Valant CEO Ram Krishnan on EHR Tech and AI

April 24, 2023

#FuturePsychiatryPodcast discusses novel technology and new ideas in the field of mental health. New episodes are released every Monday on YouTube, Apple Podcasts, etc.

Summary

We spoke to Valant CEO, Ram Krishnan, about Valant EHR. Valant is an EHR system designed specifically for behavioral health practices. It offers a wide range of features, including documentation, scheduling, billing, outcome measures. We discussed some very interesting topics pertaining to EHR systems: heterogeneity of needs of users, measuring and incorporating usability, how to manage feedback, understanding design bias, and how AI will integrate into EHR systems.

Chapters / Key Moments

00:00 Intro

05:21 What is it like working at Valant?

08:53 Why would a psychiatrist want to use Valant vs another EHR?

14:54 How can an EHR meet the needs of all users?

17:52 How do you incorporate usability data to build a better product?

22:45 How do you decide what the next features will be?

32:38 How do you anticipate AI will integrate into EHRs?

38:02 What will be barriers to incorporating AI?

EHR Technology and AI

[00:00:00] Ram Krishnan: 50 years from now, I just hope, I hope mental health is is health. It’s not, it’s not distinguished from physical health and it’s one, one holistic view in that we’ve harnessed and wrestled technology in general in a way that is net positive for humanity. And not net negative in that the solutions and software that come along to support it are hyper-focused on, making their lives better. And that’s pretty high, like visiony set of answers. But I think that’s the best I can hope for.

[00:00:32] Bruce Bassi: So welcome to the Future of Psychiatry podcast, where we explore novel technology and new innovations in mental health. I’m your host, Dr. Bassi, an addiction physician and biomedical engineer. If you’re joining for the first time, I’d greatly appreciate if you subscribe and share with your friend network and social media. Also, additional resources, a full transcript and a discussion forum can be found on our website. Thanks for joining us and I hope you enjoyed the discussion today.

Well, today we have Ram Krishnan, he’s the CEO of Valant, which he has been since 2020. And Valant is an EHR system designed for behavioral health professionals. And I understand it was designed by a psychiatrist, Dr. David Lischner. And I think that’s a really attractive point to a lot of new users because they wanna know it was designed by someone who knows their pain points and not some corporate entity that doesn’t have the, quite the same experience.

And I hear that all the time when discussing EHR systems. So I think it’s gonna be a fun conversation because you’re the first individual I have on here talking about EHR systems. A lot of the time when we talk about treatment– future treatment plans, we talk about new medications, novel treatments, but EHR systems have such a huge role in patient satisfaction, clinician satisfaction, patient outcomes.

They’re, they’re the central hub there of all information moves through the EHR system. So I thought it was very fitting to talk to somebody, especially a CEO of a major EHR system like Valant. And I wanna pick your brain a little bit and talk a little bit about Valant itself specifically, and EHR systems.

So tell me a little bit more about Valant and why it was initially intended to focus on behavioral health?

[00:02:08] Ram Krishnan: Yeah, thanks. First of all Bruce, thanks for having me on your podcast. It’s it’s a lot of fun to get a chance to talk to people about how we think and how we build things. And also, often I think in our industry, the EHR isn’t exactly the word, three letter word that people love to, to say positive, wonderful things about.

And at the end of the day, it’s a group of people that’s our team working on and trying to solve problems. And I think exposing some of that is, is helpful not just on behalf of us as a company, but our industry because I think there are many of us that have worked at different companies and places, and we all, we are all here for a similar passion and purpose.

Even if you don’t always feel it on your end I think our intent and our desires are to solve problems and improve healthcare.

[00:02:57] Bruce Bassi: No, you bring up a good point. They get a bad reputation, I think, because people use them to do work. So no matter how good you make an EHR system, if it feels like it’s creating more work for me, I’m gonna be angry at it.

[00:03:09] Ram Krishnan: Yeah. And I was thinking about this analogy the other day and and I think it’s really appropriate for healthcare. One of, one of our customers referred to our product as the operating system of their business. And if you think about operating systems in software, like you’re on a computer right now, and so am I, there’s really. One dominant operating system out there, Microsoft, followed by Second Mac os. And then maybe a distant third, I’ll give some credit to Google Chrome the Chrome os. So there’s three operating systems that most of humanity has to learn how to use, and there are probably a thousand electronic health records out there.

And so we have a whole lifetime of practice on these three operating systems for using computers, but in the operating system of our jobs in healthcare, there are thousands. So moving, if you move from one practice to another, you have to learn an entire new operating system. I mean, people won’t switch from an Android phone to an Apple or vice versa, because they’re anchored in on the way it works.

And so that’s a big challenge to analogy overcome.

The original question you asked, part of the impetus of founding the business is we had a, our, we were founded by a psychiatrist, Dr. David Lischner, and he ran a psychiatry practice in Seattle. It’s the evidence-based treatment centers of Seattle.

Great practice. Very focused on outcome measures, and in trying to create objective evidence and measurement-based care.

One of the, early proponents of value-based care and measurement-based care and he just felt like, this is I say this all the time, it’s that tale as old as time, right? There wasn’t a system suitable for his discipline in his domain. It was, everything was generic for healthcare in the outpatient setting and not specific to psychiatry or mental health in the private practice setting. And so he had a brother who wrote code and the two of them got together and said, why don’t we take a crack at this?

And they did. And that was the founding. And their first customer was his own practice. And I think that beginning starting there with that perspective in mind makes all the difference in establishing the tone, the culture, and the way you think about designing products.

EHR technology has potential to greatly impact patient care and clinician happiness

[00:05:21] What is it like working at Valant?

[00:05:21] Bruce Bassi: That’s pretty awesome. That’s neat. What is it like working behind the scenes at Valant? Maybe we’ll talk a little bit more specifically about Valant as a company and its culture, and then talk a little bit more bigger picture.

