Dan:
Yan, it's great to have you here.

Yan:
Great to be here and thank you for inviting me.

Dan:
Yeah, it's great to chat. I know we've had a history of having really interesting conversations in the past, so I'm hoping we have another one of that in your new role, or newish role. So, can you tell the listeners a little bit about your background and how you ended up at Automation Anywhere?

Yan:
Sure. I am a pediatrician by training, so I've been doing pediatrics for a long, long time at Kaiser Permanente. I had the chance to become eventually the director for innovation and advanced technology at Kaiser for the last eight years that I was there, working with thousands of startups and looking at how do you innovate in a large healthcare enterprise? And that was very, very interesting. In 2014 I left, I became chief innovation officer at a Washington D.C. healthcare consulting firm working with the VA and DOD. Then I came back to California and joined Amgen for about two and a half years being the lead for digital medicine and for things like wearable devices, and virtual trials and that kind of stuff. So learned quite a bit about digital technology in life sciences.

Yan:
Then seven months ago, I got a call from a company in San Jose called Automation Anywhere, one of the enterprise leaders in robotic process automation or RPA, now morphing into intelligent automation using AI. And I thought that was a very, very interesting move. What they were looking for was a leader to help launch their healthcare industry outreach, the healthcare vertical, along with several other industry verticals we also got people to help lead. And so, it's been a roller coaster of a ride, because actually Automation Anywhere is one of these very, very fast growing companies that looks like it's tapping a huge need in healthcare, which is to automate and to eliminate the kinds of things that machines do best and humans don't do best. So that's a bit of my history.

Dan:
That's great. Yeah. It's so interesting. I mean, you've tapped into so many different aspects of the healthcare world and now the jump into automation. And you described it a little bit in the end of your description there. What's the simple definition of automation? I think there's a lot of things that people think of, like robots doing your job or taking your job, or computers removing entire electronic medical record keepers, all that kind of stuff. So what's a simple definition of automation?

Yan:
You're right. I think automation can mean different things to different people. But one simple definition is that automation is the use of computer software and hardware to perform the manual repetitive and rules-based processes that human beings find tedious, mind numbing and error prone. And these are things that most people would rather not be doing. In fact, there was a study by PWC that found that 63% of healthcare workers today feel that part of their daily work could be automated, which frees them up to do more productive and valuable things. I think it's interesting that we're in the sort of the fourth industrial revolution as we deal with automation, as we deal with artificial intelligence, and intelligent bots and so on, how do human beings actually create a healthcare system that's sustainable and fulfilling for both its patients and its workers for the future? And that's a huge question.

Dan:
Yeah. I saw a stat from Amazon recently where they showed the number of robots versus the number of humans that they have employed, and their spin on it was, we are learning how to work side by side with robots. And yeah, I think it's a question that hasn't been answered well yet.

Yan:
I think for healthcare specifically, as you get into more healthcare specific applications like nursing, and physician treatments and things like that, that those kinds of specialties still rely a lot on human expertise. So now we're starting to see automation come in as attended automation, which we call attended because it's like having a digital assistant at your side with a portfolio of capabilities that you can tap. And this is sort of, I think this is the near term morphing that we'll see in healthcare. People would be enabled by these bots, software robots or bots, to do more things than they ever could do, and hopefully to release some of the tedium that people are complaining about and people have burned out by.

Dan:
Yeah. Yeah. You mentioned the mind numbing. I mean, my wife works in claims at Kaiser, and I just hear some of those stories of like massive spreadsheets and manual processes that there's hundreds of millions of dollars at stake that [crosstalk 00:07:01] so much better. I wonder from your perspective, what are some of the most innovative applications of automation within the healthcare space?

Yan:
There are a couple that I can think of. One example, and one very common example, is the contact center and people starting to automate this. Traditionally patients that call into a contact center, and I've done that of course, you talk to the agent who doesn't seem to know who you are, they take a long time to verify your information and your demographics, and then you have to wait after that conversation maybe another couple of weeks to get a resolution of the original problem that you wanted to solve. With attendant automation, what organizations are finding is that where the agent has at their side an intelligent digital assistants so that the moment you call in, actually a call ID can trigger the bot to start fetching your data. And so, the agent, instead of you waiting there while you hear them tapping on the keys, we can actually have on the screen within seconds all your information, can talk to you like they know you, which is good, including your medical information, and your immunizations, your reminders, things that you need to get done, your appointments, all these kinds of things.

