PEPID’s Executive Vice President and Chief Technology Officer, Edward Reynolds joins the Pulse during HIMSS 2019 to talk about the development strategy and objectives behind the company’s most cutting-edge release. His talk, “Precision without Fatigue: Focusing on User-experience and Technology to Improve Patient Outcomes” is an exclusive preview of the technological complexities and milestones that formed PEPID PGx.
Welcome to PEPID’s Pulse for this morning on Wednesday at HIMSS ‘19.
This morning we’re presenting PEPID Pulse’s “Implementing Precision broadcast series, where thought leaders and pioneers in the practice of precision medicine come together to discuss the challenges they’ve faced and how accessible predictive pharmacogenetic tools empower clinicians to succeed in future outcome-based reimbursement.
Joining us now is PEPID’s own Chief Technology Officer and Executive Vice President, Ed Reynolds. After holding high-profile positions throughout his career in software development and quality assurance, Mr. Reynolds brought his experienced tech strategy leadership to PEPID in 2004.
He is responsible for the development and management of the solutions, content systems, business systems, and technology infrastructure that secures the company’s position as the leader in medical information resources and drug database industry.
Mr. Reynolds is responsible for the conception and development of the PEPID integration model, and for transforming the first health application into a leader in today’s mobile healthcare environment. He is also responsible for developing PEPID PGx, the industry’s first predictive drug-drug and drug-gene interaction checker, which we’ll be exploring in just a little bit.
Good morning, Mr. Reynolds. It’s great to have you on the Pulse! How are you doing?
I’m doing great, Sean. How are you today?
I’m doing excellent and you know it’s wonderful to have your time this morning.
Glad to be here, Sean.
Before we get into some burning questions about PEPID’s most cutting-edge release, I wanted to congratulate you on PEPID’s 25th year in the industry.
With that, you’ve spearheaded a number of things over those 25 years, leading all of these developments. Now that we’re anticipating the launch of PEPID’s PGx tool, how does it feel to see another pioneering IT development become a reality under your leadership?
You know, Sean, it’s nervous relief I think. Anytime you roll an application or program out into the field, you know, there’s a lot of QA that goes into it… a lot of development that goes into it… redevelopment, rescoping. You just wanna get it out there, and get it up and running, and then you’re looking for feedback, right? away. What kind of feedback do I have – is it good, bad, or indifferent? So it’s exciting and nervous and anxious and everything all at the same time.
Have you had positive feedback related to that?
Matter of fact, we have, Sean. In several cases we’ve had comments about the actionability, the action items inside the actual PGx. When we come back with the interactions – and here’s the action you need to take related to that interaction so we have had some feedback on that.
You know, that’s great to hear. Aside from learning more about PEPID’s Pgx, what else can this year’s conference attendees look forward to from PEPID? What are you most excited for during this year’s conference?
I think as far as the conference goes first, I think I’m most excited about HL7 and FHIR. As we move forward we’ve got to begin to adopt the FHIR standards, so I’m pretty excited about that. I’m glad to see they’re being adopted. I’m glad to see our industry is taking it seriously, right? And as patient records become more available, the standards FHIR become more important to us. I’m glad to see we’re taking it seriously.
As far as PEPID goes – this coming year, actually – you’ll see us adding more genes to the PGx. We’re at about 23 genes now, headed up to about 40. We’ll be retooling some of our GUI components on some of our applications, and we’ll also be looking at new dosing calculators and some general adds to our application. Along the dosing lines, we’re looking at a Vanco calculator and a couple other Pharma-related type calculators as well.
That’s all pretty exciting. I know PEPID maintains suites across 13 different specialties, so I have to imagine it’s a little bit complex – making sure everything is up to date and released and ready to go for the users.
You know, Sean, it’s a constant cycle. You don’t ever stop, you’re always approaching the next project or approaching the next roll-out. It’s a constant cycle of QA and regression testing. It just never stops. That’s how we stay on top of it, and that’s how we keep our content updated, fresh and out there. Just constantly updating.
That’s some great insight, Mr. Reynolds. So with that I’m gonna delve into PGx specifically, since this is the Implementing Precision Medicine series. With that, we’d like to know the first steps your team took to conceptualize and create a truly cutting-edge precision medicine tool.
We understand that we’ve always kind of had some level of pharmacogenomic information available in PEPID via our Knowledgebase and Drug Interactions Checker. When did the idea of developing a true standalone pharmacogenomic resource first come to mind for us?
First of all – as you said – we’ve always had some level of genomics in our DI tool. I think the actual genomics tool concept came up probably 4 years ago, when Dr. Rosenbloom started delving into it. From there I guess Dr. Rosenbloom worked the concepts, right? So about 4 years ago.
