll get it right and then I'll pronounce it well and then we'll edit that in. But I think his name is Daniel Kahneman and he brought this whole point up of fast thinking and slow thinking. So our instincts are our fast thinking and our brain is not as fast as a computer, but that's our slow thinking. But we have this limbic system. So all of our decisions are based by going through a filter of fear which comes from our limbic system and also goes through a filter of self-benefit. How can we benefit ourselves? And that's something that computers don't do.
Dr. Michael Koren:
25:32
So, using, for example, the scenario of what the robo-taxis do when somebody has chest pain in the cab, are they going to take somebody to the emergency room? And then, if people figure out that they'll take you to the emergency room if you have chest pain, then people without a heart condition that want to save money will say oh well, the hospital is closer to where I want to go. So I'm going to say I have chest pain, so they drop me off at the hospital rather than this place that's closer and I'll save 10 bucks. So people figure out how to cheat and the AI is going to have to figure that out as well. So when you get into all these things, when you're dealing with human intelligence versus artificial intelligence, some fascinating things can happen.
Dr. Mike Hafran:
26:14
Absolutely, absolutely. And first of all I have to really kind of just make a push out there that, yes, fast thinking to slow thinking, everyone should read the book and actually anyone that's a programmer or working on AI should definitely read that book. But you know the other funny thing is and I have this discussion with my brother who also is in the technology realm, not in medical technology, but it's a funny little thing, but I think it says a lot. When can a computer, can a computer ever kind of? Um, I'm kind of at a loss for words.
Dr. Mike Hafran:
27:04
The the discussion was pretty funny and it and um, it comes down to this what scenario will a computer ever agree that one plus one equals something other than two? Because it's like what can I, what can I? A computer or AI, it's, it's somewhat linear, it's massively linear in in thought. One plus one always equals two, right and um, you know it's like where can one plus one equals something other than that? Uh, like three, right, and this takes human creativity right, one plus one equals three. If you have like synergy, it's kind of a a little play on what a family is right.
Dr. Michael Koren:
27:56
No, I get it. No, it's great point, I totally get it, so it's abstract thinking and will computers be able to do that?
Dr. Michael Koren:
28:03
They'll get better and better at that. Um, we were just having a conversation that computers are now trying to sound more like humans and say, you know, they'll throw, I knows, ums into their conversation to make it make you think that you're talking to a human. But some of these other points is going to take a lot more time to really develop. They may eventually, but, um, we'll see where it goes, and it could go in a bad place, by the way. Um, so are you a star treck fan?
Dr. Mike Hafran:
28:29
um, yes, I'm familiar with the um, with the series, but both the traditional as well as kind of like the newer one.
Dr. Michael Koren:
28:37
So there's a great episode of the original series with William Shatner as Kirk, and they go to this planet and they see all these people lined up going into this building and there's literally thousands of people that are going to the building, but there's no one coming out of the other side of the building.
Dr. Michael Koren:
28:54
So they start to explore and they find out that this is a society where they were constantly having wars, and the wars were getting really, really bloody. And then they were developing computer capabilities that were getting smarter and smarter Again. This was done during the 1960s, by the way, and they decided that fighting wars was just too unseemly. They had to do a better job at fighting wars. So what they started to do is they fought wars by computer and then, after the war was over, then they would kill the appropriate number of people based on what the simulation was showing. And what Kirk and Spock saw was actually the end result of a computer war where they had to settle up.
Dr. Mike Hafran:
29:35
Wow.
Dr. Michael Koren:
29:36
So this is truly the tail wagging the dog and computers creating the reality rather than trying to help us navigate reality.
Dr. Mike Hafran:
29:45
Absolutely, and you bring up a really good question, and now I'm very curious about this too. You had talked about how the simulations are, the, I suppose, the human renderings of you know. The AI interacting with people is becoming better and better. My question is would people, would patients, feel that they're getting the same quality of care from a, we'll just say, a synthetic or a silicon-based provider versus a human provider? And we should be able to. Well, there's the emotional side, but objectively too, you know, we should be able to measure and determine how those things scale out, because that's truly fascinating, because we all know that there's a physician shortage in the United States. Is the answer, well, okay, we'll have some synthetic, um, and I I hope that doesn't sound derogatory to anybody
Dr. Michael Koren:
31:00
no, no, I think
Dr. Mike Hafran:
31:02
AI people, but you know, Is that a solution, or is it going to cause, like you said, more collateral damage?
