SIGHTLINES Episode 8: AI & Simulation: A New Frontier Transcript

Hosted by Dr. Hunter Cherwek with guests Dr. Michael Abramoff, Dr. Nicolas Jaccard, Captain Cyndhi Berwyn, and Dr. Danny Haddad.

Dr. Hunter Cherwek: Welcome to Sightlines. I’m your host, Dr. Hunter Cherwek, Vice President of Clinical Services at Orbis International. Orbis is an international not-for-profit, and we’ve been working to address global blindness and vision impairment for about four decades. That’s because there are hundreds of millions of people in the world who are blind or visually impaired, and the vast majority of these cases could be treated or even completely prevented. In this series, we’re looking at how the pandemic has affected the fight against global blindness – and at how we’re navigating the situation using different technologies, creative solutions, and innovations.

One technology we’ve been developing for decades is paying off in a big way during the pandemic. That’s Cybersight, our e-learning and telemedicine platform. An important thing to keep in mind about Cybersight is that it only exists because we began investing in it decades ago, in the early days of the internet. So today, we’re going to look at the next frontier – the most recent advancements in technology, and their potential to help save sight around the world. Specifically, we’ll look at why Orbis has added AI, or artificial intelligence, to our toolkit in the fight against global blindness. We’ll also take a look at the role of simulation – how it is helping take training for eye health professionals to the next level and how that's facilitating training during this time of physical distancing.

My first two guests are Dr. Michael Abramoff, the Founder and Chairman of Digital Diagnostics, and Dr. Nicolas Jaccard, the Principal AI Architect at Orbis International. Here’s our conversation.


Dr. Hunter Cherwek: First up, I'd like to welcome Dr. Michael Abramoff. He's joining us from Iowa, where he is a professor at the university and the founder of an incredible company that is now bringing autonomous artificial intelligence to help treat eye diseases.

Michael, thank you so much for joining us today. Maybe you could tell us a little bit about your background: where you grow up, your academic background, and how you got started in artificial intelligence.

Dr. Michael Abramoff: From my accent you can hear I wasn't born in Iowa, I was born in the Netherlands, came here from Amsterdam 17 years ago. I have an interesting background, you may say, or people tell me, because I'm also a computer engineer, I worked for a long time in neuroscience so I can call myself a neuroscientist. Actually, I have a degree in epidemiology, but we won't talk about it, and so, with that background and where I was mimicking brain cells 30 years ago with neural networks, in a very simple neural network, one-layer neural networks that essentially mimic brains, oh sorry, neurons and their interactions. It was sort of obvious, as machine learning got off to a more efficient start, that when I became a resident in ophthalmology and so all these patients with diabetes weren't getting eye exams or, on the contrary, coming in very late where they also have irreversible damage and then, knowing that it's easily preventable if you catch it early, I thought, well, maybe a computer can do a better job and become also more accessible and at a lower cost.

So we've spent a lot of time essentially figuring out, and maybe we can talk about it later, what is the right way to validate AI—how should we build it, what should we actually be doing, how do we prove that—many details that hadn't been worked out and it culminated in 2018 with the FDA, saying: Well, we authorize this autonomous AI for marketing in the US, meaning you can now use it on patients. And now, for the first time ever, Medicare decided to pay for an autonomous AI for the diabetic eye exam, so that's probably a bigger achievement than even FDA clearance, because there's so much involved in these decisions. So on one hand, it shows that the health care system wasn't prepared at all, but it's now very ready and embracing it, because, really, everyone we work with on this road is very enthusiastic and they want to try to make it work, it's just at the health care system is also very complex and so it's just hard to change what was so far done by human doctors. I mean, usually many rules and regulations say, it needs to be a doctor, a human doing this, and then, you know, how do we change that.

Dr. Hunter Cherwek: Maybe you can just take a minute and make sure everyone is clear. When you say autonomous AI—when I put my head and have a picture taken of my eye—what does autonomous AI mean? I just want to make sure that's crystal clear for everyone.

Dr. Michael Abramoff: The exam is done right there within minutes, and there's no human oversight of the medical decision. It's still discussed with the patient, typically by a doctor, so the doctor is still in the picture, but it's not an ophthalmologist anymore or an optometrist. So it means the computer makes the medical decision, makes the call whether it is diabetic retinopathy.