[00:05:31] Ram Krishnan: Yeah we’ve we’ve gone through some changes. We were like many other companies, we were based in an office in Seattle, and when the pandemic hit, we quickly pivoted to a remote culture. So our workforce is now primarily working from home and working from home, not just in Seattle anymore, but in places all over the place.

I, I’m the CEO of the company. I live in San Diego, California.

We have a big workforce in Seattle, but we’ve gotten really adept at working from home in similar to kind of our founder focused on outcomes in his practice. That’s really how we run too. We are really outcomes focused in terms of production on the team.

We do a lot of trust. We do a lot of focusing on people, being able to manage their own schedules and time as long as they’re delivering the output and making the commitments that they have for each other. We’ve made a very clear purpose, mission, and vision that we’ve set for the company, and we’ve used that to guide our decision making with how we make investments in our product and as a team.

If I can, I’ll read a couple of those to you because they’re helpful in in the way we’ve even incorporated words and language.

So our purpose is to make the world a mentally healthier place. And we state that specifically because our focus is on mental health. Our focus is in behavioral health.

If we have an opportunity in that spans multiple disciplines or goes into another area or asks us to make software development changes that help support primary care or to help support dermatology, not that would, that it’s a logical connection, but we’re likely to say no, because it doesn’t line up directly with our purpose, which is studying and becoming experts in this discipline and understanding what our customer problems are. Even in our customer base, the wide variety of challenges for us to solve are profound enough without trying to go dabble into other disciplines.

[00:07:20] Bruce Bassi: Hundred percent.

[00:07:22] Ram Krishnan: Yeah. And so we then we set a mission. Our mission is built around providing technology and services that connect behavioral health patients and their providers when and how they need care, in a way that improves outcomes for all. And those outcomes are clinical, they’re financial, they’re operational, all three.

So, and we said it that way because the places in which the care is happening is evolving, right? It’s not always office based. It’s becoming– in conjunction with office space, it’s becoming a hybrid and often remote. And I think, I really feel down the road it’ll also become sustained after the session, right?

So after the appointment or the session, there is that time between sessions where the patient or client is still needs support to ensure that whatever’s in your treatment plan continues to be top of mind and continues to happen. And whether that’s remembering to take a medication or it’s remembering to be mindful or jotting down things that you’re grateful for or checking your frame of your, all these things are what the happens in between sessions, that, that are also part of the care continuum.

So that’s our mission. And then our last thing is I’ll give our vision and then I’ll stop there. And that’s just– we stated that we wanted to be a one-stop shop for our behavioral health practices for their ever-changing needs. And that just means we have to be students of the market, students of the industry, and we need to do our best to try to be ahead of the changes, interpret the changes, and make sure we’re there to help with them as they evolve.

[00:08:53] Why would a psychiatrist want to use Valant vs another EHR?

[00:08:53] Bruce Bassi: So Valant specifically is designed more towards behavioral health clinicians and what types of features are more attractive to behavioral health clinicians? Is it the fact that there’s more measurement based, care integrated within Valant? Probably not as much image annotation types of features because we don’t really need that in psychiatry.

But anything off the top of your head about what specifically about Valant would make a psychiatrist more attracted to it versus another EHR system?

[00:09:23] Ram Krishnan: Yeah. Great. Great question. And I think I’ll even, maybe I’ll take that back a layer and say what problems are unique to psychiatry or psychotherapy or the discipline of behavioral health, what’s unique? And there are some obvious examples and then some less obvious examples. So I’ll maybe give you a few.

On the measurement-based care side, understanding that for us, a PHQ 9 is our version of taking blood pressure in a way, right? Like it’s the measure, it is using that as one example. There are hundreds of measures you could choose from, but that’s one most people know. And it’s a measurement that starts at the beginning, you take it again, months later to track improvement and understanding that needs to be executed throughout the workflow, patient flow. And the provider flow isn’t just a transactional feature that you have the ability to deliver PHQ nine and measure it. It goes beyond that.

It goes to, okay. At intake I want that to be a part of the intake packet that goes out and it comes back in. I wanna capture that discreet information and capture a score that is gonna be chartable over time so I can show progress visually to my patient. I can have that progress visualized in the chart, and I can have that progress visualized in a report that I send to a payer to get paid for, as an example.

That whole thread, designing the feature set to understand its purpose, not just activate its capability, is, it, is a fundamental thing we’re trying to change along that way. At a PHQ-9 is really simple, nine questions with the scale. But what you score on each question means something specific, potentially as an aggregate.

And one of the really, I think, powerful features, our founder really spent time understanding and building knowledge around was when the patient or client fills out that score generating narratives that are attached directly to what that score is, right? So when you’re doing your progress note or you’re documenting to have a baseline narrative already written for you that you can edit it, edit is important for standardizing care and just driving efficiency.

So you’re just, that’s not something you’ll see in a multi-specialty ehr, right? That’s something you’ll have to study your domain to understand.

[00:11:39] Bruce Bassi: Yeah. And you’d have to probably copy paste the results from that. And it’s a add added step for a lot of EHR systems. I think it’s not integrated into a narrative that you can utilize and edit thereafter.

[00:11:51] Ram Krishnan: With a chart that’s visualized the progress that you can put into a report to send to a payer to get paid more if you’re getting bonused on it, to send to the patient in an app so they can see that visualizing progress is powerful for clients and patients too, right? To see the visual. I may not feel great today when I walked in to talk to you.

But if I tracked, if I showed my that this is making a difference, your scores are declined, they’re getting better. That’s helpful.

Another really simple example, and this is more on the therapy side, is we have a very extensive feature set on group therapy. Group groups are not, you don’t do group dermatology or group podiatry, right?

That’s a really different understanding of a domain. Everything from group had a register group together that’s active together at a time to tracking a group note, a group progress note, splitting that all now into a way that bills eight different ways for eight different payers of people that are in a group session executing a group telehealth session.