Yan:
They can actually talk intelligibly with you and actually solve problems. And the problem solving itself can be based on rules, which means to block it and execute it, and the bot can do the reminders and the alerts, and finally enter the data into any kind of a tracking system, including EHR that you might want to enter it into. So, the bot can take away a lot of that pause, that pregnant pause as you're waiting.  And I think that's very exciting. The other area that people are starting to look at, is that because automation runs at a very high level above other conventional software, you can actually get the data from multiple incompatible systems. So the issues interoperability, which has been a big issue in healthcare forever. So, can you imagine a bot can get information from many, many different systems in near real time and aggregate it into an ongoing near real time dashboard.

Yan:
So for instance, let's say a case of when is a patient ready to get discharged from the hospital? Normally you've checked the EHR for the final discharge order, maybe you check all the nursing, the discharge checklist, you have to train the patient on a diabetic treatment, all the kinds of things that have to be done. But even though that's quite good, it's not good enough. For instance, if you could scan that information plus check on things like the ability of transportation, checking on the visiting nurse schedule, checking on the availability of labs and appointments in other facilities and so on and so forth, then you could have a much more useful dashboard and a way to generate alerts than you would normally have. And this is a question actually that came up, because when I was talking with one organization where they were saying, oh we have Epic, as Kaiser is Epic, whatever system you have, it already has alerts in it that you can create.

Yan:
And that's true in many of these systems. But if you want to go outside Epic, you're kind of done for. You can't really do that. So having a tool that allows you to check many different things, including things you wouldn't normally check, such as social media, such as information from cities and counties, current alerts, pollen levels, things like that, that's very powerful. People are just starting to see all the possibilities. I think that's very exciting.

Dan:
Yeah, that's really interesting. And I know we talked at length about what sits on top of all this massive data that's kind of disparate within healthcare systems? And sounds like this is something that potentially has the ability to pull it all together and actually give those clinically relevant or customer focused relevant insights for people to make better decisions on.

Yan:
And in the future, you think about this interoperability and sort of this super scanning ability. One interesting possibility as we get into the clinical realm is the ICU. One of the issues in ICU with all the monitoring, and all the charting and everything that goes on, there's just a ton of data that the attending clinician has to go through when they come in in the morning. What if you could develop a risk score based on certain things, certain items that happen through the night, so that when the physician comes in in the morning, they would have a prioritized list of the sickest patients to the least sickest? And so they would spend more time on people that needed more time. So that kind of application is probably going to take a bit longer because that's going to require clinical input to figure out what the relevant components of that risk score could be. But that would be a great time saver. It would allow clinicians to be more productive, to spend time on things that needed to be spent time on, and essentially it's an enabler. It's a productivity enabler.

Dan:
Yeah. Right. I think there's some marketing around it too, but it's making super humans. I mean, they become enabled with the right information at the right time, which is the dream of I think every frontline clinician in the world because they don't have to dig through everything, hours I think. Our mutual friend Marilyn Chow did that study where it's 36% of time of nurses is wasted in finding stuff and information. There's huge opportunities to just free people up and actually be more vigilant on the care delivery side rather than just searching for information and equipment.

Yan:
And in terms of the clinical aspect, there's actually another very interesting application of automation, which is that the software bots themselves, the software robots, also have the integrated or inherent capability to monitor in very granular detail. So wherever you stick a bot they will be able to tell who did what, for how long, who was it, whose permission did they use, what were the outcomes, and so on and so forth of what was demand on that process. So you think about what areas in healthcare require monitoring. We have lots of areas that require monitoring. [crosstalk 00:13:09] Everything in healthcare requires monitoring. It's like quality, safety, regulatory, cyber security, all this kind of stuff. So all these areas could be actually enhanced using a third party monitor. They could give you a very granular audit trail for regulators and for safety coordinators and so on. But I think there's a lot of potential, but it's just up to us, the clinicians, to figure out where we need to put this technology, and of course to have people accept it, accepting that that's part of it.

Dan:
Right. Well, yeah. I mean, I remember doing a project, and yeah, the acceptance piece of just understanding that this thing's not stealing your data, or manipulating it, or sending it to the wrong person, I think that's part of the change management processes is one of the biggest barriers probably.

