Michael D. Abramoff, MD, Ph.D., a visionary retina specialist, computer scientist, and entrepreneur. Dr. Abramoff shares his view of innovation as a series of small steps and how he works to ensure progress at every stage. He offers insight into diabetes care in the US, diabetic retinopathy, and the benefits of increased screenings. You’ll also learn about bioethics, the exciting future of digital diagnostics, and the first FDA-cleared autonomous AI diagnostic system.
Michael D. Abramoff, MD, Ph.D., a visionary retina specialist, computer scientist, and entrepreneur. Dr. Abramoff shares his view of innovation as a series of small steps and how he works to ensure progress at every stage.
More about our guest:
Michael D. Abramoff, MD, PhD, is a fellowship-trained retina specialist, computer scientist and entrepreneur. He is Founder and Executive Chairman of Digital Diagnostics, the first company ever to receive FDA clearance for an autonomous AI diagnostic system. To transform the quality, accessibility, and affordability of global healthcare through the automation of medical diagnosis and treatment. Dr. Abramoff (gold Fellow ARVO, Fellow IEEE) is the Robert C. Watzke, MD Professor of Ophthalmology and Visual Sciences at the University of Iowa, with a joint appointment in the College of Engineering.
Dr. Abramoff is the Founder and Executive Chairman of Digital Diagnostics, the Autonomous AI diagnostics company that was the first in any field of medicine to get FDA clearance for an autonomous AI. In primary care, the AI system can instantaneously diagnose diabetic retinopathy (including macular edema) at the point-of-care. Dr. Abramoff developed an ethical foundation for autonomous AI that was used during the design and validation, and regulatory and payment pathways for autonomous AI. As the author of over 350 peer-reviewed publications in this field, he has been cited over 42,000 times, and is the inventor on 20 issued patents and many patent applications. Dr. Abramoff has mentored dozens of engineering graduate students, ophthalmology residents, and retina fellows. His passion is to use AI to improve the affordability, accessibility and quality of care.
Know more about her and their company here:
------------------------------------------------------------
Episode Guide:
1:30 - What is Innovation?
2:40 - Breaking down the "steps"
6:11 - Ensuring progress in every step
10:15 - Diabetes Care in the US
12:57 - Diabetic Retinopathy
14:25 - Benefits of getting more people screened
15:57 - Bringing together different domains
17:22 - What is a bio-ethicist?
20:38 - What isn't innovation?
28:59 - What's next for the company?
--------------------------
OUTLAST Consulting offers professional development and strategic advisory services in the areas of innovation and diversity management.
Jared Simmons 00:05
Hello, and welcome to What is Innovation? The podcast that explores the reality of a word that is in danger of losing its meaning altogether. This podcast is produced by OUTLAST Consulting LLC, a boutique consultancy that helps companies use innovation principles to solve their toughest business problems. I'm your host, Jared Simmons, and I'm so excited to have Michael D. Abramoff, MD, PhD
Jared Simmons 00:31
Michael D. Abramoff, MD, PhD is a fellowship trained retina specialist, computer scientist and entrepreneur. He is founder and Executive Chairman of digital diagnostics, the first company ever to receive FDA clearance for an autonomous AI diagnostic system. Its mission is to transform the quality, accessibility and affordability of global healthcare through the automation of medical diagnosis and treatment. Dr. Abramoff is Robert C. Watzke, MD Professor of Ophthalmology and Visual Sciences at the University of Iowa with a joint appointment in the College of Engineering. As the author of over 350 peer reviewed publications in this field, he has been cited over 42,000 times, and is the inventor on 20 issued patents and many patent applications. Michael, thank you so much for joining us today. I'm so excited for this conversation.
Dr. Michael Abramoff 01:25
Jared, thanks so much for having me, I'm probably more excited than you are. Very excited.
Jared Simmons 01:30
Let's dive right in. What in your mind is innovation?
Dr. Michael Abramoff 01:35
For me, innovation is a number of small steps. It's not one brilliant insight that can be interesting and exciting. But I like to think about the impact you can make and then it's really a series of many steps. To give you an example, one innovation, you could say is to create an autonomous AI for healthcare where you have a computer making a medical diagnosis, that was pretty novel, but it's not enough to benefit patients. There are so many steps involved trading ethics, and we'll talk about it later, right? How do you design algorithm? How do you validate it? What you would compare it to? How do you make sure it works in workflow? How do you make sure it benefits patient? How do you make sure it doesn't harm patients? Who's going to pay for it? So many innovations needed to be created step after step after step after step because so many stakeholders involved. It may not be true; we discussed arts last time and creation and that's so exciting to me. That's a little bit different, maybe interesting what you would say there. But definitely in healthcare, there's so many steps involved to bring innovation to benefit patients and populations.
