The Tools to Reduce Bias in Medicine - Dr. Art Papier, CEO, VisualDX

SHIV GAGLANI: Hi, I'm Shiv Gaglani, and today on Raise the Line, I'm really happy to be joined by Dr. Art Papier, who's the CEO and co-founder of VisualDX, and an associate professor of dermatology and medical informatics at the University of Rochester. VisualDX provides clinicians with software tools accessible via computer smartphone and electronic medical records to help improve diagnostic and treatment decisions. The company boasts the world's largest, most equitable medical image collection, and tens of thousands of providers, and more than 2,300 hospitals, clinics and medical schools around the globe are users. I'm also grateful because we share investors in Coveras and have gotten to know the VisualDX team quite well, including developing certain videos with the team and working with many of the same institutions. So, Art, thanks so much for taking the time to be with us today.

DR. ART PAPIER: Shiv, it's my pleasure. I actually was not pre-med in college, I was a studio art major, not heading towards medicine and a series of circumstances took me towards medicine, and I ended up working in the 1980s as a medical student with the physician that invented the idea of the problem-oriented medical record and the SOAP note format, Larry Weed. That really got me interested in the field of medical informatics. Dr. Weed was saying as early as the '60s and '70s that it's impossible to memorize it all and that we needed to bring into medicine tools to really aid and augment decisions. I heard that message in the '80s really when computers had green screens and up and down arrows and no graphics and ended up in Rochester right when Kodak, which a lot of the young people listening into this podcast probably don't know what Kodak is, they made this stuff called film and we had these things called slides and you put them in something called a carousel, and you did a slideshow to your medical students.

Right when Kodak was announcing scanning film to digital, I turned to people at Kodak and to our chair of dermatology. I said, "This is going to be huge. We're going to get rid of our Kodachromes. We're going to have imagery on computers, and we're going to think about Dr. Weed's ideas, and we're going to bring it all together in a product." Dr. Lowell Goldsmith, being a Renaissance physician, one of these rare individuals that's a powerful researcher, educator and patient advocate said, "Sure, let's give it a try." That's how it all got started.

SHIV GAGLANI: That's awesome. The role of different mentors or advisors, or even people who just come and give a talk in careers is really strong. Dr. Weed, I'm definitely familiar with him. One of our investors is Peter Frishauf who founded Medscape and he's extolled Dr. Weed for some years as well. Can you tell us a bit about the early founding of VisualDX and the size and scale? I already mentioned you're using over 2,300 health centers and hospitals and clinics. Can you tell us about the growth trajectory and where you see VisualDX in a couple of years from now?

DR. ART PAPIER: Yes. We're really different than most healthcare IT companies and we've really been three different companies. We started in 1999 and prototypes before that. We actually started working on these ideas before the Internet existed, which is kind of hard to get your head around. Our vision always was merging a knowledge base with an image base to create a point-of-care tool. This is a real paradigm shift because we've really been framed by the model of textbooks, which is loading information to the human brain and then expecting a human, a physician or nurse practitioner or a PA to see a different patient every 20 minutes and do it perfectly every 20 minutes. The vision has always been, "Let's put this information in a tool that will work in the exam room." 

We started before doctors were using PalmPilots, and computers were slow, and really the company they should have disappeared, we really should have, to be honest with you, gone belly up because we were way ahead of our time. But we did work. Early on, the very first product was for the patient with fever and a rash. We were saying, "Where's dermatology really important?" We said, "In the emergency room." There are skin signs of serious infectious disease. So we developed this tool for the patient with fever and a rash and delivered it to the marketplace in March of 2001, and doctors basically said, "Well, that's cute, but I don't even have a computer in my exam room." There were no smartphones, and so what happened was 9/11, and then the anthrax and the mail. It's a long story. We can do a whole podcast on public health informatics, but we basically ended up working for the CDC, state, city and county health departments, delivering the fever and a rash component of VisualDX emergency departments. That has relevance today. 

There is a benefit sometimes of being old, and we're not a bunch of kids that said, "Okay, we're going to create the next unicorn. We're going to “disrupt healthcare," whatever that means. I don't know what that means. "We're going to really build something that's meaningful and have impact." It's taken 20 years to get to the point where we actually have technology that people widely use. We have technology on the phone and most doctors, students and residents are actually much more comfortable with this thing than they are their EHR, because they're easier to use. We have these wonderful apps for professionals, Android and iOS, where they can enter patient's symptom signs and get a powerful differential, and it not just be a list of words. It's visualization. 

