In this episode, we talk with Trey Holterman, Co-Founder and CEO of Tennr. Trey shares how their automation platform maps endless complexities throughout the patient journey helping infusion centers process patients instantly and get reimbursed for the work they do.
Transcript: How to Improve Efficiency, Reimbursement & Service with Trey Holterman
Amanda Brummitt: WeInfuse Podcast, Episode Number 69. Welcome to the WeInfuse Podcast. My name is Amanda Brummitt. In every episode, we give you a seat at the table as we talk to infusion center owners, operators, and experts so you can get the insight you need to run a thriving practice. In this episode, we talk with Trey Holterman, Co-founder and CEO of Tennr.
Trey shares how their automation platform maps endless complexities throughout the patient journey, helping infusion centers process patients instantly and get reimbursed for the work they do.
All right. Well, Trey, thank you so much for being here with us today. We’re super excited to learn your story and to hear about Tennr.
Trey Holterman: I’m excited to be on. Thanks for having me, Amanda.
Amanda Brummitt: Yeah, of course. Well, before we get started, can you just walk us through your background? Who are you and what should our listeners know about you?
Trey Holterman: Totally. So I’m actually, I’m an engineer by trade. We started the company actually over four years ago–the company Tennr–out of the Stanford AI Lab.
So I was a technical engineer and I was really just trying to solve a problem that my mom was dealing with. She’s in family medicine, and it was basically, she’d call it the black hole imaging center problems. Always complaining in our family group chat, where, she’d send her patients out, right? She’d really call them “my patients”, right? “I’m sending them my patients.”
And it felt like she was sending them into a black hole, basically. It would just take days for them to contact her back, contact the patients, make sure they had everything. If the patients showed up, half the time, their information really wasn’t there.
So obviously very like a technical engineering and research background. And then my sort of foray into this whole world of trying to solve the pre-visit patient processing problem is from that sort of experience there. Summer, like six years ago now.
Yeah.
Amanda Brummitt: Okay. I love how you just casually and humbly, barely mentioned Stanford AI lab. Like I’d have a sweatshirt that said that if I had gotten to go to school there.
Trey Holterman: Well, to be honest, the reason I’m sort of all shucks about it now is I don’t even code anymore. The real thing.
Yeah, the joke that I make is like the only useful thing I did was I was really a TA–teachers assistant. And so I got to see who was really good and rescue a bunch of people from going to work at like Facebook and TikTok and Palantir–Palantir is a little bit more interesting–and just got them to work here, to be honest.
And that’s why we’re based in New York. So that was sort of like my claim to fame: was like, unfortunately I’m not really working on any cool technology myself, but I’ve recruited the first 10 engineers that really are doing great work. That’s my calling card now, I guess.
Amanda Brummitt: Okay, makes sense. Surround yourself with brilliant people. Yeah. Okay, so since you’ve now told us you’re not a coder anymore, what is your main area of expertise? And what do you think you’re really good at that we can learn from today?
Trey Holterman: Yeah, well, you’ll never–once an engineer, always an engineer.
And I think that’s where, one of the things is, we never say, like, “AI”. We always talk about the models that we’ve built. So the machine learning models that we’ve built. And the reason that is, is because I feel like what I’ve become, I feel like I’m now an expert at–it took me years to really be able to bridge the gap between, hey, what are the models that people are working on and what are we building? And then how does it relate to getting a patient from point A to point B with a clean experience and reduced DSO time?
It’s gonna result in a clean claim. So in the infusions world, we call this script a seat where we’re measuring DSO. I think they now, just frankly because of our customers and really working closely with them, have learned so much of that business side of, what does it actually take to be a lean, really efficient company that can get patients in super fast.
Your referring providers love you, your patients love you, and you’re super efficient, and your headcount isn’t growing linearly. And I think because I understand the tech and I can speak to a number of ML researchers that want to work in a tech company in New York, that’s what I have been able to do.
So that’s it.
Amanda Brummitt: Yeah, that’s beautiful. We need more people like that. And you mentioned DSO, Days Sales Outstanding. Can you explain that?