Dr. Hunter Cherwek: Just so people know, I mean, you really are the first person in all of medicine, not just ophthalmology, all of medicine to have FDA approval for autonomous artificial intelligence. That's a huge milestone in all of medicine. And so, just congratulations to you and your team, we're so excited to partner with you and look at how we can do research together to improve productivity of eye health professionals around the world, so it's a really big honor to have you here with us on Sightlines. I also want to welcome the Cybersight principal architect for artificial intelligence. We also are using artificial intelligence on our platform Cybersight.

Nicolas, maybe you can tell us a little bit about yourself: where you grew up, where you are now during COVID and, very much like Dr. Abramoff, you yourself are an academic with a PhD, you're an entrepreneur, you're an engineer, so you two are going to get along very well and I might just drop off this call and let you all keep talking, because I don't think I'm going to be doing much here because, certainly, you all are going to be the intellectuals doing the heavy lifting.

Dr. Nicolas Jaccard: Thanks, Hunter. So, I'm currently based in London, but as you can hear from my accent, I'm not actually from London, from the UK. I actually grew up in Switzerland. And, originally, my undergrad[uate degree] was in biotechnology engineering, so I was actually working on things like protein production, vaccine production, and my career was set to be in pharma, but I had a chance to come to London for a six-month project at the end of my bachelor degree and I basically fell in love with London, first of all, so I decided to stay, and I was offered a place for a PhD at University College London.

So pretty much from the end of my PhD through now, I've been working with deep learning, this new kind of fancy machine learning approach. I was contacted by Orbis, and I was told that there was this position opening at Orbis. I didn't have to think much before accepting it because in my days in the first startup I worked at, I actually interacted with Orbis a bit there and the Orbis project that I've worked on became a kind of a passion project of mine and I really enjoyed the time spent working with Orbis. Coming to Orbis was such a huge amount of domain knowledge; every day I talk to someone new at Orbis and discover something new we are doing around the world is very inspiring. Again, it's been a dream job for me at Orbis and, yeah, I really enjoy it.

Dr. Hunter Cherwek: You get to work with me, I mean, what else could you want out of life? I mean, really! [laugh] So, you know, Michael, I think, I visited you in February, which is not the ideal time to go to Iowa when the snow and it was minus 10 degrees the morning I came to your office.

Dr. Michael Abramoff: Minus 10 Fahrenheit, for the listeners outside of United States. [laugh]

Dr. Hunter Cherwek: Oh, yes, yes. I apologize, that is correct. How very American of me and how very appropriate of you to correct me. I will say, you know, it was such an, just so exciting to walk in your office and feel your team's energy and how you want to globalize your product. Can you talk about how already you're not only working within the US but you have installations and partners in Poland: how has this built out, how fast is this growing, this field of autonomous AI that you are leading, how fast is this growing around the world, and how is this realizing your dream to help prevent avoidable blindness?

Dr. Michael Abramoff: So we are in one province in Poland, they're doing all the diabetic eye exams with autonomous AI. In Austria, in a pretty big way. In Holland, you know, obviously, because I've still connections there, and Germany and several places in Europe, Middle East comes to mind, of course, where we have some installations, and then, you know, very exciting. I don't know whether you want me to mention the name of the country where we're working together, but Bangladesh. What is especially exciting to me is, the reason I found that this company was to save health care costs and make quality better, and make it more accessible everywhere around the world, and productivity, improving productivity is really the key to that. We have never studied how autonomous AI benefits productivity, you have great expectations, but it would be so great to prove that, for the first time ever, and that's what we hope to do together and that's really so exciting to me.

Dr. Hunter Cherwek: Yeah, like you said, it's so exciting to see how technology is advancing so rapidly and, I mean, I can imagine scanning slides in the 80s. Now, with your iPhone, you're able to take better photos than people could 5 to 10 years ago, you know, it is exciting time to be alive with technology and you are strategically driving it to the places where Orbis works and where it's needed most, and I just want to, officially, thank you for that, because it's so exciting to be partnering and learning with you and working with you on Cybersight.

You know, Nicolas, I'd love to hear how your experience being part of this Cybersight team, how we're applying artificial intelligence to consults and more for machine mentoring—not independent, not like what Michael's doing with autonomous AI—how are we using AI and how have you seen that already helping the Orbis mission?