Tying this back to my background in art, what I think about is, how do we make information? As does Osmosis, because Osmosis is thinking about this, too. That's why this partnership is so wonderful, having two organizations that believe that it's not good enough just to have everybody read voluminous words and think that the brain could put it all together. So, there's really a great conversation we could have, Shiv, about visualization of information.

SHIV GAGLANI: Absolutely. I love what you were saying about how you guys should have gone belly up, but you managed to push through and caught different waves. We had another guest, Burck Smith, who started StraighterLine, and he has a great quote that he shared with us. It was, "If you stick around long enough, your timing is perfect."

DR. ART PAPIER: Sign me up in that club. I can really appreciate that.

SHIV GAGLANI: Obviously when you were innovating and getting things into the EHR as point of care and now with the mobile apps, you've developed this massive library and I'd love to give our users a sense of how big this library is. What does it mean to have the most equitable medical image collection? Because I know other events of 2020 led to a lot of discussion around racial justice in medicine ranging from vaccine distribution to underrepresented areas in terms of healthcare. We'd love to get your thoughts on that and then how that's empowering things like AISTA and AI Medicine, because you guys are clearly leaders in that space, too.

DR. ART PAPIER: Yes. Working backwards from AI and machine learning, of course, that's the new buzz word and everybody's doing it and there's wonderful potential, but it's only as good as your data. You have to have wonderful, precise, accurate, excellent data to train your machine learning and AI, and the medical records of course are full of mythology. There's just a lot of data in patients' medical records that's not up-to-date or not accurate. So what we've been doing since the inception of the company is ingesting images that are in case formats from people we trust. So we're not a Wiki model. We've been working with experts around the world, mostly in the US, but we're very much a global company. We love to have people contributing from other countries. We get these cases with pictures, mostly scanned, but then it grew to eye, and oral, and then it grew to radiology and EKGs.

So we've just been amassing for over 20 years what we argue is the world's best aggregated, visible light medical image collection. I mean, we do have radiology. Our focus is not on radiology, AI, machine learning. We're leaving that to many other groups that are really focused on that, but we're really on visible light imagery. So we reach out to people that we trust and know. Our experts are great people. We get their imagery and we put tags on the images. We have a team, an image department that's just focused on this for 20 years. Along came machine learning three or four years ago when the algorithms really started to be good, and we started training those images on or using these algorithms. We've made a lot of progress, and one of the things that we argue is that it's not just the images, it's also the knowledge. And how do you bring a knowledge base together with an image space? 

During this period, we've been very, very interested in why errors occur in medicine. What are the causes of diagnostic errors? We have a lot of cognitive biases. Many of the listeners are familiar with terms like premature closure or anchoring bias and all the biases, but then there's also gender bias. You assume a woman is not having an MI, you think it's only guys having an MI, so you miss that MI in a woman. Then you have racial bias. This has been an area of interest also of mine for a long time. I wrote a paper in 2006, published in the Journal of the American Academy of Dermatology, entitled “Disparities and Dermatologic Educational Resources.” We showed that dermatology resources, our books and lectures, really were not showing how disease looked in people of color.

The company has been purposely reaching out to people that have skin of color collections for years. Our internal library in general is in the hundreds of thousands of images of disease, and 30% of it is skin of color, which far exceeds any other resource. When the George Floyd tragedy happened, people started reaching out to us, including a New York Times reporter who had read my 2006 paper and said, "Oh, there's no images of COVID toes in skin of color. This is a huge problem." I said to her, "It's not COVID toes that's a huge problem. It's all of medicine that's a huge problem." I said, "Our med students and residents just are not being trained in the spectrum of disease."

We've been passionate for 20 years. Many of our users didn't realize there was a filter in VisualDX where you could say, "I don't want to look at the pictures of light skin. I want to look at the dark skin." We realized we had an interface problem. In the fall, we very rapidly brought that filter to the homepage of VisualDX analysis skin of color and people are saying, "Wow, so glad you added that." And we said, "No, it's been in there the whole time. Those 14,000 images have been in there, but you just didn't look deep enough." What percentage of Microsoft Word or PowerPoint to people use? About 10%. It's the same with any software. We've been working hard to make sure that our users know, and now they know. 