Trey Holterman: Yeah. So DSO is Days Sales Outstanding. It’s basically how long from the time to receiving information about a patient, typically via fax in our world, to the time that you have a clean bill coming in being processed on the other end.
Of course, there’s also script to seat. So how long does it take to get the patient actually showing up in your system? And of course, there’s also time to first contact. So how quickly, how long does it take you to actually contact that patient?
So those are some of the key metrics that we’re always trying to measure for customers and make sure they’re hitting them really well.
Amanda Brummitt: Yeah, I love to see you applying those in a clinical setting. That’s awesome.
Okay, let’s go back just a little bit. So you’re a coder, you’re an engineer, you’re doing all this stuff. When do you take the jump? When do you say, okay, I’ve heard about this problem from your mom. When do you say, let me go fix this problem in healthcare and I’ll do it from scratch.
Trey Holterman: It’s this horrible story of it just took so frickin’ long basically and we really have been at this for kind of years, and it wasn’t like we were doing nothing.
What we were typically doing was basically, so I go to that imaging center, right? That’s like the cause of a lot of the pain. And imaging centers in infusion, you feel the pain and experience it very similarly.
And I kind of, I don’t bang their door down, but I’m like, “Hey, what’s going on here?” Right. And it would be like if I showed up at front door of an infusion center, it’s like, “Hey, why are you guys taking 12 days on average to get this patient in?”
And what they said was they’re like, “Oh, you’re from family Pacific,”–the family practice that I was representing. And they’re like, “Well, look at the crap that your mom sends us. It’s like 30 page faxes. Sometimes it’s got all the information. We’re supposed to do the pre-auth and I’ve got to call these insurance companies.”
And so we looked at that problem and we said, “Well, listen, we can automate basically everything downstream. The hardest part is going to be reading those documents, knowing when there’s 10 patients on one document, 10 documents with one patient. Doing things like reading check boxes, really the kind of reasoning side of, given, I don’t know, Ruxians. Do we have all the clinical criteria of medical necessity?”
And so we went and built out all these models, basically spent a year and a half building, and we come back to the imaging center and we did the horrible thing that you’re not supposed to do when you’re starting a company. Which is, we just built and we didn’t really talk to anybody.
And we came back and we’re like, “So imaging center, where’s your IT guy? He’ll connect to…our APIs and we’ll have automated this process, right?
And they’re like, “What IT guy?” Right. God. So then we’re like, “All right, we have to go build this whole platform.”
And then they’re like, “Oh wait, you’re not integrated into this, archaic RIS system from the 1980s.”
And obviously this is a WeInfuse Podcast and the WeInfuse integration has been awesome. So luckily that hasn’t been a nightmare.
So then we had to build this whole integrations engine. So what started as, we thought a fun research project that we were getting school credit for, became this like, whole platform and workflow builder and a million integrations. And that’s where some of that–talk about black holes–some of that time went.
And then it was really at the middle of last year when our first case study started coming out and people realizing like, “Oh wow, these guys really do reduce that time to first touch.” The patients are coming in.
And frankly, it was also once people started realizing, we’re really able to decouple our headcount from our revenue. That was when, frankly, everybody started telling their friends and we were able to grow very quickly, just naturally by word of mouth.
Amanda Brummitt: Yeah. Yeah. I bet.
You mentioned the WeInfuse integration. How does that look for a practice that is on the WeInfuse platform versus like, let’s say, how does it work for an infusion center that isn’t?
Trey Holterman: Totally. WeInfuse is just made like–we’ve worked super, super closely with their team to build what we call a native integration.
So there’s a way to integrate with a platform where it’s not really native or they’re not really on the inside with you, and WeInfuse is just been on the inside and we’ve been able to build it really natively.
What does that mean? So it means a lot of these documents come in or get scanned in or dropped right into WeInfuse. And to really leverage all of the automations of Tennr, the ability to read those documents, we have to pull a lot of those documents from WeInfuse, say “what’s all the information”, and then go in and actually create orders, create patient objects, and do what we need to do…update charts within WeInfuse.
That’s one scenario. That’s why the integration is so important. We’re pulling those documents from, actually, the WeInfuse Q. We’re able to update those things automatically.