It’s just, it is very specific in that regard. And then maybe one last example here is because we have lots of these but a good one. I like this one because it, it was one that we like many of our ideas come from customers, right? Giving us direct feedback. And when the pandemic hit, the first thing everyone had to do is figure out how to get on one of these sessions with their clients.

And the easiest thing in their right, in front of their hands was to grab Zoom, which took off during the pandemic. And then right behind that, people started realizing they had some of that capability with teams which is a really great stop gap in the short run. But one of the things that was a, an unintended consequence was that we put an additional administrative burden on the provider.

So as a psychiatrist, if you were used to having somebody show up to your office and start with an administrator at the front desk, to ensure paper was complete copays were taken. Whatever other things you needed to get done were taken care of. And then you saw the patient. But in the advent of the Zoom Teams, we just dropped patients right on your lap and said you should start taking payment now.

You need to get this, make sure this assessment’s done. You need to make sure their appointment, the next appointment is all this went right on you. And we that was feedback we got from a client. And we started to realize that’s consistent across our customer base. And one of the things we ended up building our own telehealth solution and our telehealth solution is integrated with lots of different parts of our system. So when a patient now gets a link to start their appointment and they’ll click on it, the first thing they’re presented with is our virtual administrator who will say, Hey, before you start this session, you have a $20 copay, your credit card’s already on file. Can we take that now? And that part’s out of the way.

You need to fill this assessment before we get started. That’s was your deliverable for this week session. And if it’s quick, we just do it quickly online. Right. And any other administrative work gets done digitally. And it’s great cuz the provider, if the patient didn’t like that, then love having to do that.

The product can always blame their EHR, right? In that case. And we’re happy to take the blame in that case. Cause it just sort of takes that, that the sacredness of your relationship with your patient, it gives it back to you.

[00:14:54] How can an EHR meet the needs of all users?

[00:14:54] Bruce Bassi: In that point, specifically in that example, I, it reminds me of what you were saying at the beginning of this session before we started recording, about how the greatest feature about your EHR system is ease of use and the greatest drawback can be ease of use too. Because when you were giving that example, I thought of a couple other clinicians I had spoken to recently, how they have a patient population that is very rural and all they care about is clicking one button, one button specifically, and seeing a person there and not having to deal with logging in, doing forgot password, two token, authentication,

[00:15:31] Ram Krishnan: Yeah.

[00:15:32] Bruce Bassi: …another screener, so it’s. I love how streamlined that is, but then I know for a fact it won’t fit the needs of everybody, which is the challenge that you have. How do you work that?

[00:15:44] Ram Krishnan: It’s a great question, and I told you we were gonna go all kinds of places in the conversation and it’s a good one. It’s the, and I love that you brought that up because that is at the heart of software design here, our biggest challenge is that the provider base and the client base are not uniform.

They’re not automatons. There’s a wide variation in patient type, from the rural example you gave, to senior seniors, to English as a second language to different patients– there’s a wide range, right? And so I think it’s hard to get it right for everybody. I think what we try to do whenever we can is create optionality.

So the example I gave you with that virtual admin, you can also shut that off and you can skip it. You don’t have to do that. It’s not mandatory as part of the workflow. It’s a– you opt in for that as a feature set that you can turn on because it makes sense for your client population, and it makes sense for your providers and it makes sense for your practice.

But if it doesn’t and the example you gave, you just don’t turn that on.

[00:16:44] Bruce Bassi: Right, right.

[00:16:45] Ram Krishnan: But you can imagine as each of these, like permutations start getting built and you have multiple on off switches, you suddenly have a product and a service that’s so feature rich but is a sea of features that a secondary challenge for us is: how do you have, create all this optionality and all this functionality, but make it easy still for people to use?

And how do you, and the flip side of that is how do you get people to discover all these things that they’re there after they’ve been using it for years? And those are also subsequent challenges and they really point to that ease of use.

Some people have really fine tuned and optimize it to meet their needs. And some people, they take it out of the box and then we don’t go, they don’t go much further than that and start hitting constraints.

And it speaks to, even here, it’s a continuing education environment. I can tell you I still discovered new features in Excel to this day that I wake up and learn and go, oh man, if I knew that it was, it saved me I don’t know how many hours of of work. So, so I think, yeah, these are all some of the big challenges that we run into.

[00:17:52] How do you incorporate usability data to build a better product?

[00:17:52] Bruce Bassi: Yeah, that’s it. It’s a really interesting topic, one that I really enjoy just chatting about. Because you know what, in graduate school I took a few courses on usability and for audience members, they, you can really get very granular when you’re measuring usability. How much how often are you using the mouse, how often, how many clicks are you making per patient that are not related to the note perhaps, and how do you, measure this data and then also interpret it too when there probably is a little bit of variability there among the clinicians day, depending on who they’re seeing.

And then incorporating that into your product development side to improve your product. It’s just really kind of cool. I like that because you have a way of having a very large impact in fact, when you’re just tinkering with maybe the location of one link, but if it’s done with a million different clinicians or a million different visits throughout one day, four seconds, times a million is you’re saving the entire country like a lot of time right there.

That’s pretty, pretty neat if you think about it that way. What types of usability features have you measured with your team and what have you done with that?

[00:19:08] Ram Krishnan: I love that you’ve studied it before because it is a science. And unfortunately though, when you’re doing it here, you find that it’s not a science of the median or the mean or the, or even the mode, right? Where you are looking at, okay, if I do this, half of the people will benefit from like, the average person will benefit, uses this route in the product, and they’ll benefit from this change.

It’s also, but we live in the, in sometimes in the, on the other two sides of the distribution curve, and often that’s where you get the most noise. And so the–

[00:19:38] Bruce Bassi: The most vocal people.