So it’s clearly a tool that’s been in development for quite some time. So with PGx, as we’ve developed this, it’s clearly a data set that is just massive. I mean, we’ve got a large amount of data all over, in all different places, and this is a cutting edge new tool that’s coming out.
When you and your team were trying to extrapolate all this data, what challenges did you face? How did you solve them? What were some of the issues that came across, trying to get all this into place to function as a true point-of-care solution?
Yeah, it is a lot of data. One of the things we had to do initially was trim the fat, I guess you could call it. There’s a lot of information. In our case, we’re not using all the information so it’s a matter of trimming what we didn’t need, what we don’t want – without losing it, of course. Trimming it away so we can get to the core of the data set in our case.
That’s pretty interesting. Jumping off that, what are some of the technical considerations when you’re developing a tool like PGx that’s so cutting-edge? How do you exactly “future-proof” a solution for such a rapidly-emerging field of medicine. As the science continues to evolve, how do you plan ahead for that?
You know, on the technical side, we’ll kind of back up a little bit. One of the things we had to learn initially was, “What is genomics?”, right? We’re not clinicians or pharmacists on the technical side, but we had to learn some details about what’s a genome, what’s a genetic interaction, what is not an interaction, what are the parameters? There was some level of learning that had to occur in the technical side to at least understand the data set.
Some of the technical considerations we had to get into were: Where are we gonna put this data? How much data do we have? How do we feed it to ourselves? How do we get back to it?
As far as fool-proofing the future, when we built this – and you have to think about it certainly – we built this on a scalable model, right? The modules, the database components. We put together our modular-based, so they can talk to each other and are scalable. We also did a lot of looking to HL7 standards and the standards out there. We’re building something so we know we have to move into the FHIR environment and out into the standard environment.
It was initially a lot of research and, I can tell you, along the way we actually re-tooled more than once. We start down the path, get down that path, decide “Ah, we can’t go this way.” We back up and start again, back up and start again. So we had to do that several times along the way.
That’s pretty interesting, and I’m sure it’s no small feat to be able to get it done. Going from that, drawing from your experience as a leader in technological solutions, would you say that PEPID PGx addresses an unmet technological need in the advent of drug-gene-based interaction precision medicine?
Absolutely, Sean. PGx certainly fits the needs of the clinician workflow as we know it today, right? And you need that UI that drives the clinician to the interaction, drives him to the actionable item, drives him to patient care in the end. That’s what our PGx does. It doesn’t attempt to educate, if you will. It’s not a reference material, it’s not a 50-page report. Rather, it is actionable items for the clinician and he can take action with his patient at that point. So it absolutely meets a need.
The other thing it does, the way it’s structured, I think it removes – and I’ll use the word “fear factor” but I think it removes some of the fear factor from the clinician. When you are presented with a 50-page report on a genomic interaction and what’s occurring it’s a little bit of fear there, right? “I’ve got this report. I’ve got to dive through everything to figure this out.” PEPID kind of removes that for you. Here’s what we already know. Here’s what you need to do, right? So it allows you to just kind of move forward with comfort.
So, I guess I’ll jump into that. So how do we balance that? How do we get information that is comprehensive, usable and actionable at the point of care, and make sure that it truly is tailored to what the clinician needs? Well, not delivering a 50-page report as you mentioned.
Well, it’s just that. That’s how we approach it. So when our PharmDs are working on their projects; when the technicians are working on these projects, we all work under this idea of “We are going to present the information as concise, precise as possible, right? It is one of the building blocks of the PEPID applications, right? Every approach is that way.
That’s pretty interesting, and obviously useful for anybody in the field, right? Any practicing clinician can benefit from that kind of information. So that being said, how do you see PEPID’s PGx tool evolving over the next 5 to 10 years? I mean, we talked about the way that you’re already trying to future-proof it. We’re trying to plan towards FHIR interoperability – other things like that. What else do you see evolving in the PEPID PGx tool?
Well certainly we’re gonna add more genes to it. I believe we will take deeper dives into tumor-related, oncology-related type of environments, right? Other places that genes are going, and other experimentation and learning is going. So PGx will kind of evolve into a bigger tool certainly.
I think generally speaking… genomics, right? In the field it will just become, per se, or just become a daily event. When you are born I assume at some point you would just be tested, and PGx and genomics testing would just be a matter of fact.