However, a lot of people, especially if they’re acquiring like crazy, there are 20 different fax systems, 20 different e-fax systems.
One of, I think, the awesome things about our integration into WeInfuse is that rather than people having to migrate 20 fax lines–because we’ve built so many integrations into so many different e-fax systems–a lot of times people just integrate their e-faxes into us. It might be 10 different portals, or 10 different e-fax lines, into Tennr and then we’re able to once again, get all that information into WeInfuse where it can live for the clinical staff and the billing staff that really needs that information in there cleanly.
Amanda Brummitt: Nice. Would you say–obviously it’s saving practices money, saving them time–what’s that metric? What do you guys use to say, “Yes, this one was a success”?
Trey Holterman: This is where I’m going to get really annoying and opinionated because they’re like, obviously you have to solve a problem for an individual center or an individual practice. And I’ll talk about that problem, those problems.
But really, we look at the problem as a whole, and the problem as a whole is what we call the patient processing problem. It’s what leads to all of these very downstream, costly situations for an individual practice. So what does that mean?
It basically means, if you view the whole healthcare network as a network or a graph, referring providers should always feel like they can send to recipients of patients. And those recipients are the best possible place to send it. Meaning they’re super on the ball. They’re super fast. They’re able to get all the information they need the first time and they make it as easy as possible for the sender.
Unfortunately that doesn’t always happen and you end up with situations where referring providers, nobody’s good. And so they send them to different centers, to different practices.
And the problem we’re trying to solve is how do you make it so that the recipient of that information can reduce the amount of time, reduce the amount of friction that it takes to get a patient from point A to point B. That’s the fundamental problem.
And what we always include in any account reviews or anything like that is, how long did it take the patients to go from the time we received information about them to the time that they’re getting treated, oftentimes with life-saving drugs?
That’s like the fundamental problem. It’s the problem that many employees at our company have experienced firsthand or seen. It’s the thing that I saw.
Let’s talk about the symptoms. That’s what gets operators really excited, the symptoms of that problem.
Well, the first sort of symptom is the baseline. Again, if we think about Maslow’s hierarchy of needs. Always across the board, you have efficiency struggles, right? You just have to keep throwing more people at it because you know if those referring providers send in patients and they’re not able to be processed, they’re going to be pissed. And so you have to keep throwing people at it. You keep growing and it feels great, but how much is really being added to the bottom line if you keep, obviously, throwing more people. Efficiency is always the first thing that we measure. I talked about those case studies, it was taking that national imaging center from 70 down to seven that then people were like, “Holy crap, okay, this is really a step function change.”
The next thing is really reimbursement. A lot of people realize, “Hey, what we feel is like billing mistakes downstream. And I go yell at my RCM or my billing team, they’re really mistakes that I can fix upstream where the moment that I get a patient, I can make sure that I have everything I can actually, what we call “qualify an order,” so that I don’t go back and forth to the payer three times with an authorization. I can make sure I have everything the first time.” That’s a huge thing. So we don’t buy and bill and then we’re down 28K on a drug, it’s going to take us 10 patients to go pay back.
Amanda Brummitt: Right.
Trey Holterman: So efficiency, reimbursement, and then, of course, from there it’s all about service.
When patients, where, if they feel like they’re getting into a black hole, they’re going to tell their referring provider? Provider is going to send patients to the other competitor down the street. And then of course, when you have great service, you’re able to grow and then you get good analytics.
So I have this whole pyramid, but this is a podcast, so I won’t show it. But we’ve never not done the efficiency piece. We’ve never not done the reimbursement piece, the capitalizing, turning it into great service, like texting a patient same-day–that’s more about people’s willingness to really rally around a nice business process change with the new technology.
Amanda Brummitt: Yeah. Amazing. Infusion centers are always looking for how to grow. And for me I’d say take 10 steps back and do exactly what you’re talking about, Trey. Let’s make sure everything is super smooth and yeah, like family medicine doctors, they don’t have time to track down their patients and see where they are in the process. They just need to know they’re taken care of.
Trey Holterman: And frankly, they don’t have that many data points. And so you have to be so good because they get yelled at by some patients like, “Hey, that infusion center never contacted me.”
