Discover the future of healthcare technology with our special guest, Anthony from Dr First. Curious about how a single company can impact one-third of the U.S. population? Anthony reveals their cutting-edge approach to medication management and patient care, from pioneering e-prescribing and medication history services to creating intelligent workflows that reduce friction in healthcare processes. We also delve into their remarkable response to the COVID-19 pandemic, their robust cybersecurity measures, and the groundbreaking implementation of full electronic prescribing for REMS programs.
In the latter half, we tackle the crucial integration of quality and data security in healthcare. Anthony highlights how Dr First embeds these elements throughout their processes, leveraging advanced technologies like AI and machine learning to maintain clean data and robust security policies. Reflecting on the evolution of AI from the 90s to today, we explore the significance of rules-based systems, genetic algorithms, and autonomous agents in healthcare. Plus, get a sneak peek into Dr First’s future plans, including their exciting acquisition of MindShift, aimed at further enhancing healthcare solutions. Don't miss this episode packed with valuable insights on how technology is transforming healthcare for the better.
Discover the future of healthcare technology with our special guest, Anthony from Dr First. Curious about how a single company can impact one-third of the U.S. population? Anthony reveals their cutting-edge approach to medication management and patient care, from pioneering e-prescribing and medication history services to creating intelligent workflows that reduce friction in healthcare processes. We also delve into their remarkable response to the COVID-19 pandemic, their robust cybersecurity measures, and the groundbreaking implementation of full electronic prescribing for REMS programs.
In the latter half, we tackle the crucial integration of quality and data security in healthcare. Anthony highlights how Dr First embeds these elements throughout their processes, leveraging advanced technologies like AI and machine learning to maintain clean data and robust security policies. Reflecting on the evolution of AI from the 90s to today, we explore the significance of rules-based systems, genetic algorithms, and autonomous agents in healthcare. Plus, get a sneak peek into Dr First’s future plans, including their exciting acquisition of MindShift, aimed at further enhancing healthcare solutions. Don't miss this episode packed with valuable insights on how technology is transforming healthcare for the better.
More at https://linktr.ee/EvanKirstel
Speaker 1:
Hey everybody, we're going to the front lines of healthcare technology today, with Dr First doing amazing work, empowering patients and providers to achieve better health outcomes. Anthony, how are you? I'm doing well.
Speaker 2:
Thanks for having me, Evan.
Speaker 1:
Thanks for being here. Irma and I are always intrigued to have chats like this, particularly with industry thought leaders and insiders like yourself. Maybe start with introductions who is Dr First? And maybe describe the incredible work you and the team are doing?
Speaker 2:
Excellent.
Speaker 2:
I think that Dr First is one of the organizations in healthcare that a lot of folks don't know about yet.
Speaker 2:
We touch about a third of the population of the United States, so we are a, which I think makes us about the second largest healthcare company in the US. And so we do that through medication management, really helping on all aspects of that From our original products, where we empower and kind of actually are the embedded version of prescribing e-prescribing inside about 275 electronic medical record systems, probably have another dozen or so that use parts of our system for their electronic prescribing. We do that on the patient side, where we are, that message that's coming from your doctor that says, hey, I just prescribed this to you, you need to go pick it up here. Here's what you need to know about this particular drug therapy. And so, and throughout that entire journey, whether you're on the provider side, whether you're on the patient side, you're trying to work out eligibility, some type of specialty drug. We help try and get rid of the friction by applying intelligence, AI and smart workflows wherever possible.
Speaker 1:
Wow, that's quite a lot to unpack. Very impressive. Maybe describe the impact not just on patient care but the provider operations and the impact to both of those, and what are some of the top innovations you've been focused on the last couple of years, through this really challenging time in healthcare?
Speaker 2:
So I think, like most, we heard the call and did overtime during COVID and really tried to support with our collaboration tools with our having data available.
Speaker 2:
In more recent times, I think some darker things have come through where there have been a lot of cyber attacks and we can't provide a whole picture because we're not an electronic medical record system, but we can act as a backup to that to help you limp along digitally through some of our standalone tools.
Speaker 2:
So we have about, I guess, 80 percent of the population in which we are able to pull the majority of their medication history and that'll let you know drug interactions which are very critical in assessing care when perhaps you are unfortunately the victim of a cyber attack and can't access those medical records. And so I think there's things that we've been focused on on continuity, but I think we continually focus on where is there opportunity to get rid of friction? One of the things we rolled out this last year, I believe we are the first ever to have full electronic prescribing for a REMS program and that's the most restrictive form of specially prescribing out there, and we've been able to apply the same techniques that we do in controlled substance prescribing, with the additional consents and enrollments and things that need to go along, supplemental with every single script and enrollments and things that need to go along supplemental with every single script.