Well one thing to maybe consider is today, right? Genomics are being used. And when you go to your physician – what have you – there’s a point I think very soon that you will have your genetic testing with you, right? And your physician is gonna have to understand and be able to react to that test that you’re actually bringing with you. So while it’s kind of ramping up it is coming and you really can’t run from it.
The patient such as myself – I’m gonna have my testing and I’m gonna go to my doc, and I’m gonna say “Hey, I’ve got some testing. What can you tell me about the drugs I’m on, or the conditions I have?” So it is coming kind of whether we like it or not, and I certainly embrace it and I think PEPID embraces it, right? It means nothing but a good for patient care and patient outcomes.
That’s pretty interesting insight, and as you said, it certainly is coming and I think we’re at the forefront of it.
Yeah, you can’t stop it, right? You cannot stop it. And again, whether you’re plugged in or not it’s still coming, right? And I think you’ve seen the trend probably in the last two to three years. You’ve seen this, right? You have people like me in front of you saying, “It’s coming, it’s coming, it is definitely coming.”
Absolutely. So I guess, kind of going off of that, where do you see this tool being implemented? Particularly, PEPID’s been implemented in institutions in 159 different countries, so where do you see PEPID’s PGx going? Where is it going to be implemented that – as we’ve just talked about a second ago – improves patient outcomes. Like you said, it’s all coming. Where is PEPID’s PGx going?
Sean, I think we fit everywhere to be honest with you. We have user-facing applications, right? that’s a doc himself who’s not necessarily attached to a lab, or not necessarily attached to another organization. We have backend API’s and SDKs for our PGx, which allows us to plug into any EMR system or any system that’s being built. And then we have FHIR integrations as well that allow you to plug PGx into any system via the FHIR environment.
So when we first developed our PGx, we kind of thought at the bottom, if you will, right? We’re developing a core, and then from the core we make our APIs. We make our own application so that core and that API level already exists. It allows us to propagate to just about anywhere.
So with that, I’m gonna move on to some of the development strategy and especially the proprietary capabilities behind PGx. So obviously, PEPID has proven itself capable of easily integrating into any system whether that be through interoperability standards like we’ve discussed like FHIR, or InfoButton, or other things, or we’ve also done custom tailored integrations. What specifically from past products or past integrations were you able to learn and then apply to PEPID’s PGx to allow for this ease of adoption of PGx?
Well you hit right on it, right? What have you learned from past projects, exactly. That’s what we learned from past projects: The standards. We learned how to get our nomenclatures together. And certainly everything we’ve learned over the last 10 years just kind of moves forward every time a new project comes up, every time a new application comes out, a new process – we just move forward with that standard.
As we go through this – and I guess it’s just a learning cycle; so, as you go through this, you begin to remove the roadblocks that you ran into last time. So it’s almost a natural progression. The team has been doing this for quite a while so they almost naturally, if you will, start to remove the roadblocks as we roll through these projects.
Yeah, I bet with that many years, going in and doing things…
Right? It’s like we don’t even talk about some things. Sometimes it just comes to everyone and we don’t have to have a debate about it. And not to say there’s not a lot of research that occurs because you have to look forward at the same time, right? You can’t repeat those old mistakes I get again, so there’s a lot of research and a lot of thought but a lot of this just comes naturally to us at this point.
When we talk about FHIR and these interoperability standards, obviously PEPID has got a major shift to adopt FHIR and all of these other latest HL7 standards, including CDS hooks and things like that. So where do you see the role of interoperability standards going with regard to genomic data?
Well it has to be there, right? It has to be there. It needs to tighten up a little bit, right? And I think we understand it’s a new science and there’s new investigations occurring, new findings occurring, so the standards are gonna have to tighten up. But you have to have a standard. If you don’t we won’t be able to communicate, right? And, FHIR, HL7 – these are the best standards we have today to do that communication with.
So we adopt them freely. We also contribute to the standards, and we adopt them, right? You have to have something, and looking back on this – I’ve been doing this quite a while – if you look back 15 years ago or so, when talking about some of the API’s and some of the standards that occurred in the operating systems and applications, it happened years ago, right? So you have to adopt standards or nobody can play nicely. Then all your apps just kind of fall down and all your data falls down, right? So, yeah, we’re excited about it.
Absolutely. I can understand everybody’s got to be on the same page. In a different light, what about regulatory implications of exchanging the data? Is HIPAA actually a sufficient act in this day and age for patients, given that it was a law written a couple decades ago? How does that tie into some of the interoperability that we’ve been talking about?
You know it’s a big debate, right? And we can debate morally all day about HIPAA, but again, you have to have some standard by which you play by or it’s just chaos. So, I personally don’t believe HIPAA is kind of where it should be today but at least it’s something. And we are certainly gonna have to tighten down and we’re gonna have to look at some of the regulations we have today.