When the competitor down the street comes in with cookies, it’s just a fax. It’s literally just a fax number they have to change just to see if they’re any better.
And so you’re constantly competing on service. You’re constantlyz-and unfortunately, if you’re a really well run business using the right tech, it’s like a huge advantage. You get to say, one of the big things–everybody downloads these charts now from us where it’s like, “Hey, look how quickly we were able to text your patients and contact them and move them through our system because when sales reps show up with that stuff at the desk and it’s not just cookies, it’s like proof of look how good we are.”
It buys so much more rapport so that one angry patient comes in and you’re like, well, that can’t be the case because look, I can see that actually they contacted you four times–and anyway, so that’s the fundamental service component that ends up being really differentiating for these businesses.
Amanda Brummitt: Yeah. So what’s been your biggest challenge, would you say, in the infusion space? Or even, what are you still trying to overcome?
Trey Holterman: Yeah, well, I think you have like a fundamentally–we’re in sort of 18 different specialties now, subspecialties, and we’re processing over 10 million patients a year.
And infusions, I love infusions because infusions really understands the element of service, the element of speed, the element of efficiency. But, infusions also…if you imagine levers to these things–if you’re in infusion care you’re going to care a lot more about the accuracy. You’re going to want to make sure that the right payer guidelines are being adhered to. You’re going to care a lot more about accuracy, probably more than maybe, let’s call it an ophthalmologist is going to care more about speed, right?
That ophthalmologist really is just doing a ton of cataract surgeries. They know it’s very easy to qualify a patient for that. And I think one of our challenges, honestly, Amanda, is how do you make sure that you are always building the right functionality for the right–even though it’s like all the same problems, the patient processing problem–how do you make sure that you’re really able to nail the functionality that an infusion center cares about, that doesn’t actually really want to automate, they just really want to augment. And so that’s why we have these human in the loop functionalities where people can see everything that’s being done.
But that’s not a trivial challenge. That’s the scale that we’re running at now. So, definitely is the first one that comes to mind to be honest. And by the way, also, we have the exact–irony of it is, I always joke, we have the same problem that we solve our customers for, which is that there’s an end state. We want to service, we want to treat these underlying symptoms that our customers experience, and we want to do it as quickly as possible, as fast as possible. But we have to do all the information gathering, right? I wish we could be like a one click checkout, like now your problem is solved. Unfortunately we’re just like a center in the sense that we’re not, you have to make sure you have information gathered.
That sort of slowness to me is the bane of my existence, because I feel like we’ve already solved the problem and now it’s just about getting in front of enough people to tell them about it and show them how it works.
Amanda Brummitt: Yeah, yeah, most definitely. So what about light bulb moments? Have you had any, whoa, I was not expecting this, but…?
Trey Holterman: The checkbox and signature detection models.
Amanda Brummitt: Really?
Trey Holterman: This is where I will geek out. And this will not be, yeah, this will not be that interesting I think, but like, you would just be amazed at how signature detection is this insanely hard ML problem. And that’s where on the engineering side, we always try to say, we don’t really hire for the best altruists, but we’re usually hiring for the best technologists because you’ll run into some of these problems that simply just haven’t been solved before, like checkboxes.
The best accuracy that we were able to see was…about 70 percent on checkboxes. And we couldn’t really figure out why. And it’s because financial services is a big piece of the training data for a lot of OCR models. Whereas in healthcare, checkboxes is kind of the whole thing. But since so little documentation has been automated within healthcare, you really don’t have a good data set.
And fortunately we realized, well, we have one of the biggest data sets that were like 18 billion checkbox pairs. And once I realized, okay, we’re actually able to do really novel things that everybody would agree, like, computers should just be better at reading checkboxes. It was very like…it was like a…it sort of shattered my reality to think that there’s all these people in the world, all these great researchers, and nobody had taken the time to just build a really good checkbox that worked.
Amanda Brummitt: Right.
Trey Holterman: And then the same thing happened again with signature detection. And I think it’s because those are two very healthcare-specific, and we only work with practices and providers. So I think that made it very logical for us to dump the resources into it.
But I think just two things like that.