Speaker 3:
Wow, that's quite a breadth of services you provide. So clearly, navigating the healthcare market is no easy feat. What specific strategies have Dr First used to successfully bring new innovations, new software into the market, and how do you ensure that you meet specific needs on the provider side and then, you know, improving patient's care in the end?
Speaker 2:
Irma, that's a really good question, one that we constantly are working on ourselves, because we have tried to bake innovation into the formation of our company. It's core to what we do. You'll just, if you slowly watch us, you'll see that we regularly I think this morning I got notification of three more patents we just, consistently, are adding capabilities, and we do so by talking to our customers, understanding where they're having pain, understanding where they're having a problem, delivering care and trying to do the Venn diagram over great. Not only can we solve that, but should we, because it's part of our core competency, and then, when we can, we'll help. An example example that would be.
Speaker 2:
Unfortunately, the last couple years have seen a rise in fraudulent behavior. Someone's always been trying to steal the script pad and get access to, but it's gone to a whole nother level, to where it's almost and sometimes it does look like state level actors are involved in trying to game the system in one way or another to get access. And it's gone beyond. Just let's have some type of access to opiates so that we can sell them on the black market. And recreational? No, this is. How do we have our weight loss drugs? How do we have our AIDS medications, our cancer treatments and taking them to off markets or other like fraudulent behavior. And so we started about two years ago noticing an uptick in identity theft, specifically where a perfectly awesome medical provider was doing their thing, just like any of us maybe would understand, on a credit level, and you pay your bills on time and your mortgage oh, that's great. But then someone does something bad by taking your identity and they do it somewhere else, so it's a little bit hard to see. They're not within your organization, they're not using your same medical record system. They have completely taken your identity and been able to get on to a prescribing network, adhering to all the compliance requirements, getting through identity proofing and recognizing that.
Speaker 2:
We realized that, touching about one in three prescriptions, we had a unique view. So my data science team and I had an idea of how we could combine a rules-based system and a self-learning ML model to analyze and group people's digital fingerprints. Essentially, how are you prescribing? Not because you say you're an oncologist, but from the actual scripts that you've done on the network. Let's have a look at how are you prescribing and we model over. I guess we're up to about 280 relationships now, and a relationship could be anything from? What's the distance between you and the place where you want this dispensed the pharmacy? What is the type of drug? Something obvious like that, but it could also be. How do you type?
Speaker 1:
What's your physical?
Speaker 2:
hardware that you're using and in particular, we found success with a model that's called ECOD and what it does is it uses each one of the relationships that you pair to create groupings or clusters and say here are providers that sort of behave the same, and then from us we look at the deviation to see how far away are you from your closest or the group you belong to the most, and then rules will come in and do more of the business logic of if someone's trying to get codeine promethazine for recreational purposes, we can put rules on top of that as well, and they blend together to end up creating a scoring system. So we use that internally first and we were pretty successful, so much so that we were able to get a couple hundred folks off the network that we decided they were just entirely a poor risk profile and we didn't have anyone call us and contact us of like, hey, why'd you shut off my account? We're like, yeah, we probably got something real. Now we're starting to work with pharmacy chains and we're going back and forth and sharing that information in real time with them so that they can decide for themselves.
Speaker 2:
Is this something I should dispense? And we do believe. We call this TrueRx, and we do believe this is an industry first where we not only have a measurement of trust on the NPI kind of level, but actually on every single prescribing event level we have an outlier score. So they can understand. Should I give this one a little bit more attention and then create rules around it?
Speaker 3:
Sorry, this is so interesting. I guess we all knew that patient data is a high, you know good commodity to sell on the market and just we haven't seen that kind of inside of you of how you can prevent identity theft on the provider side. You know, I don't think people typically think about that so clearly. You mentioned partnerships. So Dr Furse has worked with major clients. You can mention names if you'd like, but I'm interested can you share maybe another story from your work where your solutions had a specific impact on client operations or maybe patient outcomes? Just give us a little bit more color.
Speaker 2:
Absolutely, Absolutely. We have a couple of different products that help on that outcome side. So first we try and have fully clean data, which is the right thing to put in front of a prescriber at the moment they're choosing a drug therapy for your patient and it's sort of the unsung hero where we've had AI running on that for about eight, nine years now and so that they have perfectly clean, normalized data, which doesn't always exist in the U, unfortunately in healthcare and that's so we get them up to that moment. Then we have tools that help us go through the prescribing sequence efficiently. Where they have the right information can make the good decisions, understand what are the benefits for the patient, what co-pays might be. Let me select this therapy versus that one because it'll be a little bit cheaper, but it's the same drug therapy I'm trying to propose for this particular patient in my care.