Jared Simmons 02:41
When you think about innovation and those steps, as we compare it to art, are you thinking about the steps in terms of progression? Or do they weave through different teams and people and things like that? What does a step look like in your mind?
Dr. Michael Abramoff 02:55
Ultimately, it's about more than your shelf. In my case, I had to create a system of ethics for AI in healthcare. I did this with other experts and bioethicists. We had to create a regulatory framework started with regulators like FDA, we had to create a reimbursement framework to make sure it got paid, and patients didn't have to pay for it themselves. So that involves payers like Medicare, and CMS and American Medical Association, professional societies, patient organizations. But again, no one has ever thought about how to reimburse for AI. No one has ever thought about how to regulate autonomous AI to design a clinical trial. How do we deal with bias and prevent it from occurring? I see. There's so many examples. It's many things that you need to innovate on and solve. How do you fit it into the workflow. Typically, if you're in clinic or in healthcare, you probably know that it's very resistant to change, it's very hard to change because time pressure, monetary pressure, little resources. So to change something, really make sure that you understand what's going to be ongoing and that requires many innovations as well. We spent years after we designed the AI and the algorithm, and we knew that that had a high accuracy higher than the clinician like me. But so making sure that someone with minimal training with a local skimmer, I could operate it because you can say, well, I have this magnificent AI and it performs really well. But it only works on the million dollar camera, very expensive hardware, that's not going to happen where it's needed, which is an under resourced areas where the clinics have no money and they cannot afford an expensive camera. So now you have nothing, right? No, it needs to work with a lower cost camera with someone who may be very skilled and under in health care, but those specifically taking images in our case or the retina.
Jared Simmons 04:41
So there's an accessibility aspect of it that has to be integrated.
Dr. Michael Abramoff 04:46
Oh, yeah. That you need to solve and that you innovate on, exactly.
Jared Simmons 04:50
So it's not just invention. It's not just being able to bring it to life but as being able to bring it to life in a way that makes it accessible for people to be able to use
Dr. Michael Abramoff 04:59
Yeah, there's 25 patents along the way. I mean, these are invisible innovations you call them or inventions but they are very different. In my case, I have a background in clinical medicine, but also in computer engineering, (and) in neuroscience. It took all of that, not necessarily every time, but definitely, in my case,
Jared Simmons 05:20
Because each of the steps in the innovation process that you were talking about, demanded a different type of expertise, or a different set of skills or a different view of the problem to kind of progress the overall effort. Is that fair?
Dr. Michael Abramoff 05:32
That's fair to say, and understanding what stakeholders care about. So patients care about certain things and when it gets health care providers and physicians look at it very differently. The payers right, typically will pay for the health care, we don't pay for ourselves to have yet another perspective. At this just look at it in a different way. You need to not only understand all of that, but also be able to work with them to implement it. Because if only one of the states in healthcare specifically, if one of them says no, we think this is unsafe, or we cannot pay for it is too expensive, or we don't trust it, you're done. over. They'll not get to patients.
Jared Simmons 06:11
So understanding an innovation is this stepwise process, and that it involves all these different skills, perspectives, stakeholders, how did you through this process, measure and ensure progress? But how did you know things were moving forward? And what did you do to make sure that things kept moving forward?
Dr. Michael Abramoff 06:29
Yeah, great question. So part of it is you need to be very deliberate about all of this. But part of it is also a little bit of trial and error, because it had never been done. Again, I went into the US FDA, (who) regulates medical devices, including AI in healthcare, where it can talk about that. It says in 2010, hey, I want this computer to make a diagnosis, because we knew from the scientific aspect that the algorithm works. He said, hohoho, let's talk and it spent eight years making sure that regulators were comfortable and we were comfortable with solving, how do you do the software development? How do you do the clinical trial? One big question we need to answer is what do you compare an AI to? Because typically, it's compared to physicians. Oh, wow. But physicians are not always correct. And in fact, I and my colleagues differ with each other in about 20 to 30% of cases, I differ with myself, when I look at the same patient two, three weeks later. From time to time, right? I'm not consistent, whereas a computer can be right. So when you compare an AI, the outputs, if this diagnostic AI to one or more clinicians, how do you know that the clinician is right, and AI is wrong? Right.