Dr. Nicolas Jaccard: Yep. So, as you mentioned, we're not touching autonomous AI for now, at least for the foreseeable future. We are mostly focusing on two aspects, which is providing clinical decision tools: 1) tools that help healthcare professionals in low- and middle-income countries to make the best possible decision when they see a patient, and 2) tools that can be useful for training or mentoring.

The way our users interact with AI in terms of diagnostics is so far through Cybersight Consult, which is our telemedicine platform, whereby a user, which is typically a healthcare professional at a low- to middle-income country, can submit a patient case to the system and, typically those are complex patient cases, where the local professional needs some help to really understand the case and come up with appropriate diagnosis and appropriate treatment for the patient.

And once the case is in the system, it is paired up with a mentor, typically an expert of technologist, for example, someone in the UK or somewhere in the US or Europe, and together they work on the case. So the benefit of this is two-fold: you have better outcome for the patient, through this collaboration, and also, through this collaboration, you have a kind of a mentoring and training of the local healthcare professional.

The way we integrated AI in this workflow is, instead of when you submit a patient case, you can opt in, so it's opt in, it’s not by default, you can opt in to also submit ophthalmic imagery, which are typically back of the eye image, for automated interpretation by the system. And what that does is, your case still goes to a mentor and the mentor will also see the AI outputs as part of that workflow, but, before you get a response for a mentor, you'll get an AI report, typically within a minute of submitting the patient case, so not only is it useful for diagnosis purposes but to show, to give the best possible care to the patient; it's also useful for the local health care professional to learn from it.

Dr. Hunter Cherwek: And so I think that's where the partnership with Dr. Abramoff's company works so well with Orbis's, you know, we're looking at AI for training and building up human resources and customized learning on Cybersight, when that's all being supervised by a human. Dr. Abramoff is really looking at patient care, direct diagnostics and, without a human supervisor, that his is autonomous.

You know, Michael, you obviously have been not just driving this from a medical standpoint, but from an ethics standpoint, from, you know, a regulatory standpoint, from a security and patient privacy standpoint. You all have been, both of you, have been looking at how AI is affecting the entire ecosystem. And so, maybe we can just talk for a bit. You start, Nicolas, because I know Michael talked about this with the evolution with his first ever FDA approval for AI, the ethics, security and privacy. Maybe we can just talk about how, as we bring out any new technology to the market into the public, we always need to think about those things. So, Nicolas, maybe you can start?

Dr. Nicolas Jaccard: I agree that these three things are some things that need to be thought about. As Michael said, it's not only something you say about, you know, at the end when you have something and, you know, you want to productize it and have us to use it. You want to think about it from the design phase, you want to build ethics, security and privacy into your product from the start into all your processes and this is something that part of it is also forced by regulation, so if you have a quality management system—some of these aspects are mandatory that you have to make sure that, you know, for data security and privacy concerns and so on— you have some regulations around these aspects.

But I think really the most important, I think the most spoken about of the three, is probably ethics and I think it's a very important to take the time to talk about it. We've seen, until even very recently here in the UK, when you had an algorithm, you know, giving grades to students and if he was a student from a deprived area, you will way more likely to get lower marks by this algorithm or lower grades compared to someone going to a prestigious school in London, or in the more affluent areas of the UK. And, this is an example of, basically, massive biases in your data sets. Just because, you know, historically in deprived areas, they didn't do as well, it means that, according to the algorithms, that they will never be able to do better because it only looks at, you know, recent historical trend or how it extrapolates.

I think this is very important to be able to think about these aspects from the very beginning of work on the new AI product. There are many different biases you may want to think about; one is, I think it's important healthcare, some measurement bias, whereby even before you go to the machine learning or artificial intelligence algorithms, the way you measure something can be biased depending on body ethnicity, so it's known that some imaging using dermatology, for example, will not give as good results when it was on a darker skin tones.

Dr. Michael Abramoff: I started worrying very early on about racial ethnic bias, and bias in general, that we built into our machine learning algorithms. For me, it was pretty obvious that if you build an AI for detecting diabetic retinopathy or other diseases, that you look for what clinicians already do, which is look for hemorrhages [unintel], so that's what we build machine learning, deep-learning-based detectors for. But that builds in a variance to race and ethnicity in the background color of the retina, because it doesn't matter what the background color of the retina is, if you have a hemorrhage and an [unintel] is likely diabetic retinopathy, so you avoid this whole problem of making sure, which is what in my view is impossible to guarantee that your training set contains all variance that this there in a network population, so that you are sure you are not excluding a certain race or ethnicity from, you know, from high performance. So that way we are able to address it even from the design, rather than just in the validation, which is as important of course, so I embrace very much this biomark approach, I really think that's the way to go where we can.