Speaker 2:
We have a new orchestration platform coming out for the getting it there I described a little bit like the REMS program of where it's an unusual script and not the standards, and so it needs a little something extra. And then we have the patient side and we have two elements to that One which helps pre, where we actually identify patients that we predict are not going to be adherent, they're not going to pick up their script and surface that up to health systems so that they can choose to do outreach to that. We also have just purely a lift that one of the exciting stats we have is on those messages we send out on behalf of your doctor. Exciting stats we have on those messages we send out on behalf of your doctor to say, hey, I might prescribe this particular drug therapy. This is what you need to know about it. It's at this location you need to go pick it up Is that we're having about an 85% click-through rate on that and that's something that we started many years ago in the, I think, upper 30s and we've just progressively gotten better and better and better at that and what we see as an effect is that that lifts the entire population so that they are actually going in picking up their prescriptions that the doctor said, hey, you need to do this.
Speaker 2:
They don't get lost in their life or get a little uneasy because they weren't quite sure what it was going to do and what the effects were, because we have the information for them on their phone where they just click no apps, nothing necessary. You just click and go. You need your date of birth and from that I think that's one of the areas we're really really pushing. Outcomes in general is that we're able to help people understand why they have particular drug therapy and what it's going to take for them to get it, and then often help them, if at all possible, with some kind of coupons or some kind of way to have a discount on that. If it's available, We'll match them to programs. Other times we're just giving them the helpful information they need to understand where is it?
Speaker 1:
Oh, fantastic approach, Switching gears to the providers. You work with IT leaders and CIOs to deploy this technology and, as you know better than us, there's a lot of legacy, a lot of technical debt. The industry is fairly slow, conservative, let's say, to adopt new technologies. New technologies how do you help them understand the business impact or the ROI even of investment in new technology? And then how do you measure impact, success and communicate all those metrics to stakeholders? That seems like quite a job as well.
Speaker 2:
It is.
Speaker 2:
Luckily, most of our tools are delivered web-based and so they rely on us to be that front line that is, keeping it up to date, and that's one of the reasons that so many electronic medical record systems like using Doctor First is we keep them out only for the portion of medication management current on technology, but we also keep them current on regulation, and so whenever there's a script, standard or update, they don't have to think about it, we just take care of it and it magically shows up.
Speaker 2:
And then one of our primary use cases for our actual e-prescribing software embedded that all skins and looks like the tool that you bought as a provider. So as a provider, you don't even know you're using Dr First software. The workflow is unique to your setting of care and the way that your vendor of choice for your electronic medical record system has presented it. You get to customize it further through what we offer, and for them it's fairly seamless, and I think that's where we have the largest impact on keeping them current on technology is we do most of the heavy lift for them.
Speaker 1:
Well done.
Speaker 3:
Interesting how you are behind the scenes sometimes, where even providers might not know that they're using your product. So now let's talk about data security, specifically in healthcare. Obviously it has to be a top priority. I already talked about some of the measures you take. So how do you address that data security in the wider sense and the compliance with regulations? You mentioned how you help providers kind of follow them seamlessly. What innovations specifically in that arena have you implemented?
Speaker 2:
Yeah, so I think that's a great area of focus that we just constantly have to build in. It's not something you add on. I'm a classically trained manufacturing engineer and they drilled into us 35 years ago when I was doing that that quality wasn't a step at the end. Right, data security isn't a step at the end. It's something you bake into every step of the process, and I think that's where we continue to innovate. What we help in particular is by having clean data. By the way we integrate to our customers. Their data is secure because of the policies we've baked in.
Speaker 2:
So it's less of a data security innovation and more of simply how we do business and very much a consistent process, and where we literally have AI that detects anomalies in data and then can flag that right. That's what they need to understand in order to be able to have good understanding of data where it's propagating to and I think some is in our own work to, where we tend to maybe be a little bit slower in healthcare at adopting some of the newer exciting media generating technologies like generative AI, but that's done because we are always wanting to apply those in controlled, hipaa-compliant environments so that data doesn't propagate where it shouldn't that we understand what's happening with the data prior to that, and that's why we concentrate heavily on first our own internal models, so just ML models running in our own environment that we understand, and then, more recently, we've started to use large language models, but ones that are already HIPAA compliant.
Speaker 1:
Well done. So looking at your bio, I noticed you were doing quantitative trading with AI and machine learning back in the 90s, so we've come full circle. How big of an opportunity is this in healthcare now to leverage not just LLMs but AI and ML in general, and I think this is a much more satisfying use case of AI than leveraging it with hedge funds and trading. No disrespect to that it is.