There are all these medical schools and users out there that are hungry to reduce bias in medicine. We have to get the word out there. We're getting these emails like, "We want to create a skin of color atlas." Our team here that has been working on this for more than a decade is going, “Why do you want to create a skin of color atlas? First of all, VisualDX is 14,000 images of skin of color. You don't need just a skin of color atlas. You need an integrated system for all diagnosis, that light skin and dark skin.” One of my African-American dermatology colleagues said to me, "Why would I want just a skin of color atlas? I'm African-American but I see plenty of white patients. I need something that handles all of my patients." We, as an organization, VisualDX, the one thing that we need to do better is to get that message out there because so many people don't know what we do.

SHIV GAGLANI: Absolutely. My hope is the podcast and our audience can be part of that as well, as well as all the collaborations we've done. I don't know if you've actually spoken to or know, but there's a medical student at St. George University of London, Malone, who wrote that book, Mind the Gap.

DR. ART PAPIER: Yes. It's interesting; I got a lot of press, but there are African-American dermatologists that wrote skin of color textbooks. There are many physicians outside of dermatology and inside of dermatology that don't know about Susan Taylor and Paul Kelly's second edition. Paul is deceased, but Susan, who's a professor at U. Penn, has been a champion of this. She helped to found the Skin of Color Society. So, there are wonderful books. There's wonderful information. I think it's difficult for people that have worked hard in this field, many of these professors that have written these skin of color textbooks, to see the press think that the wheel has just been invented. I mean, people who work in this field for 20 years, 30 years or longer creating these books.

SHIV GAGLANI: Yes. That just shows the gaps sometimes that occur between academia, things that are published and validated, and then people being willing to accept them, or interested in them, at least. George Floyd's case obviously helped bring that to the forefront.

VisualDX, obviously, you guys have a lot of traction. You were saying after September 11th with anthrax and looking at rashes and the emergency room, et cetera. COVID is why we launched this Raise the Line podcast. Can you tell us a bit about how VisualDX has adjusted to COVID? Then you as a clinician, what are your thoughts on some of the lasting changes that will occur over the next couple of years because of the pandemic?

DR. ART PAPIER: For me, it's like déjà vu all over again, because 20 years ago we had a national emergency. Nowhere as bad as COVID, but people were really scared when anthrax was being spread in the mail because it followed 9/11, and there was collective anxiety. Now we have the same level, much more actually fear in our patients. One of the things that happens when the system is stressed, there's more error. So, people are being harmed because we're really almost in a wartime environment in our emergency rooms and ICUs today in most states. I mean, it's horrific what our colleagues that are on the front line are doing, having to deal with, and they are true heroes and there's a lot that's going to be written about this. But how do you make decisions in stressed environments?

As the acute phase of the pandemic comes down, I think we're going to see a flood of people deferring normal care. There are a lot of unfortunate stories now of people that have put off their symptoms and are getting diagnosed really late with things that could have been diagnosed earlier, because they've been afraid of the hospital. So as this acute phase winds down because of vaccination and herd immunity, and now we're back to medicine, I think diagnosis is being elevated. Obviously, digital tools and telemedicine has been elevated. There's a lot more telemedicine, tele-education, digital education, but we really have to live up to the promise of what these tools can be and we've fallen so short.

I mean, billions have been spent on the electronic health record, and it hasn't really improved care. Those of us working in this field have this collective responsibility to really deliver meaningful tools. Our take on it from a COVID point of view and a future prep point of view is, "Look, we're going to have big change starting on the 20th." Reason and science and data and rational thinking is going to return to this country. It has to.

When that returns, we're going to look at the budget for the Centers for Disease Control. We're going to see that we've been giving the Centers for Disease Control — to do not just COVID, but all bacterial and viral, fungal disease, motor vehicle safety, tobacco, vaping, everything they do – we've been giving them $11 billion. The revenue for Humira, arthritis and psoriasis medication annually is $20 billion. So, our society has committed half the revenue from one drug to the organization that's responsible for pandemics, epidemics and public health. That colossal mistake will end up costing us probably $5 trillion in harm to our economy because we didn't do preventive medicine as a society to invest in the Centers for Disease Control, city, state, and county health departments. We just haven't done it. We will wake up to that fact now, and just in a week or two people will say, "We can't starve public health. We have to invest in it." That's going to be the number one change. 

From a technology point of view, where we have a lot to add to this conversation is out of that experience working on anthrax, we realized that there are two separate silos. There's what happens in these county health departments that are starved. These heroes that work for very little money, these commissioners of public health, the contact tracers that are now volunteering, they have been operating on fumes. So they're doing their thing in one silo. And then you have these overworked doctors who are starved, who are pressured too much in emergency rooms and in offices. These are two separate silos, but shouldn't the busy doctor know what his county health department is alerting about? Shouldn't that busy doctor know how to easily report infectious disease to the county health department?