Dr. Nicolas Jaccard: This is becoming as important as the actual machine learning engineering is to really understand these ethical biases in the data and in a machine learning algorithm themselves.

Dr. Michael Abramoff: I think ethics is a tool, it doesn't tell you what to do, it's just a way of analyzing what you can do, so you still need to make decisions afterwards of how will I apply ethics to, in this case, AI. So, we've already talked about the biomarkers, which is, at least, you know, a safer way of doing it and the other way is, of course, validating. You can never examine your AI [on] all the population, it will be used on its 40 to 50 million people in the world, 30 million people in the US. So, you have to take a sample and that's accepted, but a sample, you know, it will never represent all of 30 million people, so you need to choose the sample wisely.

Dr. Hunter Cherwek: I think that's something, you know, I appreciate you both commenting on. Obviously, we need to think about this as humans as well as technicians and clinicians

I'd like to hear from maybe Michael first and then Nicolas, where do you see the field of autonomous AI in the next 10 years and where do you see it's applications and impact for what we do at Orbis and global ophthalmology?

Dr. Michael Abramoff: I expect to have autonomous AIs for at least, maybe, 10 diseases and let me explain why I said it, and then expand a bit. So, it needs to be scalable, it needs to be a massive problem; otherwise, it's so much investment and effort and cost in validation and getting the training data and doing the clinical trials that it may not be worth it for, i.e.; Leber's Congenital Amaurosis with 3000 subjects in the entire world, maybe 5,000. And so, you will see it for diseases—like glaucoma, diabetic retinopathy, maybe a sickle cell, maybe, you know, AMD and other diseases like that, malarial retinopathy—those are the ones that caused the most suffering and the most, in this case for Orbis, the blindness. So, I think that's where we will be going, that's where it makes the most sense. And then, aside from that, there is the other challenge that yes we are now really good at diagnosing people who are at higher risk—like I mentioned earlier, we still need to find solutions what do we do these people, wants they're identified, do we have the healthcare, the ophthalmology system.

And that's, I think, you know when residents and fellows asked me: What should I do, you're creating these AIs and they're taking away all the jobs. I say: No, because we need more people, in fact, to treat these patients and to treat them well in an evidence-based way. So, I think that's where more and more focus on Orbis will be because, you know, like you're already doing, educating the best of ophthalmologists, the best other ancillary personnel, so they can deal with all this massive wave of patients that deserve to be treated.

Dr. Hunter Cherwek: No, I appreciate that. Nicolas, maybe you can tell us, you know, where you see both the field of autonomous AI and then, as Michael has alluded to, where you see Cybersight AI for training. Maybe you can talk about both those for diagnostics with autonomous and training and machine mentoring with Cybersight AI, maybe you can talk about that, where you see us in 10 years?

Dr. Nicolas Jaccard: I do think that will be, kind of a period of time where there will be limited improvement of the technology itself until there will be our next boom, maybe in 10 years or 15 years, 20 years. I do think, though, that where all the improvements will come is how we apply this technology that we already have, but also how we integrate it in, you know, day to day workflows of clinicians, be it in Western countries or more in developing countries or low- and middle-income countries; how do we make it as streamlined as possible, so that using it is actually not a bottleneck or yet another step in the process, but just built into the process as they get the benefits of it without any negative impact on their day to day work.

So, I think understanding how AI helps and the way it helps, as it was mentioned a few times during this conversation, is that there's low hanging fruits and we know that having AI for certain things like DR diagnosis will help a lot. But I do feel like there is a lot of areas where we think AI is going to help—is the effect will be actually be negligible and other areas where we may not think that having this type of technology can actually be helpful, we turned out to be where it actually makes the most difference. I'm sure in five years, AI will be embedded in learning and teaching in ways that we can't even imagine now, because, as I said, I think we are biased in our ways that we think we know how we want to apply AI, it will turn out that there will be many, many use cases that haven't even cross our mind at this stage, and it will seem so obvious then, but I'm hopeful and excited that Orbis will be part of that journey on discovering where AI will actually be useful in our mission to really teach and mentor these healthcare professionals in the communities.