[00:19:40] Ram Krishnan: The most vocal people are on those edges, and they can sometimes drive you away from what the median or the mean is. Right? It’s a, it’s really important to have objective evidence and data to help you make design decisions, or you will end up listening to the loudest people on the fringes and make everything inefficient for the median in the middle, right?

Like, and so we do a handful of things and I’d say there’s some that are big activities, some that are kind of standard activities.

And so we recently redesigned an entire patient portal experience. And that was how we rebuilt it, we relaunched our new version in the fall. It’s a mobile app and it’s a and it’s a web-based as well.

So we have both components to it for the patient to interact with our providers in the practice. And we hired a third party to be frank, like we hired a third party.

We told, we had them interview our practices and the providers and their patients, whatever they signed up for. We had them do the interviews.

We, so we had to take our bias out of it, right? So our design bias and our historical bias, we had. Go interview them all. And then come back with the top five– what are the most important work streams that we need to build, how to optimize them based on this feedback?

And they, came back with a design paradigm for us, and they also tested it back with their survey users. And so that was an important way to design a feature set on a strategic basis.

And in usability, there are things you do where you articulate what the task is and you measure the time it takes people to do the task, and that removes your design bias, right?

I love the way, like this image looks. I love the beauty of this design, and it turns out nobody– nobody can figure out what to do. Like it looks beautiful, but no one has any idea how to do the job, the task. And then you have like a Google, which is the most rudimentary basic design, but everybody knows how to do it.

So that you look at task time, you look at task error rate you look at a handful of these types of metrics that give you some indication objectively whether your design is consistently achieving the result on the other end. And so, like–

[00:21:42] Bruce Bassi: Mm-hmm

[00:21:43] Ram Krishnan: that’s a series of ways we did it on a big design.

And then we actually use a user flow tracking software in our product. So we have click paths that we can aggregate and see, like, what page does everyone start on every day, at what time of day, what day of week, what type of user, it’s all aggregated and like anonymized, but it’s it gives us what the usage flows are and then what do they people tend to do in our product. There’s five ways to do things. Which one is the one most people are using? And then going back eventually to think through how do I optimize that?

It’s a tricky thing because like I was saying back with that, that like, that getting used to your operating system, one of the things, even though people like want things to be easier, they also just sometimes don’t want things to change.

So knowing that, I go ‘boop’ every day, and there’s muscle memory to that. Like when you suddenly have me just go ‘boop.’ Just need different spots.

[00:22:35] Bruce Bassi:  You’re hitting all the sensitive topics here for a clinician.

[00:22:40] Ram Krishnan: Yeah You can see how that like it is, you watch a clinician, it’s a lot of muscle memory.

[00:22:45] How do you decide what the next features will be?

[00:22:45] Bruce Bassi: Your CRM is probably filled with a lot of users, maybe giving you features, new feature requests, and I, there are a couple companies, and you’re probably familiar with this, but maybe some listeners aren’t, but you can essentially crowdsource new features where a user will post something and then other people can up vote. It kind of like a Reddit style thing where you up and down vote particular features. What is it like in the backend in the office where how do you decide upon where am I gonna steer the ship now? Like, you probably can, there’s only so much time and resources that you have as a company and you then there’s a million infinite number of different ways you can improve a product, especially given how fast things are moving.

Now with AI, how do you decide what’s important?

[00:23:26] Ram Krishnan: Yeah it’s a great question and it’s not an easy answer. I think we have multiple ways we do it. So we do similar upvoting kind of style, except that we in our crm we’ll categorize what all the feature requests are. We’ll document them, and then every time a request comes in, we’ll add a vote to it.

So that’s one input for us is that kind of classic style. It’s not necessarily opened to the entire, as a website like Reddit. But we do that internally inside of our crm.

We have a number of user groups that meet independently of us that are organized, self-organized and managed, but they offer us participation in them and they’ll aggregate requests at times.

And then we, the third area is we have an advisory board, and that is a representative sample of the market. So it’s not all gigantic multi-state, the biggest, largest practices. It spans those all the way down to a solo practice. And we have tried to get that group together at different times and get them to react to things that are happening in the industry and the market at large.

Right? So, a trend is coming in: how many of you care about this? How many of you’re worrying about this? How many of you think you’re gonna do something about it? How many of you want us to do something about it? And it’s and flip side of that is, what are your top three challenges and what do you struggle with?

And it’s important. I’ve done these a bunch in my life because what happens is you walk, you as a, some, a member, a participant will walk in with what you think is the most important thing until you hear other people bring up their most important things. And suddenly you realize that actually I’d prefer that you worked on that than.

And doing it in isolation, one-on-one, you miss the richness of the conversation, the debate, and the hashing out of what’s really valuable and important to the group at large. And for us, it profoundly reduces the chances that we make a mistake in what we build. Like we build, we spend all our time and money in the wrong thing.

And it opens our eyes to things we never considered.

[00:25:23] Bruce Bassi: That company that you used, they, you mentioned one of the first things they did, they interviewed clinicians. Did Valant incorporate any of that style or approach into gaining feedback, or is that something you might consider in the future?

[00:25:36] Ram Krishnan: That’s a wonderful follow up. I didn’t bring that up when we, I joined, I took over the company in October of 2020 and brought in some kinda experienced veterans that I’ve worked with that have done a lot of this in healthcare before. We’ve gone through a life lifetime of mistakes.

Okay? So we’ve gone through and built stuff that nobody wanted, nobody bought cause we thought we were cool and got humble and learned that building cool things is not necessarily what people will benefit from.

And one thing he put in place, our CTO, is on every like, moderately sized feature that we build, any, like that’s going not a one week job, but a multi-month like, product enhancement that we do.

Each one has to have five customers attached to it. And so five customers somewhere out there have to care about it enough that they’re willing to sign up to be part of a group that will review designs before we start writing code. And we’ll see early iterations of the product to give us feedback about whether it’s any good or not.