Yan:
Right. And I think one of the things that people are starting to realize when they look at these enterprise automation packages is you need a very, very strong central control center. Obviously you don't want to have people building bots all over the company and running rogue processes, sending data out of the company, that kind of stuff. So obviously if you buy a package for automation, you need to make sure that they have this totally controlled governance, and security, and audit ability and everything else. And I think that's one of the things is sort of the table stakes for healthcare is cybersecurity. We hear about a breach almost every day. It's crazy.

Dan:
Yeah. So that brings up a good point. What do leaders need to know or how do they even start approaching the idea of automation? Is there an industry standard they can go look at, or is it kind of one off innovators, early adopters at this point, or is it more widespread?

Yan:
I think right now it's let a thousand flowers bloom. It's like the early days of innovation at Kaiser, everybody's experimenting with it because they want to learn. And so it's an exploratory learning process, and part of that is that you may not know which direction you want to go, you just want to learn it now so you have some idea of where you want to go. But I would say one of the 800 pound gorillas, and there are a couple in the closet that needs to be acknowledged, is that there is a lot of pushback, a lot of fear about job displacement. And this is something that's happened with the last industrial revolution. And so for leaders, basically if you want to automate, you need to think about re-skilling or up-skilling. That's just part of the package. In fact, most organizations according to some of the analysts, they don't replace people. That's not how they're getting the ROI.

Yan:
Most organizations, I would say over 90, 95% are not laying off employees, because they're finding that using employees, humans, for what humans can do best actually gives your company a lot higher ROI. And for instance, you want to deploy human beings to show empathy and to look at care interactions with patients, you want them to solve complex problems, you want them to deal with exceptions, things like that, that machines can't do well. So the thing that leaders have to keep in mind, is even though this many times unspoken, it is a definite concern by their company and they need to address the re-skilling, up-skilling piece. There's also responsibility that leaders have to monitor and govern automated processes. So, as I mentioned, they need these products, they need to have enterprise graded security and governance, especially in healthcare, which is being governed by HIPAA, and by GDPR and several other places, several other organizations. The last thing you want is a process running unsupervised somewhere in your organization. So you need strong control in the center. But, right now what we're seeing is that healthcare organizations are just experimenting. They're doing proof of concepts, and pilots and things like that

Dan:
Squarely in the IT realm, or is there some, like the chief technology officer's office, or do you see engagement from both the clinical leadership and the IT leadership in some of the early adopter space?

Yan:
Well, I think the IT folks are generally always involved, because they're worried about the impact on the networks, and the storage and so on and so forth. So they are definitely always involved, but more from a technical point of view. The other side, the business side, what we're seeing is we're seeing a lot of interests from CFOs, from CHROs, CIOs or the business people, because they see the immediate return, and the returns on the order are 20 to 40% typically on productivity, on efficiency, reduction in delays, in processes, reductions in claims processing, things like that, medical billing, physician credentialing, prior authorization, all this kind of processes that are essentially electronic paperwork they were seeing a lot of return. We're starting to see interest from the clinicians, which is where I come in. And I love talking with them about all these other healthcare specific applications like monitoring, like inter-operability, that you don't think of normally for automation.

Yan:
But I think if you think about it as a toolbox as a way to enable different things, it's a good basis to start thinking about how to re-engineer your entire organization so that, and I asked this question of people, I say, when you think about the 21st century, we are in a 21st century, what do you think healthcare organizations should look like? I mean, for a lot of us, a lot of these organizations, they're still operating like they're in the 19th century. The services are provided very labor intensive, the delays. Things that's still very, very manual. And so, if you really want to compete and to prove yourself to be really competitive in the 21st century, there are a lot of technologies that you can tap, and automation is just one of them.

Dan:
Right. Yeah. No, I think there's, like you said, there's a ton of opportunity. One of the things we're doing here at Trusted Health is building software from the ground up using 21st century engineering techniques. And when we compare ourselves to some of the competitors, they basically built their software in the 90s, which was taking paper processes and digitizing them. And we're just coming at it from an actual intentional what software can do best and what people can do best, and actually splitting those roles out, which is I think healthcare has a long way to go in that space.