Dr. Hunter Cherwek: I just want to take a minute and thank you both. Like I said, I learned something from you all every day, the Cybersight team is just such a fascinating team and not trying to separate the digital divide, but close it so that we can democratize and globalize game-changing technology, like what Dr. Abramoff, and what you are doing with the Cybersight team. For me that's the greatest pleasure I've had in medicine.

Dr. Hunter Cherwek: That was Dr. Michael Abramoff and Dr. Nicolas Jaccard, two absolutely brilliant minds. I want to thank them both for joining us on Sightlines.

My next guest is Captain Cyndhi Berwyn. Not only is Captain Berwyn an incredible FedEx pilot who volunteers her time to fly the Orbis Flying Eye Hospital, she is also an expert on simulation. I’m so pleased to have her as a guest today to discuss how simulation is playing an important role in advancing training and education—even from afar.

Cyndhi, I can't thank you enough for joining us today. How are you, and how are things going with you and your family during this COVID time?

Cyndhi Berwyn: I'm doing really great. I just had a new granddaughter that was born August 30 in North Carolina. So now I have three granddaughters and the family's doing great. You know, the virus has slowed things down, but not for FedEx. We're still busy working, so my paid job is still keeping me really, really busy. We get to do a lot of humanitarian relief, that's kind of cool, sort of relates to the Orbis mission. And, keeping my flying current, that's for sure, I'm still all over the world.

Dr. Hunter Cherwek: Yes. Well, tell us a little bit about your background: how you got started in aviation, what drew you to the field, and what keeps you going, because you've accomplished so much in aviation, what keeps you driving as a perpetual student of aviation?

Cyndhi Berwyn: So that's always a hard question for me to answer with a smooth answer, because I can't remember exactly what got me interested. I remember as a child, dreaming that I could fly—you know, personally fly—and maybe that was a safe space for me, you know, because it's a way of getting away from people that are not so nice to you, or whatever, I don't know. It just was really a cool thing to look down. I used to love to hike, I still do, but just kind of a neat thing and then the opportunity presented itself when I was a teenager to go out and learn to fly. Joined Civil Air Patrol, got involved with the glider flying and small airplane flying, started earning my certificates, paid my way through college teaching flying. And then, the year I was a senior in college, the Air Force decided to accept women into pilot training, and I applied in March, graduated in May, and was accepted with the first class of women that went directly from civilian, there were a couple of classes ahead of me that were previous military, but the first class that came from civilian to enter as a pilot.

[I] did that, top flying in the Air Force for a few years, then ended up flying with the Reserves— the Air Force Reserves— the KC 10s, the similar airplane to the Orbis airplane, and got hired by FedEx to fly for them. So, I've been with them [FedEx] for 35 years and I've been volunteering with Orbis since 2012, I believe it was. So, just love it, love the aviation, and I've done a lot along the way—I've flown seaplanes, sail planes, hot air balloons, the Goodyear blimp, helicopters—got ratings in a lot of different things. It's just fun.

Dr. Hunter Cherwek: Well, Cyndhi, you know, one of the things I've admired and learned a lot about you, is how you'd use simulation in aviation training over the several years you've been with FedEx, but also throughout your entire career. Could you tell us what is simulation and how you as a pilot and as a trainer in aviation use simulation?

Cyndhi Berwyn: So there's a lot of different levels of simulation and one is a very basic level, what we call a chair flying, and that's where you simulate in your own mind going through an event and you go through step by step: what would I do, what would they say to me now, you know, what would I say back, and it involves nothing artificial; it’s just you and your mind, or you and somebody asking you questions. And that simulates a flight by chair flying, by preparing and going through it in your mind, so that's the most basic level.

Everybody knows now about video games, that's a level of simulation. There's increasing levels of video games: more interaction, more or less, so you can work with just the keypad, you can work with just the mouse and move things around and simulate by putting your little arrow somewhere and say, okay, this is where I would want to go, but then, you become increasingly complex by putting together—what we have at FedEx and what they use at all the airlines and flight safety training departments and the military and stuff—and those are full flight simulators, so they've got ones that have no motion, but they have all of the buttons, dials, and controls and they've got ones that come up on motion and, when you're flying one of those with the visuals and with all of the components, you don't know you're not in an airplane. It is so real –the feelings, the G forces, the turbulence, the vision—everything feels real.