[00:26:31] Bruce Bassi: So that’s even before it’s rolled out to the public.

[00:26:33] Ram Krishnan: It’s before it’s rolled out–

[00:26:34] Bruce Bassi: Nice, that’s awesome.

[00:26:36] Ram Krishnan: We use an agile development approach, which is we basically do is we deliver code every two weeks to the product. We update the product every two weeks with whatever’s been complete and is a hundred percent done.

We, we up, we launched to the market. Some of these take multiple iterations. It’s really tough for me as a CEO to plan exact dates things will release because I have to give the team the grace and the space. If the customer group comes back and says, you really, I won’t use this without these three things that you didn’t think of when you did the design.

They’re critical. And that feedback is, becomes vital to us saying, okay, we then should push this out a month, incorporate their feedback because it won’t get used without that. And consequently we will always go to, we’ll, we will start with a list of. 500 things that we could possibly put on there that our team internally is really passionate about and thinks are the right thing.

Inevitably the customers will find three we didn’t think of, and they’ll take the 500 on the list and tell us, most of them we don’t need. And that also saves us time and money from building the wrong things. And it increases the probability of us being right when we take it to market.

[00:27:40] Bruce Bassi: Mm-hmm

[00:27:41] Ram Krishnan: The other thing we do is we then, after we release to market, we know we’re gonna miss stuff because the market is varied and until it’s out in the wild, people who start using it then start bringing up all kinds of marginal use cases or even obvious use cases we all missed, including the group that was with us.

And we try to leave another month to two months after the product’s released to do a fast follow up on all the feedback we’re getting from the market. So those are the ways we try to incorporate customers in our, and it’s a, it’s an evolutionary process, right? Like you get better at it over time.

[00:28:11] Bruce Bassi: Yeah that’s really interesting. I like those examples that you gave. Were there any others that you felt like would be a absolute hit and maybe it was just dud?

[00:28:21] Ram Krishnan: Yeah. I think I think it’s more sometimes like the timing of things, is item A more important than item b?

I would love to give you the inverse of that because I think that was one that, like, that we didn’t appreciate. We launched a mini CRM in our product this fall, and when we got our advisory board together, that was probably number 10 on our list, or 15.

It was pretty low on our internal list we thought was important. But from the small practice to the large practice, every single one of them brought that up as a market need that we were missing– that the market was missing.

And this is capturing just capturing perspective patients upfront in the, off your website, off of a web form, off of a third party site, bringing them into a holding area and being able to manage them effectively before turning them into patients. Either putting them in a holding pen, putting ’em in a wait list, being able to communicate with them in that while they’re in that state.

We’ve done some additional work that also takes the criteria that the patient filled out in the form and understand criteria that describes the provider and tee up the ide ideal match. So that says patient, has this, these indications is on, this insurance is looking for a male, an Asian male provider.

Just what, whatever’s both demographic and treatment type and payer type in some cases and making sure we’re matching that with the right provider in the group that would fit.

We had that way down on the list, but that’s been the one of the most widely I think anticipated and well received feature that we rolled out and spent a lot of time and money on that we would not have had at the top of our list before.

[00:30:03] Bruce Bassi: Interesting. You mentioned an example of how– you mentioned there were three operating systems, basically, windows, Mac, maybe Chrome.

[00:30:13] Ram Krishnan: Yeah.

[00:30:14] Bruce Bassi: Everyone has to just basically get used to using those. But I think the analogy is good, but the operating systems, they also have third parties that allow you to build apps and software to improve upon the user’s experience of that operating system.

And I know some EHR systems, they allow third party integration. Now does Valant allow that? And, I, this probably can be a larger conversation because it introduces a lot of other challenges when you start to allow third parties to work within your EHR system. But what decision has been made about using third party apps or software and integrating it into Valant?

[00:30:54] Ram Krishnan: Yeah, I think it’s hard to be completely restrictive. I’ll give you a, I’ll give you a spectrum here, there is the spectrum of my EHR and practice management solution do these base sets of things and everything else will just open it up for you to plug in any third party you want, so that we, if you want to do the zoom example I gave you, just use Zoom if you wanna– we’re not gonna build any of that stuff. You can plug in all these other 50 things into your product and.

[00:31:20] Bruce Bassi: There’s so many. There’s so many out there now.

[00:31:23] Ram Krishnan: You’ve solved them at a checkbox like level, but you haven’t really solved any workflow.

If anything, you’ve made workflow more challenging, you’ve made it multi-system. You’ve got data in lo like 50 screens. And so that’s an end of the spectrum. And then the other end of the spectrum is we’re completely closed box. If we don’t build it, you don’t get it. And I think we’re trying to find the balance in, in, in between there.

And we don’t have a wide open API platform that anybody can code to and do whatever they want. We have been thoughtful about the things that we’re trying to build that we inside of our product that you could acquire as a third party because we think the out outcome of workflow is better if it’s native, doing appointment reminders, doing payment collection, doing telehealth, doing measures and forms and that kind of thing.

We think those are, you can define and design all kinds of specific workflows that make your life better if they’re all well integrated. I think we’re, we’ll work our way towards a middle ground where we have some set of APIs, we have a couple APIs today for patient insertion and things like that.

We don’t get enough consistent demand on a lot of the other ones to build out a fully fledged, wide open API platform. But we’re not philosophically opposed to it. We are just gonna let the market demand drive our building of them. Does that make sense?

[00:32:38] How do you anticipate AI will integrate into EHRs?

[00:32:38] Bruce Bassi: Yeah, that makes sense. Has there been much user requests for AI integration and another probably podcast in itself? How is AI going to help assist a clinician and patient experience one of your mission statements was allowing them to connect. And I know one potential feature for AI that’s maybe a low hanging fruit is keying up a message for a clinician after they’ve seen eight hours of patients.