Jared Simmons 07:45
Right and it's also the accuracy versus precision question buried in there as well.
Dr. Michael Abramoff 07:50
Exactly. So what we shared is what matters to you as a patient is your clinical outcome, do I get better? What do I need to get better? How do I not go blind or die? So there's clinical outcome, and you have the same for populations, groups of people who also want to do better on their health. They care about that, payers care about that, they don't care so much better doctors agree or disagree. Do you care?
Jared Simmons 08:16
No.
Dr. Michael Abramoff 08:16
You don't want to know, you dead already, right? So make me live, tell me what to do. What we said and together with FDA, we continue to work on this regulatory aspect, together with FDA and other regulators around the world is, can I compare it to clinical outcome? Can it compare to AI that it actually predicts clinical outcome, and then you don't need to compare it to physicians anymore? And then actually, you can say, this AI is better, more accurate than clinicians.
Jared Simmons 08:44
So instead of saying is this, I'm going to use general terms. But if the human clinician is a tool, and AI is a tool, instead of saying our tool A and B comparable, you're saying do tools A and B generate comparable outcomes?
Dr. Michael Abramoff 09:00
Yeah, a little bit. But also, is the tool a actually good at making the right diagnosis?
Jared Simmons 09:10
Ah, so the implied assumption is that the tool a is good at creating those outcomes?
Dr. Michael Abramoff 09:17
Correct. So you don't look at whether it compares favorably or favorably to a clinician, you say, Well, does it actually positively affect outcome?
Jared Simmons 09:17
Hmm, that's so important. That's great.
Dr. Michael Abramoff 09:21
So then you a) you can say, well, it can actually do better on certain aspects. And b), you don't get the variability and conditions because I can tell you if three clinicians look at a patient, they will differ and 23% and so how do you manage that? There's way to manage it but typically what we said together with FDA is superior to compare to clinical outcome than to physicians agree or disagree, or other providers. I mean it's not about the physician. It's about the expert, right?
Jared Simmons 09:58
That's right. So let's take a half step back and just if you can frame up what we're talking about, just in general terms, we dove right into the definition of innovation being around a series of steps and those things, what world are we talking about steps in right now?
Dr. Michael Abramoff 10:13
Absolutely. So what we looked at was a very specific thing, we said people with diabetes, very common disease, bad disease, can have complications, including in the eyes of vision. It's called diabetic retinopathy, very bad disease, most important cause of blindness in the working age population in the US and many countries. So bad disease, we also knew already that to prevent it, and actually prevent it in almost all cases, and make sure these people don't go blind, we need to do diabetic eye exams. We need to get these people examined every year and that's the standard of care and it is not happening. Only 15% of the people who need these, all these people with diabetes more than 30 million in the US, only 15% so less than 1/6th. More than 80% are not getting the exams they need every year.
Jared Simmons 11:04
Unbelievable.
Dr. Michael Abramoff 11:05
And that's a needless preventable cause of blindness. Can autonomous AI, meaning an AI that makes your medical decision by itself helps with this? Can we bring the high quality diagnosis into places where the patient is rather than a patient having to come... I'm a retina specialist so I was trained to be the retinal surgeon and I see these patients and then examined them and usually they're fine, they don't need to be treated or manage. Some cases they do and you don't know before you examine them. Now we do it with an AI but that means the patient doesn't have to come to me. I'm expensive. You spent a long time waiting. I'm in rural Iowa but in a city, Detroit, wherever there's a lack of access. People don't have access to high quality eye care. So can I encapsulate this into a computer and have this done? Wherever is an outlet where the patient is, right where they get their diabetes care? Yes, that's what we did. But it took a long pause. So we said, well, here's this algorithm, can it do it? Can it do it accurately? Yes. Okay, is it safe enough? Good. Let's go to a regulator like FDA. How do we get reimbursed? How do ethicist feel about the computer rather than a human making a diagnosis? How do patients feel about it? Those are the innovations, the steps we were talking about at the beginning. So we now have a computer making a diagnosis anywhere the patient is getting diabetes care,
Jared Simmons 12:26
right and it's not a multimillion dollar machine that only works with a special person running it or anything like that. It's something that's been democratized to the point that it can be used.