Speaker 2:
I made the jump into healthcare almost stumbling. I responded to a Craigslist ad and ended up starting a healthcare company in 2011. And I fell in love with that. And you're right, it was just. It was the human aspect of being able to actually impact people. To your question glad you asked it that way. Yes, I entered computer science field enterprise computing at the first run at AI, our Bayesian networks and our expert systems and applied those in nuclear and satellite environments first and a lot of control systems and manufacturing.
Speaker 2:
What's interesting to me is to see as much power as there has been in manufacturing, specifically ML models and I'll just loosely call those neural networks. We have gotten away from some of the rules and I mentioned that a little bit on the fraud, that often in order to do something well, you need to not only have the machine do the magic logic and learning but then you need to constrain it and say, hey, this is the objective, this is the goal, and rules often are a good way to reinforce that, and we found that the interplay between the two works particularly well and so, for instance, on that 2RX system, we have about 50% that is a chaining rules system, expert system system and 50% that is really ML. And so using your what is now almost ubiquitously called AI but it's the blend of those two and the piece that I think folks forget a little bit, maybe because we have so much data that flows through our world now is there's another technique that I particularly favor, which is genetic algorithms, and it's almost fallen out of favor, but I like it because you let the computer simulate many, many, many evolutions with a little bit of randomness thrown in there, like nature gives us, and almost design the best path, the ideal flow. And you can see that with our creation of synthetic data, with ML models, that we often are achieving a little bit of that. But I think that's an area where there's still opportunities to take some of the lessons from the old days.
Speaker 2:
And then, most recently, I would say I sort of giggled at myself that I dusted off a textbook from my library on Blackboard systems, something that had its moment for maybe a year, and I don't know, maybe 93, 94, somewhere in there and the idea was how do you have a bunch of autonomous agents do their thing?
Speaker 2:
They do, but not necessarily talk to each other directly with APIs, and that's exactly the problem that we're trying to solve now in AI as we're moving towards these autonomous agents enabled that they loosely know each other, but there's a place they can share, like a literal blackboard that they're writing each other. But there's a place they can share, like a literal blackboard that they're writing up with chalk on what the problem is, what needs to be solved, and be like oh, I'm an expert, I can do that little piece. That's something I know how to do and I think you'll find, paying attention to some of the more recent developer AI conferences, that right now is one of the quintessential problems we're trying to re-identify how to solve. I also think it's hilarious that in that same timing, everything was an agent and it was all agent-based architectures, and we're there again.
Speaker 3:
Yeah, what's old is new again. Yeah, so maybe my prior life. I was a software engineer for 15 years and then. So Evan shouldn't make fun of me when I still keep some of my old textbooks right Because they could come in handy.
Speaker 1:
Throw them away. Oh my gosh. Dozens of them, Old textbooks from the 90s Unbelievable.
Speaker 3:
All right. So back to you, anthony and Dr First. So how we would like to know maybe some of your future plans near future, company-wise, personal. Are you going to the conferences? Where can people learn more about Dr First? What are you excited about for the next few months?
Speaker 2:
All right. So in the next few months, things that I'm the most excited about is and I think you actually spoke to our chief medical officer, dr Columbanis, on this a little bit about our acquisition with MindShift and they had some pretty novel AI that basically took the friction out of prior authorization. So definitely excited with that that now in the next couple of months we get to start marrying that into our core volume and so we get to bring that into our prescribing network, because we've done the tooling and we're now ready for that integration. I would say another area that we're exploring a little bit more futuristically is everybody is working on an ambient listening. Yes, but what we focused on was we have started to create a context window for a patient which then allows us to reason upon.
Speaker 2:
What do we know about them? Not that we're the master record, but it's the best known good for this patient. What can we do to help? Where are there gaps in care? Where are there gaps in adherence? Where is there some type of an issue that we should alert the clinician that they can make an informed decision, and I think that that's going to bear a lot of fruit.
Speaker 2:
First place that our own users will see it is very simplistically and being able to have voice prescribing, and that's great, but we've had dictation for quite some time, so it's not as exciting for me for quite some time. So it's not as exciting For me. What's exciting around it is shifting the method in which that's done to where we truly have a unified patient context window that goes across our entire platform. And so, whether we want to personalize the engagement for the patient, to make it a little bit stickier and for our patient side of the house are doing some really, really interesting work with John Hopkins University on that of what works and how to personalize information. All that comes back to having a what do you know about this person and what can you do with that? And since we touch each piece, we recognize that, instead of just helping with prior auth or prescription or medication history or the engagement method, we actually know all the stakeholders involved in that patient's journey, and so hence, let's put them together in one context window and be able to reason upon that.
Speaker 3:
So interesting.
Speaker 1:
Yeah, fascinating work. Congratulations on the amazing work and the mission. Can't wait to catch up at one of the many meetings and events this fall. Thanks, anthony. Thanks for joining you, bet. Thanks for the time.