They probably are responding to 10 to 20 of very similar types of questions, refills, I’m sure a lot of them get categorized into a couple of buckets of basic types of questions, some of which can possibly a message can be started by an AI system and then hopefully somebody will review it and edit it and then send it out to the patient.

Maybe that can save some time, but what are your thoughts on ai, bigger picture? Maybe we can start off there and how to integrate some EHR systems.

[00:33:38] Ram Krishnan: Yeah I think I think the example you gave is a great one. It’s a low hanging food opportunity. It’s very similar to the auto narrative generation example I gave earlier when a measure comes in and a note is pre started with a automatic narrative in most of us, we often do prescription requests through a secure message of some sort.

So having that’s something we build, built ourselves with exactly what you said in mind, which is understanding what the type of request is and having templates that you can use to, to send your response to so you’re not typing it up every time. And also just making sure your inbox isn’t full of administrative questions.

If it’s a billing question, it should route differently than it does a prescription refill. Understanding that kind of thing I think is really super low hanging fruit and independent of the chat GPTs and all those announcements coming out.

I’ll give you an internal one and then I’ll give you a patient provider one.

I is, I’ve, we’ve thought about, we’re already doing the data analysis on streamlined workflows and as this technology evolves, it can write code also. And being able to feed it our workflow data on the paths and have it come back and optimize them would be an interesting internal science experiment for us, to start applying and seeing if it can come up with better objective, faster routes for people to do tasks.

And that could rapidly accelerate the amount of flexibility we can offer from client to client. It’s hard to do it when you’re writing, you have labor that’s writing code and ha you can’t create 50 permutations, but if you have code created code like that opens up a whole world of possibilities.

[00:35:11] Bruce Bassi: that really reduces the bias that you were talking about earlier. One goal is to eliminate bias of the attachment bias that you have to your product. Right.

[00:35:20] Ram Krishnan: a hundred percent. And yeah, and even attachment bias to the way people give feedback, right? They often give us feature feedback and not a problem they’re trying to solve. That’s the one thing we were always trying to get to. Okay. What problem are you trying to. I know you want faster horses, but maybe we’ll invent a car, instead of a faster horse as the old Ford example.

I think we, we’ve thought a lot about, I’ll just do like food, advanced food for thought here on where it could go. There’s so many places and it’s every day. I think we’re coming up with new theories. I saw a post yesterday of of someone using ChatGPT to play Dungeons and Dragons. Like they used it as the dragon dungeon master to get geeky if any of your audience is geeky at all. And it runs through an entire campaign with prompts, which is profoundly just interesting in that application.

In, in our world imagine like the current model the big advancement with ChatGPT, it doesn’t have custom data sets. And so I think the world over the next three years having this release to the wild will be people taking and training custom data sets using that AI technology to advance and improve something specific to the domain.

In our world, we have diagnosis codes, we have treatment plans, we have procedure codes. We have progress notes and you really have telehealth scripts. Just like we’re recording right now. This thing is converting our audio into text. So is every telehealth session ever, you could take the transcripts of those and feed a learning engine.

And the outcome from that in the, in that first mode would be back to your example of helping pre-build some of the note is being your assistant, your second set of objective eyes so you can really focus on the patient, take all that into account and pre-build the narrative for you to edit, highlighting keywords, maybe things in the second and third wave of this, incorporating the actual audio for tone, for inflection especially if you can compare it to prior.

Sessions where tone and inflection work a certain way odd with video it would be expressions changing. And then the most invasive and kind of scary model would be a little agent sitting on your device saying like, you have, your client hasn’t been mobile for two weeks, like they’ve been in the same room, right?

Like you, you can just, if you add all this up, you can imagine that either we will have 10 times as many data inputs to do something with, or the machine will assist us with all that. So we can be more present and empathetic with the client in front of us while the machine gathers all the insights and presents us with the analysis.

And then you have quantitative and qualitative coming together to finish your your documenting your session.

[00:38:02] What will be barriers to incorporating AI?

[00:38:02] Bruce Bassi: It’s an interesting perspective, idea, of incorporating ai, but I think what’s gonna happen goes back to what you said earlier. I think some of the resistance is going to be from clinicians, basically resisting change. And I think that’s gonna be the biggest barrier because I, I think actually a lot of patients do like new ideas of how to improve efficient care.

But it reminds me of how prior to Covid, one of the biggest barriers that I felt as a clinician was not only like intraprofessional stigma, they always felt like a telehealth doctor was just kind of, in the wild, wild west, they didn’t, they weren’t as thorough perhaps. And then the other barrier was payer based, but a lot of patients actually wanted it. They felt like it was more convenient.

I feel like we’re gonna go down a pretty similar path with AI, where maybe people like you and me will be interested in it, but to get the mass of clinicians on board with something, listening to your conversation and writing your note for you and like taking it the next step, I feel like will be a really slow endeavor.

But I think we have to start with that low hanging fruit probably. What do you think?

[00:39:11] Ram Krishnan: Yeah I agree and I’ll maybe give you another flavor of this outside of mental health because, some of that slow endeavor has been, had started 20 years ago. And I’ll give you a couple examples there. I’m a, an advisor in an AI company in radiology for breast imaging. And this isn’t new technology. Some of that stuff we called clinical decision support in the early days and then called it big data and then machine learning. And today we call it ai and then now it’s AGI, which is actually really is a different thing, the degenerative learning. But we have, on the breast imaging side to, to speak to the resistance cuz you’re spot on like the breast imaging.

AI isn’t visually looking at images, it’s looking at the pixel and voxel data to find patterns over time. So it’s not even using eyes, right? It’s just looking at the ones and zeros and finding patterns. And it has a higher sensitivity and specificity than the human does, looking at the same study, but it’s still been resisted and hasn’t been adopted or approved for widespread use in a way that’s been meaningful yet, even though that’s been around for a while. And it’s a lot of the same resistance that you bring up.