Dr. Michael Abramoff 12:36
Exactly. There's the word, democratize. So what you're saying is so right rather than requiring specialists, either someone like me, or highly trained operators in the primary care clinic or endocrinology clinic, where to get to Diabetes Care, essentially, anyone can be trained to do this exam and to this diagnosis and we know it's accurate.
Jared Simmons 12:57
As you think about accuracy, are you thinking about it in terms of the AI looks at it and says this definitely is not diabetic retinopathy? Or this definitely is diabetic retinopathy? Is it a positive or negative? Does that question make sense?
Dr. Michael Abramoff 13:12
To make it very simple? No, you're so right. Because initially, we considered with FDA decades ago, maybe it needs to give a degree or some complicated number but primary care providers and people will provide in these underserved resource places, they have no time to consider these difficult things. So it's better if it just says yes, no. If it's yes, the disease's there, which is not a few percent, go get an eye exam, you probably need to be treated by someone like me. But that's a very different story to the patient because if you tell the patient, you know, your diabetes as well don't smoke, he's losing weight, they know, the usual things, then you also need to get an eye exam, they don't go; we know that's 15%. But now, if you say, tell those patients who actually needed Hey, you have a high risk of going blind to the next year, you really need to go. That's a very different story. We're seeing that in studies that, indeed, the patients who do need to go after the AI diagnosis in detail, mostly go, it's really changing.
Jared Simmons 14:14
Now. Wow. I would imagine you're learning a lot from this and it could potentially change the way you think about diabetic retinopathy, like detection and things like that. Could it have benefits to understanding the disease itself to have more people screened for it?
Dr. Michael Abramoff 14:29
Well, yes, I mean, mostly, it's to make sure that people who the need the care they deserve, get the care. One big thing is health disparities, where some races and ethnicities and rural groups do even worse because they don't have access to care. Now with AI, we have shown actually, that almost 100% get access to care where previously was it was like 6%. Yes, that is interesting, and new, and novel, and very exciting because again, remember at the beginning, I said it's a series of innovation steps, there's algorithm, but you need to do all these other things. Now we see yeah, was it worth it? Yeah, it's starting to be worth it. You mentioned research, we learn so much along the way how to implement AI in the workflow. What works, what doesn't work? That's why I founded the company in 2010, the digital diagnostics, because you need so many experts in so many different fields that all come together more than 100 people, very different aspects, clinical workflow, algorithm design, user interface design, etc. You need all of that, one person can never ever do this,
Jared Simmons 15:32
So building on that, is there a secret to being able to bring all those different types of expertise and experience together in a way that's productive and doesn't drive stagnation? We've all been in the room where different experts and different people have different focus areas and different priorities, and it creates gridlock. How were you able to bring folks together from so many different domains? And have everybody working productively and in a positive direction?
Dr. Michael Abramoff 15:57
Well, so two things I mean, digital diagnostics is very successful, because it operates so well. Once we say, well, we know now what to do, how to bring it into workflow, we're doing that very successfully everywhere. Before that, I was trying to explain in terms of stakeholders, you need to get their acceptance that means listening. So when you talk about, especially a few years ago, now we'll check UBT and all the LLM things that you heard about the last few months or last last half year? You saw it again. If you talk about computer's doing health care, or anywhere people get concerned, is this safe? Is there racial bias? How are we going to pay for it? What about legal liability? How do we know it actually improves outcomes rather than make them worse? What happens to my data? Who owns all of this? So many different issues that people have, you can say two things, you can say, I don't care? I love the technology, let's just do it. That may work sometimes but then become very reactive, because usually there's a problem later on and then you say, Oh, well, we should maybe have foreseen it. Let's try to solve it now. You're being reactive, like we saw on social media. We started with social media. There's some evidence of harm in some areas and now we're trying to solve that, instead of from the outset saying, Well, if we think we're going to do this, what is going to come up? That's where ethics comes in. Ethics is 1000s of years old.
Jared Simmons 17:26
I was just going to ask what is a bio-ethicist?
Dr. Michael Abramoff 17:30
I'm not a bio-ethicist.
Jared Simmons 17:31
No, yeah, I know, you mentioned that you could collaborated with them.