So, the beauty of that is that you can practice everything—you can simulate loading somebody up when they get really busy and something goes wrong and how they would handle it— and there's an advantage to practicing that. I heard one of the NASA astronauts say, one time that they practice, they simulate training so much that train till they can't fail, they do it so much they can't fail. And, you know, we do that with walking as a child or with talking, we train so much till we can't fail, til we can walk across a room or, I mean that's very basic, but you can use that for any skill set, you just practice it and how to close a door quietly. I used to make my kids close the door 10 times quietly so they’d quit slamming it, because then you build up some motor memory there and there's just ways of using simulation or practice to build your muscle memory and your thought process and do it right every time.

Obviously, we can't afford, in an airline training environment or even military training environment, to train till you can't fail, so we train till you can succeed. So there's a balance between the time and the cost and what you get out of it and you have to weigh all of those things as you develop your training programs to what an average learner would pick up. How much is going to be enough practice until they know what they're doing, and then, of course in life, they'll continue to practice as they operate aircraft throughout the world, and they'll develop the skill set even more.

Dr. Hunter Cherwek: Yeah, and I think one of the things I've always appreciated is—and this is when I first stepped on Orbis One back in 2005—I saw so much you have wired and hardwired into your team, into your thinking, where when you encounter a challenge you literally stop, first control yourself, and then walk through an algorithm, and so that it's like a checklist. And, I think, we're now beginning in medicine to learn a lot from you and aviation, the use of checklist: are the wings? check. wheels? check. Now we're doing that with: this is Mrs. Jones? Check. We're doing: the right eye? Check. Cataract surgery? Check. But also that team approach, where certainly there's control in the cockpit, but there is everyone's checking, everyone supporting each other; we're doing that same thing now, taking that cockpit and putting it in the operating room. As you've said, Orbis is now really leading the way. We look at things like Xbox, how they can make these video games so life like, we want to look at how you can take the Xbox into the OR [operating room]. So, yeah, I'm really excited before where the future of training is going, and I never thought when I was training as a resident, almost 20 years ago, we'd be sitting here talking about simulators and artificial intelligence. It's just such an exciting time to be in training.

One of the things, obviously we've all been impacted by COVID and this new normal that we're living in, do you see with social distancing where, maybe the number of flights have gone down for certain airlines or the number of procedures that doctors are doing have gone down as hospitals were slowed down because of the lockdown, do you see simulation as a way of maintaining skills? Or as a way of allowing people to practice in a safe and socially distanced manner?

Cyndhi Berwyn: Sure! So, the FAA for pilots in the United States has a requirement for landing currency: you have to do three takeoffs and landings every 90 days. Well, the Orbis airplane hasn't been flying, and so to maintain currency in the airplane we use the simulator to go in and practice three takeoffs and landings. Well, we have the simulator for a four-hour block and you're not going to be a very proficient pilot if all you practice is takeoff and landing. So we practice crosswind takeoffs and landings, we practice engine failures, we practice emergency procedures of many different types, looking for errors, looking to make sure we're still very competent with our systems knowledge and don't become complacent. I think the airlines that are suffering from a lack of flying right now are also very concerned about what it's going to do to their pilots and their proficiency, so they will come back and, once they start flying again, they have to determine how much training is going to be required, how much simulation.

I don't know how you handle that in the medical business, you know, if you're a surgeon and you don't get to operate for a year on some specialty, how do you bring that back? Is there simulation available, because it would be very beneficial to practice with a knife on a simulator before you hit a person, you actually have to have the fine motor skills, right?

Dr. Hunter Cherwek: No, you're exactly right. With that tactile sensation— the haptic feeling of what does the tissue feel like, what does this, what pressure do I need to apply there. That's something Orbis is doing. We're also working with a company to build simulators, so that we don't have a lost generation of residents who have not gotten the right numbers, who have not gotten the critical mass of cases, that, you know, as we've talked about with simulation, we can sit there and encounter every complication in a weekend on a simulator, where it could take three years of a residency to see all the different complications, and it's in such a controlled environment, people feel calm, they can discuss it rationally and build up that assessment and judgment and execution skills that you have spent decades trying to build in pilots and aviation team members.

So, obviously, it's very rewarding from a technical standpoint, flying this very special plane and going to airports that probably you would not normally land in as a FedEx pilot. What are some of your greatest memories, the memories that really have stuck with you over the years of volunteering with Orbis? What are the special moments for you at Orbis?