As you know, Sean, patient information is becoming accessible, right? I as a patient, I’m allowed to have my record. I’m gonna be allowed to share it, and it can’t be blocked. Anybody can get to it. I can do what I want with it, and it belongs to you, and it belongs to me, right? I should be able to do what I want. So, there has to be standards to lock it down. At the same time though, I have to be able to open it up a little bit, right? I have to be able to provide to my clinician, or whomever I want, my record and they have to be able to get to it easily, access what they need, and then get on with what they need to do, which is treat me. So, I think it needs some work, but you have to apply yourself to some standard at this point or, again it’s just chaos.
Absolutely. I mean, that’s the end goal, right? Being able to move these records around with the end outcome or the end goal being for patient outcomes really, right?
I’m gonna shift gears a little bit. When we talk about the PEPID PGx tool, we hear the word “predictive” a lot when we’re talking about it. So can you describe what that means in relation to PEPID’s PGx tool and how were you able to accomplish it?
Yeah, so the way the predictive side, or the technical side anyway works is you can think of drugs and genes that have relationships on a maybe one-to-many level, and depending on where the drug in the gene fall within these relationships you may have an interaction. It’s a predicted interaction. So another way to think of it as kind of like “mechanisms”. Certain drugs fall into certain mechanisms, so whether or not you’re directly related to that interaction or that gene you fall into that mechanism, right?
So, when the interaction engine fires, it pulls up everybody: Every drug, every gene in this mechanism and then it’s a predictive outcome, right? So, within PGx, inside the interaction monographs you will see an actual interaction, or predicted and actual. So you can see things happening in PGx but it’s kind of a one-to-many mechanisms-related type of thing that occurs.
That being said, we’ve talked about how it’s both predictive and actionable, right? Obviously that is a key feature, but elaborate on that key feature, and what are other key features of the PGx tool?
Well, certainly predictive. You don’t have that. You don’t see that in the world today, so predictive is a big feature. Tt’s a big component of the PGx. Most of the DI engines and the genomics engines you see today are actual, right? So they limit themselves. So that is certainly a big feature.
Another big feature the PGx has is workflow. Again, Sean, we’ve been doing this for 20 years, right? 23 years, right? We understand the workflow of the clinician. We understand the workflow of the pharmacist, so the workflow is built into the PGx as well, right? We understand you’re treating a patient. You’re not necessarily doing research, and that’s what our product is about. It’s about fast access, accurate access, and treating the patient. So that is a big feature of our product.
And then, in general the UI. We built the UI to be kind of fun, if you will. I know it’s not necessarily a fun job but, you know, we work with applications all day and it should be enjoyable to some degree, and it should be fun to some degree, and it should entice you to use it. So, we tried to do that too, right? We tried to make it cool. We tried to put in the, right features, the right information, the right person, right data, right time.
I think the focus on the user role is important, right? I mean because ultimately that’s what matters. It’s being able to get these clinicians to use utilize the tool to improve outcomes, really.
So, I know that as a component of the user interface, you can actually just hover over an interaction and find alternative treatments, which is no small feat obviously, right? Can you tell me a little bit about the development of that algorithm and the power and the speed of the alternative treatment feature of PEPID’s PGx.
Yeah, that’s a cool one, right? That’s my favorite, by the way. So, the PGx by default comes back with alternative drug suggestions, right? Here’s your alternative, so we’re looking at this inside the PGx and we go, “I’m gonna give you two or three alternatives, but there may be multiples of alternatives and there may be multiples of categories of other alternatives, right? As a physician or a pharmacist, you may want to move to a completely different category for your alternative, right? So I can’t show you 10,000 drugs.
What we did is, basically, we start with… “Here’s a couple of known alternatives you might want to go to, and then we share what we call our “table of contents” of the drugs and categories. Then you go to the table of contents – a listing of drugs – and as you scroll over those drugs we actually populate the alternative back into the PGx result, and you can see on your screen – real time: “Here’s the other interactions that you’re about to cause if you replace that drug with this alternative.” Right?
So, you scroll down through your list of drugs. They’re just popping on the screen: here’s what’s gonna happen, here’s what’s gonna happen, what’s gonna happen. You can imagine the speed of that all, right? You just popping down through on the screen and then you decide what you want to do. It’s an amazing feature, right? It’s almost a standalone feature if you think about it, but it’s part of our PGx.