If we didn’t have the ML talent, it would have just stopped us in our tracks. All of it would basically be for nothing because you wouldn’t be able to read checkboxes. And so, I don’t know. I will scream from the mountaintops of the importance of being able to read checkboxes and detection of signatures.
Amanda Brummitt: Yeah. That’s shocking.
Trey Holterman: That was a weird “aha” moment.
Amanda Brummitt: Yeah. Those seem like the simplest things that you’re doing. So fascinating.
Trey Holterman: Totally. And luckily we have this really close relationship with Amazon Web Services and Azure, both because we share some cloud computing with them, and spend enough money and they suddenly give you a lot of attention.
And yeah, we try to like really understand. We even met with their teams like, why have any of none of the off the shelf models solved for this? And they explained the data set problem.
And anyways, I’m sorry. This has got on a weird technical tangent, but it really was the most like, whoa, you need to be able to build cool things.
Amanda Brummitt: I expected you to just say physician handwriting.
Trey Holterman: Yeah, no…we joke cause my dad got his thesis in reading handwriting in the eighties, OCR. It was literally a research project at MIT at the time. It hasn’t actually gotten any better, frankly. In the last 40 years, you’ve seen very marginal improvements in actually reading handwriting, but it’s very good, very, very good at reading handwriting.
We haven’t had to improve anything on the handwriting extraction or image extraction. Frankly, it’s all just about being able to discern if clinical criteria are being met based on a drug and based on an order and based on a payer selection.
That’s, gosh, sorry. Yeah. Another thing that you would think that selecting a payer is easy and then you actually study it and you realize that selecting the right payer is frankly, one of the number one things slowing down the health care system. It’s what leads to so many denials, it’s what leads to so many angry providers–is the fact that nobody can agree on the best way to identify a pair: based on PO box, or is it little things on the card? If it’s–
Anyways, I can nerd out about those kind of weird little nuances because I think they’re weirdly just bottlenecks to good solutions.
Amanda Brummitt: Right. Well, thank you for nerding out on them so that our infusion centers don’t have to. So, you definitely–people can’t see your face, but you’ve got a palatable energy about this. You can tell you’re excited about it.
What are you most excited about?
Trey Holterman: I am like, I am obsessed with good solutions, I would say….Obviously I even struggle to appeal to the…obviously it really matters that we build great products for service for patients.
But more than anything, I think there oftentimes exists the right way to do things. And what I love about the infusion space and why I’ve taken such an obsession to it is that it, it is the–not for any actually familiar relationships or anything like that–
It is the right solution. You should not, as a patient, in my opinion–I know we’re largely talking to ambulatory infusion centers–you shouldn’t have to go into a hospital to go be administered a drug. It should be this lovely retail experience, in my opinion, where you’re really taking care of and you’re really doted on.
And I really think that specialization, meaning if somebody runs an infusion center, all they care about is that experience of walking through the door and getting a life-saving drug and being cared for, you’re going to do so much better of a job.
I really believe in the fragmentation of services like this.
And fragmentation almost sounds like a bad word. It’s actually a really good thing. It means we’re letting specialists specialize and it means we’re letting people build the optimal solution to this problem.
I think just the excitement, honestly, within infusions is palpable. There’s growth left and right. People really understand that it’s the right solutions, the right patient experience, it’s much more cost effective than sitting in a hospital.
I think it it’s also spurring great pharmaceutical innovation. Once again, another thing that like gets a bad rap–
Obviously people, rag on pharma all the time. And yet when they’re actually building really key drugs and this new sort of vehicle and medium that’s really being ushered in and actually a good way to deliver life saving drugs–
I don’t know. It’s hard not to get excited about that. Like that is the right solution. That is like one more sort of patient visit where people aren’t going to think, “Man, healthcare is broken.”
They’re going to think, “Wow. First of all, I’m impressed. They texted me 30 minutes out of stepping outside,”–stepping out of the referring practice, they’re not going to call it that. “And then to think, wow, I got scheduled in four days later. That was really fast. They were able to get that insurance process through and now I step into this really nice retail experience.”