Dr. Michael Abramoff 17:34
Yeah,
Jared Simmons 17:34
I understood all the different roles. How does a bio-ethicist play into this role? What is it? And how does it factor into what you do?
Dr. Michael Abramoff 17:41
Ethics, in a way is across cultures and across generations, for 1000s of years, how we thought about how to organize life around us. What is good and bad, what is right and wrong and religion plays a role there. But mostly, it says, seems like the golden rule, don't do to someone else what you don't want to have it to yourself. So there's things like benefits do no harm. So individual benefit, there's something called justice, which is, how does it affect groups? Is there a difference? You know, do you only care about one person? Or do you care about the group, the social aspect? autonomy? Can you make your decisions about in this case, your own health care, or is only someone else making those decisions? So bio-ethicists recognize these principles. What we came up working together with the bio-ethicists is well, if an AI, you can say, well, this is an ethical AI. But as an engineer, that is not enough. How do you measure that? You want to be able to measure it. It's just words, if you say, well, we did this ethically. Well, what do you mean, right? Show me? Can I measure it? The big step we took was say, well, we can measure how much it benefits groups. Justice aspects. We can measure how much it affects autonomy, how much it affects patient outcome. We already mentioned that and many other aspects. You can never be perfect. You can never hit all of these three at the same time.
Jared Simmons 19:07
Of course not. That's right.
Dr. Michael Abramoff 19:08
You need to find a balance and that's not, ethicists don't do that, that is society as a whole or the healthcare profession or whatever. But we at least were able to now, measure it and we were able to say, well, in this ethical principle, we're so good. We're about 80 here, about 95 here. Now you can see things coming like the issues I mentioned at the beginning was going to happen to my data? what about autonomy of the patients? etc, etc. We could see them coming and we could build AI so we could afford those issues from the start rather than having to catch up at the end. I think that was different. With that and having built all of that, which is digital diagnostics. We went to the different stakeholders, we said to payers, look, here's this AI makes a medical decision. I know that sounds scary but here's what we did to make sure it's not scary and they got it. It made sense and we'll work together payers, patient organizations,
Jared Simmons 20:07
That's such an important point there is the importance of crafting a story that resonates with different stakeholders, because each of those groups that you mentioned, has a different perspective on the problem and a different mindset about what they think is going to work or where their issues would potentially be with what you're describing. I would imagine, I mean, I know that there's an art to having that conversation the right way with the different stakeholders on what they care about.
Dr. Michael Abramoff 20:34
Yeah, exactly.
Jared Simmons 20:36
So we talked a lot about what innovation is, what isn't innovation, and why does that matter?
Dr. Michael Abramoff 20:42
In healthcare, right? Because again, I'm sure interested in your thoughts about art.
Jared Simmons 20:45
Yes. Oh, please, for sure. All of it.
Dr. Michael Abramoff 20:48
What I call glamour AI is really cool. I love coding and like you. I mean, we talked about this. So glamour AI is where you have a really cool technological solution, which doesn't benefit patients where there's no evidence, where there's no proof that it actually does anything, move the needle on patient outcome or population outcome, or health equity, or cost, or whatever, or quality of care. Glamour innovation is like this cool thing. Everyone said, I admire that, Oh, wow, that you did that and it never gets to patients and there's so much of that. That's why I mentioned the beginning that it's the steps that many things need to be innovated to make real change happen.
Jared Simmons 21:27
And to recognize what's truly innovation is what gets to the patient and generates an outcome not what's new or novel or leverages a new technology, that in and of itself doesn't drive the value in the outcomes.
Dr. Michael Abramoff 21:41
No, because of that specific design of the algorithm and how we mimic what the brain of a clinician does. That was several patterns. But you could call an innovation, maybe that would be glamour innovation if it remained at that, but then with digital diagnostics, we had to work on how to make sure that this works on the locals camera, in an environment where people aren't used to doing this and don't do it all day in like, academic clinical environment. So these were our other innovations. Then no one has ever thought about how to reimburse this so we had to do other innovations.
Jared Simmons 22:16
That's exactly right. What's great about that is, it's not just about accessibility. The democratization isn't just about accessibility. It's also about robustness. Because to use a lower quality camera and to have higher variability in the users, you have to have a more robust solution. That's also a core element of innovation that I think a lot of people overlook, I made this new amazing thing that does this amazing thing, but it only does it if you do this specific thing, that specific way, stand on one foot on Tuesday afternoons, and then it does this great thing and that's not repeatable. It doesn't lend itself to democratization.