So, the challenge is there, and it took us quite a while to get this, right? The challenges there, obviously, [are] the processing. I mean, you’re processing through potentially thousands of interactions, millions of records to try to get this information to pop on the screen in real time. We went through several different algorithms. What we ended up with in the end was a multi-threaded approach, right? So, there are multithreads happening down here – all over the place – when you start rolling down through this alternative drug picker. So, it was a challenge, right? It’s pretty fun.
We weren’t sure at first if we’re gonna be able to make it work. We had it in for a while. We worked a couple algorithms and, “Oh, man! We got to take this out. It’s just too slow.” And nothing [didn’t] work. It’s just the responses – it was not there. So, we actually took it out for a while. I had to think about it, go back through and rework it, and then, “Ah yeah, we got it.” It’s a cool feature. I like it.
I mean, I think that’s one of the most amazing features. It’s really cool.
It’s fun to watch! If nothing else, it’s fun to watch.
I mean, at the point-of-care you’re literally just able to immediately recommend alternative treatments and stuff. It really is incredibly powerful so it’s interesting that it had to go through multiple revision cycles to make it to its final form.
Yup! Hair pulling and, “Let’s think about it a couple days.” Yup, yup!
So let’s get into wrapping up here, Mr. Reynolds. You know, we’re aware of the constantly changing regulations and standards in healthcare. How does PEPID’s PGx help clinicians at the point-of-care to meet these ever-growing patient outcome requirements both today and in the future?
You know, Sean, I’m not a clinician, right? But I know for a fact that there are regulations in hospitals today, and what have you, that say you can’t be readmitted. If you are readmitted and the insurance isn’t going to pay, there are several – I’ll use the word “penalty factors” – that occur for readmissions. There are certain penalty factors that occur when a patient is in the hospital inpatient scenario, and you have another condition occur – like, a blood clot or something like that. So, I know these things are happening, right? And these penalties are being posed on hospitals today.
PEPID and PGx helps you avoid those, right? Our products are applications that help you avoid these issues, and helps you navigate the regulations that are being imposed today, right? It also helps you keep current with them. It doesn’t necessarily mean you have to know all of them. Although, I suspect most clinicians do, right? If you work in a hospital, you know it’s happening. But PEPID helps you stay in that track, right? Helps you make sure that you’re getting the right patient care, and helps you stay within the regulations themselves, and there are new ones, and they make sense to me [that] if you’re in hospital you should not be readmitted for the same condition within 24 hours. You shouldn’t have another condition occur to you while you were there for the first condition, so it all makes sense to me. So, PEPID helps you with that by providing that the right data to the right person at the right time, right? So, that’s how we’re kind of helping with those regulations
On the IT side, I guess, of the house – I think I’ve already referred to this – we certainly do regulate ourselves with HIPAA standards and HL7 and where they’re headed, so we are taking care of our own regulations on that side of the fence, and trying to help the clinicians, and make their regulations on that side as well.
I think that’s some great insight, and I think you really hit a key note there which is “Right data, right person, right time.” It really is the goal for PEPID, and obviously you are responsible for making sure that is achieved.
To stay on that for a second, again 23 years – if you look where we started, Sean, we started in emergency medicine in the Docs’ hands, right? Right there at the heat of battle. So that’s what we did. We need the right person, right time, right information, so we absolutely know what point-of-care is, right?
Yeah, you don’t have another option in emergency medicine.
Yeah! There isn’t an option. You’re there, man. You gotta do it NOW. So, we know what point of care is, and I think PEPID was point-of-care way before buzzwords “point-of-care” came up. We were point-of-care first, I think. Absolutely. Again, right person, right information, right time.
It’s ingrained in our corporate culture, the way that our data is delivered is obviously amazing for the industry and amazing overall. So with that, Mr. Reynolds, I just want to thank you for your time and for giving us an exclusive look inside PEPID in our development of PGx tools, and in overall industry knowledge for the last 25 years. I’m looking forward to having you back, as well as Dr. Rosenbloom on the PEPID Pulse over the next day and a half here at HIMSS 19.
So, with that, again, thank you Mr. Reynolds for your insights and your knowledge, and I’ll talk to you again soon!
Alright, Sean. Thanks for inviting me, and it was my pleasure to be here. Anytime. Thank you.
For 25 years, PEPID has been a leading global developer of clinical and drug information resources and mobile applications for healthcare providers and institutions. Their trusted, validated content and intuitive workflow provide clinicians peer-reviewed support for diagnosis and treatment at the patient-point-of-care. It’s accessed individually or integrated into any healthcare system.
To find out more, go to www.pepid.com
Questions or comments, e-mail Anna Tan at firstname.lastname@example.org.