Yan:
Yeah. And I would totally agree. I mean, you look at the early software that went into healthcare like Epic and Cerner, all of them written in the 90s. And basically what it is is an electronic version of a paper chart, which if you would design software today, that's not how you design software today. But basically there are a lot better ways to do it where you use a technology to make the software contact sensitive to provide only what you need and to do what you normally do, and not to give you everything in the world like current software does. So, a lot of room.

Dan:
A lot of room, yeah. Well I'm interested, Automation Anywhere is not just a healthcare focused company, and so you must be learning a ton about other industries. What industries are doing automation right? I mean, top of mind comes manufacturing, maybe some legal or financial industries. But what are you seeing as industries leading the pack in that way?

Yan:
Yeah, that's a really interesting question because you're right. Like I mentioned, a lot of the early applications actually run across all industries. So the back office functions that include sort of HR, finance, electronic paperwork, processing, those are all obvious initial targets for automation, they cross all industries. But, I would say industries that have really taken it on wholeheartedly are the banking, financial and insurance industries, which are all kind of combined together in BFSI.

Dan:
A lot of paperwork there.

Yan:
A lot of paperwork, a lot of processing. If you look at Yelp, for instance, people rate these insurance companies to healthcare providers. One of the number one concerns is the delay that it takes to get anything done. And this is where rules-based processing can come in. You can do a lot of automation in this area. So banking, finance, insurance industries, including health insurance. Telekom is another industry, very transactional, very rural space, and many times not requiring human intervention. And of course, automation happens in the consumer goods area as well, retail. Anything that makes the consumer experience a better one.

Yan:
And I think that's something that those three industries ... Manufacturing is starting to pick it up. I mean, but you think about manufacturing, you think about hardware, you don't think about the robots. I'm basically talking about software automation.  Yeah. But yeah, I think those industries have natural. But I think when you start to look at the future of automation, I think industries is where it is. So people are starting to look at what are the industry specific use cases that we should be looking at, and we could really help healthcare to be much more sustainable.

Dan:
Right. Where do you see it on the adoption curve? Is this in the plateau of productivity yet, or is it in the hype cycle, or where would you put it based on Gartner's trough of disillusionment?

Yan:
Right. Right. No, I think that's also a very good question. I mean, the pace of adoption, it's really interesting. When you think about the Gartner hype cycle, I don't think RPA, robotic process automation, software automation, actually went through the hype.

Dan:
They just got productive real quick.

Yan:
Yeah. Yeah. Yeah. They just go ahead right to the plateau of productivity. It was just really something. I mean, it's an illustration of not everything follows the curve. Maybe it was a very sort of abbreviated curve.

Dan:
Fast, yeah.

Yan:
Yeah. Because they started 15, 20 years ago, and so they started as testing software and software to test code development. And so it's been a very gradual up brand. But as people started to use computers more, they can see the computers are not the best thing in the world for humans so they can start to automate them. And in fact, automation has been happening for a long time. So maybe that's why we don't have a lot of hype. In fact, that's why when I joined Automation Anywhere, I said, why is this such an explosive industry given their valuations, the spread, the adoption rate, and I've never heard of it? And I'm in innovation. So it's very, very interesting question. I've heard that from a number of other people. But when they really look at the industry, it's a stealth industry.  People see the return immediately. They don't hype it, the media doesn't cover it.

Dan:
Right, right. Yeah. Media likes to hype up the stuff that maybe sort of delivers value, not quite.

Yan:
Right. Right.

Dan:
This is boring. Oh you save $10 million now, this is boring. Oh my goodness. So kind of wrapping this up, you've held a lot of roles in a lot of different areas of healthcare, and now with automation. You mentioned the ICU and some clinical applications. Where do you think the, I don't know, most out there, blue sky application of RPA is?

Yan:
Well, I think RPA is going to change. The robotic process piece, which is just automating road processes, manual process, that's going to reach a saturation level, I believe, because there's only a certain number of processes that you can really automate. And when you get to that point and you're automating like crazy, that's about it. So you hit a plateau in productivity, the next step is going to be actually requiring redesign and re-engineering our healthcare so that you can automate the pieces that makes sense to be automated that people don't want to do and patients don't want to deal with, and you put more productivity in the areas that people value in terms of human interventions. So, I think that piece is going to take a lot of thought. I don't see very much discussion at all about automation and healthcare. Just the industry itself has been kind of a self industry. So not a lot of talk. Just like I don't see a lot of talk in many other technologies in healthcare because the clinicians are too busy. If you ask clinicians, what's on the future? There's two things. They say, oh, we should make the EHR more friendly. And number two, we should talk about telemedicine. Those are technologies that have been around for 20 years.