Cyndhi Berwyn: So, flying the airplane is great— delivering the vehicle, picking it up, you know, moving it— I love going into new places, I love being able to bring the airplane in and that challenge that it brings, so operating in to unusual places is a joy. But working with the people! I love working with the people and helping people feel calm and cared about, you know. I've seen patients walk up and they're lost, they don't know what they need to do, they don't know if they're going to be seen. I can't speak their language, they can't speak my language, but I can reach out and hold them by the arm and lead them to the next room that they need to go to. I can sit with them. I can help them feel calm, that they're being respected and going to be seen and going to be taken care of, and I love that. I also—the joy, the results of one of the operations, what you guys do for people is incredible.

You also know, I have that granddaughter who was born blind and she does have some vision but not much; but I know the joy that comes from a child seeing their parent for the first time. Ah! Or a grandparent being able to see their child or grandchild for the first time. Those are incredible. Or for somebody to be able to work after they've had that life-saving surgery. It doesn't, maybe you don't think of it as life-saving because it's just an eye, but it allows them to work and support their family and not be a drain on their community.

You know, I also like working with the donors. So, I've stayed in Ethiopia, for example, and worked with some donors and given them tours of the airplane and showed them what we were doing and what we care about and for me that's really fun, to meet the people that are blessing us with their assistance. That's an important, important factor, because they're making a difference in the world, they're allowing us to operate. I think somebody who— we all donate what we can and I don't have a lot of money, but I can donate my time, I mean, I can donate my skills to flying the airplane, if I had a lot of money, I'd love to donate that and I do donate cash, of course, to Orbis – but people who are able to give generously and make such a difference in the world are really neat people, and to get to know them and to show them around, it's an honor for me to work with them as well. So I've loved interacting with some of the donors that come out and see what we're doing out on the road.

Dr. Hunter Cherwek: Yeah, and I think not many people know this, you know, speaking of donors, it was actually the employees of FedEx donated the current Orbis plane.

Well, I definitely want to end today’s episode with the exact same way I started by saying thank you. Thank you for all that you do for Orbis, how you inspire so many people, but especially women and young girls to pursue their dreams, whether it's in aviation or medicine or going around the world, just thank you!

Cyndhi Berwyn: My pleasure, you know, that's one of the reasons that I like to wear my uniform as well— I wear my uniform when I'm out in speaking engagements, I wear my uniform when I'm helping there with Orbis on the screening days. I hope that inspires them, I hope that inspires some of the youngsters that we see coming on the airplane for tours or having a family member that's coming along for surgery to maybe pursue a career field in medicine or in aviation or whatever it is that interests them.

The things that we get to see from the cockpit are amazing. Looking down on the earth and seeing the world—the things that you get to see, I think, in the operating room, are equally amazing as you see people develop and blossom into a better society, a one world they care about each other. I think that's an important thing that Orbis brings is a world that cares about each other.

Dr. Hunter Cherwek: I don't know if I could have said it any better.


Dr. Hunter Cherwek: That was Captain Cyndhi Berwyn, a leader in the field of simulation. I’m so grateful that she’s a part of the Orbis family, and grateful that she was able to join us on Sightlines.

Finally, I sat down (virtually) with Dr. Danny Haddad, Orbis's Chief of Program, to talk about how simulation training is playing a key role in Orbis’s training programs around the world. Here’s that conversation.


Dr. Hunter Cherwek: Well, Danny, it's really exciting to see where we are today at Orbis and, you know, one of the things you've been heavily invested in is simulation. Can you just talk about your experience with simulation and how you want to bring even more of this technology to help our mission?

Dr. Danny Haddad: I think when I did my first surgeries, I remember how I was just—petrified, and one of the things that really impressed me was work that colleague of ours, Dr. Emily Gower from University of North Carolina, did with developing a simulator for trichiasis surgery and the amazing thing is that I actually have it right here on my shelf.

Dr. Hunter Cherwek: Right next to Seymour, the teddy bear.

Dr. Danny Haddad: Exactly! So, the amazing thing is that what they developed was this head with eyelid units in there that you could practice your surgery on. And the nice thing is that after you've done the surgery or you could take the eyelid out, you can actually see what you had done. One of the most amazing comments that we've received was that surgeon that had done this for many years, when they saw what they have actually done that: Oh, wow, I had no idea, that I was doing that; and this is also in a time where the quality of surgery led to large numbers of people getting trichiasis again, this tool really improved the quality of the training and the quality of the surgery of the surgeons.