We started thinking about that 20 years ago with anthrax and we developed in VisualDX, a  parallel database system technology to customize the knowledge that has the local health department information in it. When we come out of this acute phase of this disaster that we're living through right now today to this phase where we're going to say, "Okay, how are we going to prevent the next pandemic, the next epidemic, and recognize it early and be rational?” Digital health is part of that. We have a lot to bring to the table in that area.

SHIV GAGLANI: Absolutely. My hope is that we don't forget a lot of things that we've learned over the past few years. I hope that we don't have too short-term memory and move on to the next thing.

DR. ART PAPIER: That's a great point, because bioterrorism, that memory span was five years. We invested in that for five years and then it tailed off. So sustainability of what you do — I know Osmosis is thinking about the educational paradigm. How do we shift the educational paradigm? We're both involved in this shift, really a paradigm shift and a moving away from just reading books or reading online, to figuring out, how do we get students to be engaged in residence and doctors over time, to stay current? How does education dovetail into tools that you'll use? You wouldn't ask your pilot to memorize the route from Chicago to LA and memorize the route to Miami. They have instruments and then you have a cockpit simulator and you train them in the cockpit, instruments with simulators. So Osmosis and VisualDX, our charge is to think about what's the cockpit, what's the simulation, what's the education and how do we change the paradigm?

SHIV GAGLANI: Absolutely. Great analogy. I know we're coming up on time, so I had two more questions for you. The first is, given that you're a physician, entrepreneur and still in academia, what advice would you give to students considering a career in healthcare right now about meeting the challenges of COVID and beyond?

DR. ART PAPIER: I still enjoy clinical medicine. Unfortunately, I'm only able to see patients one clinic a week, and it's from seeing patients that I understand what the problems are. It just can't be read about. You have to touch it. So, students that are in healthcare have this incredible opportunity to get involved in transforming healthcare. What I don't recommend is that you go to med school and then you maybe do residency or you quit residency, and then you say, "I'm going to start a unicorn based on the fact I went to med school." I recommend that you stay clinically engaged so that you understand medicine. I think that understanding medicine is not just a book experience; it's a real world experience. Most learning in medicine happens in residency. It doesn't happen in med school. So do your residency and stay in it.

SHIV GAGLANI: Absolutely. We had a guest, I'm sure you know Dr. Sachin Jain who ran CareMore and now runs SCAN. He's also a vocal advocate for people who want to innovate and improve healthcare to maintain their clinical practice as well as he does. My last question is, is there anything else you'd like to share with our audience about you, VisualDX, the future of healthcare, that we didn't get to?

DR. ART PAPIER: Wow. Do you have another two hours? We're at an inflection point at VisualDX where the world has changed. We're now going to be helpfully living in this post-COVID world. We're right in the middle of COVID, and we have tools that exist today that can be helpful to anybody that's excited about digital health and anybody that wants to have real world impact around racism. 

One of the things that one of my team members said is exasperating, is the fact that there is a tool here that will bring equity to knowledge around skin of color. We've been saying to healthcare systems and to the payers, "Look, it's great that in a white coat, your healthcare system got down on your knee and did solidarity for George Floyd and Black Lives Matters. Nobody's disapproving of that. But the real question is, what are you doing that is meaningful? What are you doing that's meaningful for your hiring? What are you doing that's meaningful for pay equity? What are you doing is meaningful in terms of how you present cases in your medical school?” We've been saying to them, "There is bias in the educational resources, and we have a solution that exists today, so get beyond doing the talk and let's do the walk together." 

We'd love a reach out from anyone that's listening that wants to learn about our solution that can help with reducing knowledge biases and bringing more equity to medical knowledge. We're staying focused on that right now, because it's so important.

SHIV GAGLANI: Again, thank you for that work. We will, in the show notes, make sure people know how to get in touch with your team at VisualDX. You have a wonderful team; I had a chance to interact with a number of them like Damon and Rory, and of course yourself. So with that, our thanks so much for taking the time to be with us here on Raise the Line, and more importantly, for the work that you've been doing over the past decades to make medicine more equitable and reduce diagnostic errors.

DR. ART PAPIER: Well, Shiv, it's been really a pleasure and we look forward to future innovation and in collaboration with Osmosis. Thanks for having the conversation.

SHIV GAGLANI: Thank you. With that, I'm Shiv Gaglani. Thank you to our audience for checking out today's show, and remember to do your part to flatten the curve and raise line since we're all in this together. Take care.