Dr. Michael Koren:
31:11
I think it's spot on. I think one of the mistakes that were made, certainly in the United States, is this notion that we can provide everything for everyone. I don't think we can. I think that maybe we decide what is baseline decent healthcare, and maybe baseline decent healthcare is an AI bot. Maybe that's what everybody gets as their primary care physician, and if you want more than that, you pay up for it. Maybe the world turns into that. I'm okay with that. Actually, I'm okay with that. Believe that it's going to take a long time for a bot to reproduce the experience that human beings have with each other. That's my own personal opinion, but let the test run and see.
Dr. Mike Hafran:
31:51
Well, now this is a question of how much is technology or culture or other things pushing things without the validity of more like an objective, real measure? So there are companies out there currently with you know, health bot apps that are trying to grow their market share and positively impact people. I do know that most of those companies have been struggling and I'm not saying that's a bad idea. I don't think you're saying it's a bad idea.
Dr. Michael Koren:
32:28
No, not at all
Dr. Mike Hafran:
32:30
Can we actually objectively A: measure but also understand what the impact is? You know it's probably more like a combination, but a narrow range of where the bot works and a range where the human works.
Dr. Michael Koren:
32:52
Well, I'll give you an example. I manage some of my own financial portfolio, and when I first started learning about this, I would call up a broker and I would explain what I was interested in, and then I would ask the broker to execute the trade for me, and that would take 10 or 15 minutes and you go through the whole rigmarole about making sure you heard everything correctly and putting it into a system, et cetera. Nowadays I do it all online. I don't use a human being, and I think that's probably better. Now there may be circumstances where I want the advice of a human that's an expert in the field and have a bunch of questions, but I should pay up for that. But in terms of day-to-day interactions, I'm very happy to use the computer rather than a human for that. So I think this will all play out in a similar way
Dr. Mike Hafran:
33:37
You know that's kind of what our approach was with the med tech company I was involved with.
Dr. Mike Hafran:
33:55
It's utilize a very robust and reliable technology to manage day-to-day as well as keep an individual motivated and excited. Well, to make sure they were on task. When it came to motivation and excitement, that really came from our nurses and nurse practitioners reaching out and interacting with the patient, but again, with a limited number of human beings, that was perhaps maybe only twice a month, versus the AI was able to operate on a daily basis, regardless of what the patient's schedule was. So you know, that is another great way where, even looking to integrate some of the AI and technology into clinical research uh, now, I know we we were just talking about we should do. We should be seriously involved in clinical research on AI and its integration, because, uh, financial technology went the same way, just a sudden shift over. Is it good, as you mentioned, for you and actually for a lot of people? There are some very good aspects, but can it also be bad, such as the rapid sell-off that was AI-driven and then they had to shut down?
Dr. Michael Koren:
35:23
Yeah, and the technology doesn't always work. So I'll share a quick financial services anecdote with you. So I had an old brokerage account that was fairly inactive, that had six positions, and I called to liquidate the account. So we needed to close out each position and then to send me a check for the proceeds. The person who I spoke with said oh well, our price to do that with the assistance of a broker is $800 per trade. So that would have meant about $5,000 cost to close out this account with only six positions. But the person said if you get your online capabilities up to speed with this account, then you can do it yourself for $20 a trade. So I said well, that's a no brainer, just get me up to speed. I'm familiar with online trading, I can do this.
Dr. Michael Koren:
36:12
So unfortunately, it took about three or four weeks to get everything in place to do online trading for this particular account, multiple conversations being on hold, et cetera, et cetera. Finally, we got to the point where it was operational. So I get on to do it and I can't figure out how to do it, even though I've done it in other platforms, and there's this error message that I keep on getting. So I call the help desk again and I explain what's going on, and the person there says oh yeah, well, we've been having a glitch in the system. Let me just help you do it and we'll waive the $800 fee.
Dr. Michael Koren:
36:47
So after a month of trying to get this thing up and going and use it myself, at the end of the day they did it the old way and didn't charge me. Of course that could have been done in the very beginning and saved all the rigmarole for everybody. But these systems don't always work. But that's the beauty of clinical trials, moving in that other direction is that we record everything.
Dr. Michael Koren:
37:12
So when it works, we record it, when it doesn't work, we record it and we look at the net benefit.