Dr. Michael Abramoff 22:56
It doesn't scale.
Jared Simmons 22:57
It doesn't scale. That's exactly right.
Dr. Michael Abramoff 22:59
It doesn't scale, man. Yeah, you can make this magnificent invention and if it doesn't scale, I don't want to pay for it as a taxpayer, right? I mean, it's right. It should be dead in the water. It needs to scale.
Jared Simmons 23:11
That's the danger, I think of kind of staying in our technological bubbles. Because we all geek out on different aspects of innovation and things. It's easy for me to say to another person who's into my kind of innovation. Oh, wow, that's really cool. That's great. And we can sit in our little bubble and pat each other on the back. Meanwhile, nothing is changing for society. No outcomes are being changed for humanity. I think it's why it's so important that we connect across domains, but I don't do anything in healthcare. So that's why I think it's so important that innovators connect across domains to have that perspective and get that push to say, oh, yeah, that is cool. How is that? How does that drive outcomes?
Dr. Michael Abramoff 23:58
I was thinking while you were saying that, it also you need to get your hands dirty. If you want this to happen. You need to be so you can design these algorithms and never go into a clinic and never know what the patient is and never know what the provider does, what the payer does. So academic, it's ivory tower. But if you get your hands dirty, you realize, well, I actually, this may be sounding like a great algorithm, but that actually it cannot work in this environment, or no one will ever reimburse it or no clinician will accept it. You need to solve for all of that, you need to get your hands dirty, and really feel what all the stakeholders are dealing with.
Jared Simmons 24:34
Well said
Dr. Michael Abramoff 24:35
Medicare was really worried we're going to see a mess, right? Typically all the payers fall in many cases, they were really worried about blowing up the budget. If they start reimbursing AI, it can absorb the entire healthcare budget. Because there's so much cool technology. So they were really concerned among many, they were concerned about racial bias, about over utilization, about data use, but they mentioned this, how do you make sure that it actually is cost saving? and so, so many aspects but you need to get your hands dirty and be willing to get your hands dirty.
Jared Simmons 25:12
There's such an ivory tower as a perfect phrase. It's such an ivory tower kind of view of innovation and what's innovative, and even when you just say AI, I think people just start thinking pie in the sky, conceptual sorts of things. But what makes it create outcomes, what makes innovation change the world is getting your hands dirty, and getting down and understanding the stakeholders and understanding the problem that the in this case, the patient level, or to borrow the terms the consumer, the user level that's where innovation really lives.
Dr. Michael Abramoff 25:42
Yeah, you're sure right.
Jared Simmons 25:44
So the piece of our own art, because we started down that path around comparing art to technologically driven innovation and I do think there is a step-wise element to it. First, you choose the size of your canvas, and you select your paints, and you decide what you're going to paint and all those things, those sorts of steps in the process. Or if you're a musician, you have to have chosen an instrument, developed some amount of technique with that instrument, and then deciding what you want to express. Back to your definition of innovation. I think there are small steps along the way to creating art. The reason I started the conversation where we did around, you know, whether you think stakeholders are involved in those steps is I think that's the key difference. When you're trying to bring a technology to life, or a technologically driven outcome to life. I can paint a painting by myself in a studio. Well, I can't. But a person could paint a painting by themselves in the studio. But as we think about technologically driven or business oriented innovation, almost all those steps involve other people. There's this sort of misconception that Steve Jobs invented something.
Dr. Michael Abramoff 26:53
No, no, you hit the nail on the head. But I think it's closer than you think. Because you can, okay, first specific ideas about art and I love the art that you have behind you but you can make art for yourself or for your peers. You can make art for the larger public, and you can call it selling out. But what is right and wrong and sometimes it turns out 300 years later, yeah, what you made for yourself was actually now
Jared Simmons 27:17
valuable to other people.
Dr. Michael Abramoff 27:18
And it's hard to optimize at the moment. But one is not wrong and the other is right. You can make really cool algorithms, glamour algorithms. But if you want to benefit patients yet, then you need to, you know, in a way selling out, you need to really interact with those people and make it work for them rather than only for yourself or your peers, You mentioned the geeking out, which is your peers and if they say in the gallery, well, you know, your music or your your art is beautiful, but the public is not buying it. Okay. Well, it to me, it's very similar.