And then one step even further back was clinical decision support, where we just studied data and patterns. And and, I was part of a group that studied this at a number of hospitals and we put together principles for standardizing just administrative practices: should you order a an elective C-section for a patient under 37 weeks, the morbidity rate goes up substantially than if you do it after 37 weeks. Really see this most simplest of decision support, like ideals, right? In an ordering. And that had massive resistance because every market and every facility was different.

They had a different patient population, different providers, different, there’s a lot of the, there’s a lot of the art of healthcare that, that I think pushes back against some of this on that end. So I think you’re a hundred percent right. It’s up to us to find the low-hanging fruit that we can apply all over the place.

I will say radiology is another great, like, place to look at some of this. They’ve been dictating their reports through a voice recognition system that has been applying AI to that that text to speech for a long time. And they’ve done some, I think, really powerful things where you dictate the report, the report comes up and the thing, the critical findings are all highlighted. So if something you’ve said in the report dictation indicates something that needs to be, then– it’s essential, it wasn’t part of your study, it wasn’t what you were looking at, but needs to be communicated back to the referring physician that’s highlighted and a decision point has to be made there.

I think there’s a lot of that technology that exists in other disciplines of healthcare that has been testing this for years that might accelerate bringing that into our market as well. And so that that’s maybe another example or twist there that I think gives me hope that some of this can be applied here because we’ll have other proof points to point to within healthcare.

I love finding the ways to automate as much of the work that people don’t value doing first because that’s the easiest thing to get people to buy into with respect to change.

[00:42:05] Bruce Bassi: Yep. Yeah, it’s interesting watching the market move. In regard to AI, I always feel. There’s two camps: there’s like people who are really judicious and cautious about moving forward, and then there’s the startup companies that wanna be disruptors and they just go and do it anyway.

Maybe patients shouldn’t be using them directly, but there’s like basically AI out there for treating depression, doing mindful meditation, like you think of it, there’s probably somebody who’s already creating an app out there for it.

So it’s like, why not? Maybe we can get some of those judicious people to start the ball rolling because they’re gonna do it cautiously and maybe they can come up with ground rules and start to help build trust there between that company and the patient and the clinicians as well. And hopefully move the needle forward a little bit at a time and sit back while some of those startups go under perhaps.

[00:43:04] Ram Krishnan: Yeah. And I think what’s what’s neat about the ecosystem is to your point, people will, will just plow forward with the idea, irrespective of whether or not anyone’s gonna be able to adopt them and or will take them, right? And they’ll flare up into the sky, get lots of media coverage, and then they’ll fizzle back down just as quickly as they, they went up.

But in there are always nuggets of innovation that come out of those that I think tell us things that if we’re listening, particularly those myself and my peers running EHR companies and such we can make, instead of fearing those technologies and instead of your practices fearing those models. Like you’re doing great, plow forward. We’re doing great work, we plow forward. And we try to think through how to extract the best outta that and bring it to our providers. And so like you have the Calm and the Headspace of the world, exactly to your point, there’s a thousand mindfulness apps out there and there’s, mindfulness is a core part of our industry, right?

Mindfulness and and meditation and a lot of, like all the neuroscience kind of apps out there, there are elements of those that are really valuable to and a core part of what providers are deliver, delivering in a treatment plan to their patients. We are sitting there with the EHR between the patient and the provider.

And I think as a trusted relationship, bringing those tools into that relationship could be powerful. When you discharge a patient, giving them a prescribed mindfulness app to continue on in some form of practice afterwards could be a way of ensuring that after you’ve driven change, you’ve given them tools to keep them in control.

And that would be, to me, a very logical and powerful extraction of that, that technology set and that content set, and now integrating it into the care path of what you do today.

And, for us, we, we launched an app with that in mind, which is that, we think there is lots of neat things that have been developed, there’s a way to bring that directly to your practice and connect, keep you connected with your patients during the course of care, but also after care to make sure that they keep on it. You know what I mean?

[00:45:08] Bruce Bassi: Mm-hmm. You mentioned clinical decision support in the context of that radiology example. Was there anything that Valant is thinking about incorporating for decision support making for psychiatrists in the EHR system? Because that’s a whole nother industry and of itself. It could be with prescribing, could be with the note itself, could be with the messages. What are your thoughts on that?

[00:45:30] Ram Krishnan: I think some of that stuff the problem set that I articulated before in the hospital setting is even more profound in the outpatient setting. So if it’s difficult to get five hospital systems to agree on a set of standards, getting, 5,000 independent psychiatrists to agree on a set of standards, is that problem just exploded tenfold.

So I think like the potentials there the change management problem has just grown exponentially. That’s my 2 cents on it.

[00:45:59] Bruce Bassi: Yeah.

[00:45:59] Ram Krishnan: I don’t know. What do you think?

[00:46:00] Bruce Bassi: I, so the EHR systems that I’ve worked with, I feel like part of it is just staying, having that competitive edge. You need to do something that shows that, Hey I get my finger on the beat here. AI is where things are moving. We hear you. And we want to try to start to incorporate that, even if it’s one feature or something or other.

And I think that could be very attractive thing for certain people, but I it gets back to how heterogeneous the customer base is. Can certainly see people not caring, some people being like, oh, that’s weird, and some people being like, oh, that’s really neat. I want to, I’m interested in that.

So hard.

[00:46:41] Ram Krishnan: I think our focus will be in, in the areas where we can eliminate repetitive and un like non-value added work as best as we can. And being an aid and an assistant in all the documentation, this face-to-face you and I are doing it just gets crushed if my head’s down typing in any way, shape, or form, you know what I mean?

And then the end of your day, having to do that any of that week and we can work on, I think that’s gotta be our primary area of application of all of this.