I don’t know. That just sounds right to me. And you can always tell [when the] whole engineering team shares some itch in the brain. And I think for good solutions–I probably won’t say on this podcast–but we all get really satisfied by those types of things. Seeing it at a macro level, the way infusions is, it’s just exciting.
Amanda Brummitt: Yeah, most definitely. So Trey, as a business owner, as somebody actively helping and improving the infusion space, what’s one last piece of advice you would give our listeners?
Trey Holterman: There is a very clear–
We only do one thing, right? And it’s, how do you solve the pre visit experience?
And it just so happens to solve a bunch of things downstream, billing, DSO, and of course, time to first touch, script to seat, what have you.
But the number of practices that we speak to where I feel like people have…they’re super excited about their MMA, they’re super excited about the clinical team they have, and they’ve…started to ignore the blocking and tackling of being a really great operator and making sure that the way that information and patients flow into their system is really well optimized.
There’s been such an advance in technology that I would really say, really study the process, really take a good hard look at, watch over the shoulder of your most junior data entry team, how your RNs are spending their time, how your clinical team is spending their adamant time, because we do that, right?
Honestly, as part of our onboarding, I forced the CEOs of these big companies to, “No, you’re going to watch and you’re going to see what your team does now because we’re going to show you what they do in three months.” And I would just say it’s just blocking and tackling and…just really understand your processes, understand where those bottlenecks are, understand where those weaknesses are.
I still have to do the same thing, right? I’m constantly trying to figure out, why does it take so long to get somebody signed on to get them implemented, get them integrated in a given case, right? And I’m always measuring these things.
But especially when the business is about service, especially when it’s about providing great care, not getting lost in driving that patient experience from the moment that they’re not just a blob on a piece of paper. It seems very hard to fail if you do a very good job of that.
If every referring provider knows if I send patients to Amanda, she’s going to be white on rice all over it. I just don’t understand. And if you’re very accurate, right? And you’re not making mistake, clerical errors that are gonna cost you tens of thousands of dollars per patient, it’s very hard to fail.
I think it’s very hard to fail is what we’ve seen. And those types of businesses grow very fast. When I talk to the operators and they really understand their own processes or they have a really tight loop with who they’ve delegated those processes to, man, they seem to just be running on a different level.
So I guess I don’t really have much advice, honestly, outside of the small domain of expertise that we’ve curried over here, but that would obviously be it. I’m sure there’s much more important things, though, that people can spend their time doing. But, uh, that’s mine.
Amanda Brummitt: Yeah, the piece that jumps out at me is you mentioned when people are further along down the road, once they’re already busy, having that CEO take half a day and go sit with their intake people, once they’re already busy, that’s huge because you can identify so many things that are inefficient. You can identify people are doing a great job and praise them and I think that’s brilliant. Yeah.
Trey Holterman: I’m doing an hour of data entry every Sunday. It’s like mandatory, because you can’t lose. I’m going to go to a call center–as I was in a big call center and working it–
Like if you as soon as you forget what is actually being done, it just seems really hard to make good decisions and know where pain really is coming from.
And I don’t know, all these things sound so basic, right, Amanda, when you say them out loud. But then it’s like, it’s the weekend and it’s like the last thing you want to go do. But that’s how everything is, right?
Amanda Brummitt: Yeah. Well, Trey, thank you so much for your passion for the industry. Thank you for all the information. And we appreciate you making time to be on the podcast.
Trey Holterman: Oh I appreciate it too. And what WeInfuse, obviously has done for the industry, we’re hoping to catch up in the next 10 years or do something as notable, but it’s really, really great to be here.
And thank you, Amanda for your time.
Amanda Brummitt: Wow. That was really great information from Trey Holterman of Tennr about how they’re helping infusion centers improve efficiency, reimbursement, and service using automation and machine learning. If you’re already using the WeInfuse software platform, be sure to ask about the Tennr integration.
And if you aren’t familiar with WeInfuse and RxToolKit’s web-based resources, I encourage you to schedule a test drive. These tools can save you time and money in your practice while making infusions safer for patients and caregivers.
My name is Amanda Brummitt, and we’ll catch you in the next episode.
Guest Speaker:
Tennr Co-Founder and CEO Trey Holterman.