Jared Simmons 27:49
that is a great parallel because it's information, right? It's, I made the art for whatever purpose. But the separate question is, is it useful? And to whom?
Dr. Michael Abramoff 27:58
Yeah. And define useful?
Jared Simmons 28:00
Yes. And define useful? That's exactly right. I'm sure there's 1000s of masterpieces that we'll never see or hear that never see the light of day, because they're not made for us. That doesn't mean they're not useful. And that doesn't mean they didn't serve a purpose. They were just useful in a different context. Like you said, with these algorithms. I'm sure there are algorithms doing a lot of different things out in the world. Some of them are useful to society, and some of them aren't, but who's to say what useful means and how someone should spend their time. So yeah, I can definitely see the parallels there.
Dr. Michael Abramoff 28:35
In healthcare, I would say patient outcome, patient outcome.
Dr. Michael Abramoff 28:39
I mean, there. That's right, then, because you do it for someone else. Clearly, healthcare is not preset.
Jared Simmons 28:39
That's right.
Jared Simmons 28:46
That's right. Innovation writ large. Yes, you can do but you want to do for whomever. But when you bring it into the healthcare space, and start consuming resources, taxpayer dollars, other things, it has to be linked to patient outcomes. So Michael, we've talked about the work you all are doing in diabetic retinopathy. What's next for you and your organization? Where are you bringing this AI driven solution to next?
Dr. Michael Abramoff 29:11
Yeah, absolutely. So living in a score, which is this diabetic eye exam with with AI, a, make sure that every patient has access to it either the eye exam in the US and elsewhere. So that's first order of the day. And the second, we have two more AIs for AI diseases that we hope to put through trial soon. One more for cardiovascular disease that we've been working on for a while. Wow. Hopefully also starting trials very soon. So yeah, absolutely. The plan is to use the knowledge we now have about all these things, right. Stakeholder acceptance, reimbursements, regulatory, part how to build ais, in the right way to make sure it applies to other diseases, because there's so many diseases who can benefit from better access, fewer disparities, better outcomes, and AI can do a lot there. Not in all cases, but in many cases,
Jared Simmons 30:03
right. And the fact that you took such a process driven approach to the first approach should theoretically make these new applications. I won't say easier to minimize the work, but it should make stakeholder engagement a bit more of a streamlined process, I would imagine.
Dr. Michael Abramoff 30:20
Yeah, digital diagnostics, there's a fair deliberate step by step process on how to approach all of this. So yeah, of course, you need to prove the first one. We did and now expand beyond that, of course, it's the most exciting thing.
Jared Simmons 30:32
Fantastic. This has been an amazing conversation. Do you have any advice for innovators that based on everything you've done in having built this company and all the different domains of expertise you have? What advice would you have for innovators who want to make a difference and drive outcomes?
Dr. Michael Abramoff 30:49
So back to healthcare, at least tech, I think is, you can't do it alone and that sometimes it's hard in the beginning, I mean, I'm so proud of digital diagnostics, and these people are way better than I am at anything they do, right? I mean, it's just amazing to see 100 people all so brilliant, it's amazing,
Jared Simmons 31:08
unbelievable.
Dr. Michael Abramoff 31:09
But initially, you come up with these ideas or these innovations. And then typically people will do innovations, or let me speak for myself, I find it harder to work together with other people. So you need to reach out, you need to reach out and get your hands dirty. Because this academic ivory tower algorithm may sound great, but if you want to change the world, then you need to get your hands dirty. And then, in my view, the innovations will come because once you get your hands dirty, you realize, oh, wow, we need to do better here. Another innovation, another innovation, step wise
Jared Simmons 31:43
well said, gotta get your hands dirty. Understand the problems you want to solve and that will lead to the outcomes and highlight other opportunities for innovation. Just like you said, Michael Abramoff innovation is a series of small steps. Thank you so much for your time, and I hope we can stay in touch and hopefully chat again soon.
Dr. Michael Abramoff 32:02
Yeah, thanks so much, Jared I love your program and I'm very, very glad to be here.
Jared Simmons 32:06
All right, take care.
Jared Simmons 32:12
We'd love to hear your thoughts about this week's show. You can drop us a line on Twitter @outlastllc, or follow us on LinkedIn where we're @outlastconsulting. Until next time, keep innovating! Whatever that means.