When it came to Orbis and the work that we were doing, we really saw an amazing opportunity for us to look at establishing what we call wetlabs, and doing simulation surgery in a wetlab. So, a wetlab is basically a room where we have the operating microscope, where we have the surgical instruments and the rest of them can practice the surgery in a safe environment. And one of the crucial things that I always find is that people learn the most from their mistakes, so being able to make a mistake in a safe environment that's, of course, how you learn to avoid that when you're doing the surgery on your patient.

What we've been able to do at Orbis is to create partnerships with groups, like the Phillips Eye Studio, around using artificial eyes within these wetlabs to train people how to do surgery and develop that skills level that you need to have to be able to provide high-quality surgery to the patient, because that is, of course, our ultimate goal. We want to make sure that everybody that receive surgery, the surgery is done in the best quality possible and, therefore, the outcome is the best possible outcome.

Now, we've been able to do a huge amount of work around simulation surgery, but one of the things that I'm personally most excited about is an opportunity that we got through funding from a really visionary donor—Dr. David Chang and the ASCRS Foundation [American Society of Cataract & Refractive Surgery] through our Sobrato Silicon Valley Innovation Fund—that funding allowed us to develop a virtual reality simulator, so the nice thing is that this is focusing on gaming hardware in 3D virtual reality. We've been working with a group in England to develop the software so that you can actually do the surgery in virtual reality and get feedback of where you go wrong. So, as a resident, you would try cataract surgery and it will tell you when you make the incision— when you're stabbing too deep, or even when you’re holding an instrument at the wrong angle— this provides an opportunity to provide a training institution with a very sophisticated tool that is very affordable, it's less than $15,000 for one unit, that could benefit all the residents for years to come in an eye center, it is portable so you can even share it between training institutions.

Dr. Hunter Cherwek: I fully agree, as someone who went through a residency without real simulation programs, and really, we didn't have these unbelievable technologies where you put on the oculus headset, and it looks like you're in the OR. You sometimes forget that you're in a video game and you can do and practice one step, like you said, a 100 times in a day or two, where to do 100 cases could take six or 12 months.

Dr. Danny Haddad: At Orbis, we've always been on the forefront of innovation and we have that strong link with aviation as well and, I think here, we've been able to also learn a lot from our partner FedEx. Mid-last year I had the incredible pleasure of visiting the FedEx training facility and seeing how FedEx is investing and structuring the training of pilots and how they are so many similarities—when you're looking at that combination of theoretical knowledge and the actual skills that you need to have when you're flying a plane and how to translate to ophthalmology. We've been able to learn a lot around how, a group like FedEx, is training their pilots, using different levels of simulation and we are trying to duplicate that model with virtual reality simulation. Afterwards, you can go to a much more sophisticated simulation with artificial eyes that are so realistic now that it almost feels like you're doing surgery on a real patient, so, I think, that there are such an enormous way for us to embrace what is already being done by so many different industries.

And that's also where our great partnerships with the academic institutions come into play, we have an amazing network of Volunteer Faculty from state of the art universities, where we are in a position that we can learn what is being done, what is happening in those institutions, what is the latest in artificial intelligence, what is the latest in simulation. And how can we translate that in the world that we do, as Orbis, and making sure that we keep it on the forefront with providing our trainees, our partners with all that technology, all that new innovation from the start.

Dr. Hunter Cherwek: Again, I really am excited to see how we're democratizing these technologies: we're using the best evidence, the best training tools to give best outcomes, both for patient care, but also for education and training, and again, it's just an exciting time to be at Orbis.


Dr. Hunter Cherwek: That was Dr. Danny Haddad. I’d like to thank him, and all of our other guests today: Captain Cyndhi Berwyn, Dr. Nicolas Jaccard, and Dr. Michael Abramoff.

Thank you for joining me for Sightlines today. I hope you’ll join us for the next episode and for the entire series.

If you would like to learn more about Orbis and the Flying Eye Hospital, please visit us at If you’ve enjoyed this show, please subscribe to our YouTube channel to watch each episode and check out many other videos about our work around the globe.

If you’re listening to the podcast version of the show, please hit subscribe, so you don't miss a future episode; and if you're listening in Apple Podcast, please consider rating or reviewing the show. It really does help others to learn about us, about Orbis, and our sight-saving mission.

Until next time!


Close the modal
Sorry there was an error.
Try again