Dr. Mike Hafran:
37:14
Absolutely, absolutely. And now you made me realize something else, and I think there's a lot of concern around security and safety of information, especially with a lot more programs, hospital programs and everything utilizing large databases that continually seem to get hacked. I think that also there's a larger challenge that really we need to also look into, and that's kind of like the, the algorithmic black box, and meaning you know, algorithms are really programmed right now by humans, maybe with some artificial assistance behind that, but any biases and kind of I'm going to just they're not quite errors, but any sort of particular thought processes that are programmed in will replicate and spread and sometimes get enhanced within the program.
Dr. Michael Koren:
38:28
If there's a problem in the system, it can be repeated over and over and over again before somebody realizes what's going on. That's a great point. And then, we have so much software that comes from different places that we have a lot of incompatibility problems. I'm sure you've had those issues.
Dr. Mike Hafran:
38:45
A hundred percent, a hundred percent.
Dr. Mike Hafran:
38:49
Yeah, Go ahead please.
Dr. Michael Koren:
38:50
I was going to leave you with one really interesting example of studying AI in a clinical trial, and then I'll give you the last word.
Dr. Michael Koren:
38:57
Okay, so we're working with a company that is very interested in using an app that a patient talks to on a daily basis and determines if the person who has a history of congestive heart failure is having a worsening of his or her situation by virtue of changes in their voice pattern over a period of time, and I think computers are really good at looking at changes.
Dr. Michael Koren:
39:23
So, for example, I went to a nice meeting about a year ago showing what AI could do, and they were pointing out that if you look at a skin lesion and want to determine if it's melanoma or not, that AI can do just as well as a dermatologist these days. But what it isn't doing yet that AI would do better than a dermatologist, quite frankly, is to look at changes over time. So you can't see your dermatologist every day, but you can take a picture of your lesion every day on your cell phone. So here's an example where there may be an intrinsic benefit to using technology, and that's what we have to help the AI developers do is to develop these things in ways that make sense from a clinical mindset.
Dr. Mike Hafran:
40:13
Absolutely, absolutely, and that is really fascinating, some of these technologies you're bringing up, because I've been exposed to them as well and involved in some peripheral ways with some development of these things. But you know, I know GE has always been interested, or has been interested, in developing kind of like the smart patient room. And think of this as a hospital room where it's digitally driven, almost like the way your house is nowadays, but of course the drivers or the owners of the room would be licensed medical professionals utilizing technology in a way that's standardized, and it monitors not just everything about the patient in the room but it also monitors, hey, what supplies are being used for financial reasons, so that the hospital can bill appropriately. And that's another wow, that's a whole other realm just getting it to artificial intelligence technology and appropriate billing. And we all know that the US, our health system, is kind of really facing a financial crunch in the next few years.
Dr. Michael Koren:
41:40
My bottom line is that both of us, our organizations, are open for business, whether it's AI, whether it's infusion of products that are not traditionally studied by the FDA, whether it's gender reassignment studies where you want to do randomized studies, Whatever it may be that's nontraditional we're willing to take this concept of a randomized clinical trial and apply it to that space for the benefit of science and for our patients, and that's the bottom line is we're able to do that. So why don't you give us the last word on this, Mike? Again, I appreciate your insights. It's been amazing.
Dr. Mike Hafran:
42:15
Well, thank you and I'm not sure. One I really enjoyed the experience. And number two I'm not sure I can beat your bottom line there, because it truly is in alignment with what I believe Our job is as modern physicians. You know, we're no longer like in barbershops, sawing off legs and stuff.
Dr. Michael Koren:
42:43
I did get approached to do a study about that, but I rejected it.
Dr. Mike Hafran:
42:48
Our job is, I believe, as clinicians and physicians, our job is always to take care of people, our patients in the world and the situation they find themselves in. In today and the modern practice of medicine in the United States, there is an increasing utilization of very intelligent technologies as well, as we'll say, kind of like non-traditional approaches, and that is how things move forward. But I suppose I'm asking that you know, we as clinicians also play a role in determining the validity as well as the usefulness and effectiveness for the safety of our patients. And very much just echoing what you said, you know I am involved with an organization that's very interested in continuing to push the frontiers of education and research and really critically look at some of these things that go beyond kind of the traditional aspect, because that's where we need to go, that's where we need to meet the future.
Dr. Michael Koren:
44:14
Absolutely Well. I started our conversation by saying we were kindred spirits and I'll end it with the same sentiment. Mike, thank you for a great interview, thank you for your insights and we'll do this again sometime soon.
Dr. Mike Hafran:
44:27
Awesome. Thank you so much, really appreciate it.