Dan:
Right. Yeah. Yeah. I often say healthcare is about 20 years behind the rest of the world.

Yan:
Exactly. Now, I was just at CES, which is this big convention in Las Vegas, to look at the latest and I can see a lot of potential. For instance, one of the biggest things was a company called Neon, which is funded by Samsung Ventures. Neon is a startup that's creating full body, full size human avatars that are responsive, emotionally responsive, that react appropriately, that can engage you in conversation. Those kind of avatars, if they ever get into production, probably in about 5 or 10 years, they get sophisticated enough, I mean there are some studies that show that consumers sometimes prefer to interact with those folks as opposed to real human beings. But the key for us as healthcare sort of thinkers is that it can allow healthcare to be more scalable. And that's something, given the shortage that everybody's predicting for clinicians, for nurses, for healthcare workers. In fact, I think nurses are going to be the greatest gap in terms of opportunities for jobs in the next 10 years. It's crazy.

Yan:
So we're going to be short of people. How do we create systems that can actually do well, perform well in terms of quality, safety, patient care experience using technology? And it's interesting, I did a HIMMS webinar a couple of months ago, and I asked the 160 people on the webinar, what's your top of mind priority for technology automation? What's a pain point you want to address? Number one was patient care experience, which was kind of surprising. It wasn't cost savings, it wasn't addressing quality, safety, whatever, it was actually the care experience. And I guess I could see that. Yeah. So things like the avatars, and the intelligent chat bots and voice response system, those kind of things are going to go a long way I think to help people really get a better experience from the healthcare system.

Dan:
Yeah. Do you think we need to start over? Is this possible to lay on top of the 1990s technology that's running most of healthcare now? Or do you think the advantage goes to someone like an Amazon or a Walmart who are kind of starting that new?

Yan:
I think they do have an advantage. They have a green field because they're starting from no legacy baggage. But the problem is that eventually they'll hit the legacy ... Right. Right. Right. So they'll go as far as they can from the retail medicine point of view. They'll go as far as it can from what they do well, which is a deliver sort of the retail experience. But I think that if you look at how long the FDA took to figure out whether a mobile app was a medical device, we're going to take a long time before we can really get into the nuts and bolts of healthcare that they can actually influence that. On the other hand, I have to say the FDA has been quite responsive for the FDA in terms of realizing how fast technology is moving and trying to get a handle on some piece of it.

Yan:
They are doing things like a fast tracking certain medical equipment. They're doing fast tracking for drugs. They're trying to rush things that could be very beneficial to the market without giving up safety and efficiency issues a review. So I think everybody realizes, but it just takes time for human beings to really think about it.

Dan:
Yeah. Wrap your head around it.

Yan:
So you're talking FDA, you're talking about regulatory ethical. Lots of issues with ethical. All kinds of issues that need to be discussed, and it's just takes times for human beings to do that.

Dan:
Yeah. Yeah. No, that's super interesting. Well Yan, this was a great conversation and I think gives the listeners a good overview of RPA and its applications to healthcare, and some of the things to have leaders to think about and how it may be impacting the industry. If people want more information, do you have any good resources or ways to get in touch with you too if they're interested in chatting with you about this?

Yan:
Yes, they can certainly LinkedIn with me. I'm on LinkedIn. Yan Chow, Y-A-N-C-H-O-W. Or you can send me email. The best email may be my personal email, which is Y-A-N-C-C-H-O-W@yahoo.com. Y-A-N-C-C-H-O-W. And I appreciate being on the show. It was great to talk to you, Dan, and catch up on future medicine.

Dan:
Yeah. It's always fun. Yeah. We're partners in crime on that for sure, Yan. So I really appreciate having you. Thank you so much.

Yan:
You're very welcome. Thank you.

Dan:
Thank you so much for tuning in to The Handoff. If you like what you heard today, please consider writing us a review on iTunes or wherever you listen to podcasts. This is Dr. Nurse Dan, see you next time.