[00:47:10] Bruce Bassi: Whats that Nvidia I think they have that new AI where it moves your eyeball to the Have you seen that? And that just probably eliminated all of those companies that have the camera dropdowns.

[00:47:24] Ram Krishnan: Yeah.

[00:47:25] Bruce Bassi: But now it’s like almost too creepy. So that’s a really good point of how like, AI went way too far, and now it’s just like staring at the camera nonstop.

[00:47:33] Ram Krishnan: Yeah. It’s a wild world we’re in, man. I and I, it’s, so I’ve spent a, I left healthcare for a little bit and spent a little bit time in the gaming industry as well, and there are a lot of if you just went a different direction with this, there’s a lot of, I think, value in, in understanding so much psych, like, like neuroscience and psychological, like understanding goes into game design because you’re nudging behaviors constantly.

It’s all nudging behaviors and there are really powerful elements of there. We bring ai, but I think we’ll bring some of that back into our tool set also. Right? Like in, in the language you choose for onboarding clients and getting them to do tasks, the what’s the, you they’re so, it’s so mechanical right now and software driven instead of empathetic with with the way a game would work in terms of how they talk to you and kind of take care of you through the journey? We don’t, we could, we, there’s a lot for us to gamify in the experience. Like, like you’re working your way up the ladder of completing the tasks that you’ve been asked to do and you’re giving positive reinforcement for it.

That’s simply like using like common practices that exist to get people to feel motivated to, to do the tasks before them. And I think our patient population more than any, would benefit from that along the journey. And I think bringing those tools in adds as much sort of different market value into our space.

Like back to your kind of marrying what’s cool and hot with what’s really delivering value. And what’s cool and hot from like a label perspective, but really isn’t doing, at the end of the day, much more than another bullet you can throw up on your website or your talk track.

I think that’s where I think like the next iteration of value-based care, if you actually did sign value-based contracts, we would value the adherence to the treatment plan. We’d value the adherence the motivating people to do the things that we’re asking ’em to do along a care care path with a patient because you’d get some point in time, a real value-based care contract would pay you a fixed amount, and you’d want all these tools to ensure that the treatment was being adhered to.

And I think that’s, if we’re skating where the puck will go at some point, that is gonna be really important. And it’s one we have our eye on to make sure we get ahead of it for all of y’all.

[00:49:46] Bruce Bassi: That was my final question that I wanted to leave you with was where do you see things going in 50 years from now? Do you feel like the number of VHR systems will continue upward? I feel like we gotta reach a critical mass at some point, right? There can’t be hundreds of thousands of EHR systems in 50 years.

Do you think they’re gonna consolidate into a few big players?

[00:50:07] Ram Krishnan: I I’m just thinking through through 50 years from now, what was 50 years ago? 1966, 68? I guess. 1968 was 50 years ago. And think about sitting at

[00:50:18] Bruce Bassi: how far we’ve come.

[00:50:20] Ram Krishnan: and looking ahead that’s a, yeah, it’s a big distance between the two. I can’t even foresee what will happen in 50 years. First of all, I’ll be dead.

And second of all, in 50 years, I don’t, for, for the way things are advancing with deep Link and Deep Mind and just being able to connect directly I, all I know is it’s gonna be an exciting, crazy place and the people who are leading the way won’t be the folks that are doing it right now, right?

Like, I know we look about a year out to two years out, and that’s as far as we can without like the world changing rapidly 50 years from now, I just hope, I hope mental health is is health. It’s not, it’s not distinguished from physical health and it’s one, one holistic view in that we’ve harnessed and wrestled technology in general in a way that is net positive for humanity. And not net negative in that the solutions and software that come along to support it are hyper-focused on, making their lives better. And that’s pretty high, like visiony set of answers. But I think that’s the best I can hope for.

[00:51:27] Bruce Bassi: I like the way you put that. I like that you’ve mentioned that mental health should just be health. That’s a great point. And I feel like maybe 50 years from now, EHR should be more patient-centric, where an EHR is basically your EHR as a patient and you go to a clinician or another clinician and they basically have a way of inputting data into your EHR.

So that way you can keep it and you can store it and you can understand it and have more value rather than, hey, saying, Hey, can I get my data from that place and bring it over here? And are you gonna do it on time? Is it gonna be a nice easy path over, but it, so it would be nice if there was a little bit more universality of how they all speak to each other.

And I could probably speak to that as a clinician too. Patients come from one person to another and I’m basically re-collecting a lot of work that a lot of other doctors probably already did. How much waste is there with doing that? That would be something that needs to be improved upon, but hopefully down the road.

[00:52:29] Ram Krishnan: I love that view and I think it’s it’s to take that two steps further. Not only that, all my biometrics will probably be collected in advance too. Right. So you’re not collecting all of the vitals and you’re not analyzing any of it. It’s all like kind of pre telling you exactly. And in many ways, hopefully we’re taking care of our own health because it’s so easy because your, to your point, everything is patient centered in that capacity.

I like both views. I think they, they are totally complimentary.

[00:52:55] Bruce Bassi: I appreciate this conversation. It was really fun. Fascinating. I really like your viewpoint and the principles and philosophies that you have within Valant. I can tell when you mentioned, thinking about how words are used on the EHR system and thinking about user behaviors is just a level of understanding of the user that I feel like we need right now in, in EHR design moving forward.

So I really appreciate your insights today and thank you so much.

[00:53:22] Ram Krishnan: I appreciate the chance to be on the show and to talk about it. And it’s been great. You’ve been a lot of fun to talk with. And love to do it again someday.

Resources

To learn more about Valant EHR:

https://www.valant.io/

TAGGED UNDER: ai | EHR | valant
Are You a Journalist Writing About This Topic?
Are You a Journalist Writing About This Topic?

If you are a journalist writing about this subject, do get in touch – we can provide a comment for you.

Pin It on Pinterest

Share This