Designing the AI Native Clinic with Dominik Middelmann, co-founder and CEO of mdhub

[00:00:01] Welcome to Expanding Access, a podcast by Behavioral Health Tech, where we explore the cutting edge of mental health, substance use, and autism and IDD care. Each episode features insightful conversations with industry leaders who are driving real change through technology and innovation. Whether you're a provider, payer, or just passionate about expanding access to behavioral health care, you won't want to miss these stories. Let's dive in.

[00:00:30] Hello, everyone. My name is Solomay Timaboo, host of the Expanding Access podcast. And as always, we have an amazing guest, expert, the intersection of mental health and AI. Thanks so much, Dominik Middelmann, for joining us. Just to jump right in, can you start by telling a little bit about yourself, how you got interested in this space? And of course, for those unfamiliar, what is mdhub? Yeah, absolutely. First of all, thanks for having me.

[00:00:58] I've been following you, and we've been at BHT now two years in a row, and I'm excited to attend this year as well. Yeah, so thanks for having me. A bit about myself and mdhub. So, yeah, my background really is in building and scaling product and operation teams in hyperscale environments. At my previous gig, I was one of the founding employees, and we actually scaled the company from zero to 4,000 people and more than 300 million in revenue.

[00:01:28] So, this was a fun journey. It was an unrelated industry, though. And my passion was always for behavioral health. And during the COVID times, this got enhanced even more as some of my colleagues and close family and friends started struggling really with their mental health. And mental health. And this really showed me the gaps in the industry more and more.

[00:01:53] So, I then decided to apply my knowledge of building a product and tech to the space. And then really how the stars aligned was when my co-founder, my now co-founder, Efren, who I worked with for the last eight years by now, mentioned that he's working on a side project with a psychiatrist. And he was like, Dom, all these psychiatrists are working with Penn & Pay. The tools they use, they look terrible. Let's build. And that's really how mdhub started. Amazing.

[00:02:23] What a journey. And, yeah, tell us, what is mdhub exactly? Yeah, absolutely. So, really, mdhub has evolved quite a bit over the years. But in a nutshell now, we essentially have built an AI operating system for behavioral health clinics and to run in one system end-to-end. So, we essentially bring together the concept of AI workers across admissions, clinical operations, and then also on the billing,

[00:02:51] allowing basically clinics to replace legacy CRMs, EHRs, and RCM solutions so that they can actually operate end-to-end on mdhub. It's been a fun journey over the last three years. We have now more than 10,000 clinicians actually on the platform. And clinics like Tokayatry, the Family Care Center, or interventional clinics like Radial that are just getting started as well operating on mdhub. That's fantastic. And big.

[00:03:19] That's a really big deal, what you're saying. In fact, I've heard you argue before that operations are really the bottleneck to care and behavioral health. Can you talk a little bit about what that looks like day-to-day inside a real clinic? Yeah, 100%. So, when I started to become really an expert, like a few years ago in this space, I always was surprised that even though there's obviously such a shortage of clinicians,

[00:03:47] not all clinics were operating at full capacity. So, clinicians had availability and clinics were actively looking to start, like to basically get patients through the door. And what then became more and more obvious when we started to dig into the clinics and the operations was that the infrastructure they were operating on became the bottleneck, really. And I always look at this in three buckets, really.

[00:04:16] Like if you look, for example, the front door, like a digital front door, then clinical operations, and then on the billing front. And very classic example on the front door, right? Patients all call on Monday 8 a.m. And it doesn't matter how many people you staff, you will have missed calls, right? So, then when a patient is actively like seeking help, and we all know in behavioral health, this window is also sometimes very short.

[00:04:42] And then if they're on hold or they don't get through, this is obviously a terrible experience for the patient and also for the clinic. Yeah, I missed the opportunity to get patient access to care. And yeah, the same trend really goes through all of these three buckets, right? Same for clinical operations. We all know the story of clinicians spending half their day on writing notes, right? Or writing treatment plans. So, this we really have automated through our AI clinical assistant.

[00:05:12] And then the same actually on the billing front. We work with clinics that have actually millions in uncollected claims, right? Just sitting there in the backlog. And this obviously impacts the financial health of organizations. That's, again, becoming a limiting factor of really getting patient access to health, mental health. This all sounds really great. And clearly, there is a lot of opportunity for AI to impact behavioral health operations like you're talking about.

[00:05:41] But how do we earn the trust of clinicians to embrace AI more? I think it really is about a few factors. One is obviously privacy and security is a huge cornerstone, right? Especially there are more and more startups coming into the space. They claim to be HIPAA compliant, even though it's not a formal audit, right, that companies need to do. So, you have to then really trust these companies out there.

[00:06:08] So, what's really important is to also look under the hood. Do they take a step further, right? Like becoming SOC 2 compliant and so on. Looking through the privacy policy, the terms on service. This is really, really important. And to start building that trust. And then I think another key factor that wins trust is building a system that's reliable.

[00:06:32] So, in the early days, of course, there were, it wasn't as smooth, right, when we got started three years ago. But over time, we have become so, so reliable that clinicians can just really focus on their patients. And we just make sure that everything works in the background. The same with other products that we have, right? Like with our call agent, we call her Sarah. For some clinics, actually, she performs the work of 20 FTEs.

[00:06:59] So, this is of huge, huge importance that this system doesn't just go down all of a sudden, right? So, we really invest a lot in building a very reliable infrastructure that works at scale. Wow. 20 FTEs. That is remarkable. That makes me think, how does really unifying admissions, clinical workflows, and billing all into one backbone really change outcomes for clinicians and patients?

[00:07:28] Yeah, absolutely. If you look at the industry right now, how clinics are run, it's one EHR, and then you have all these different point solutions bolted on. And a huge piece of the problem is that they oftentimes don't really work with each other or communicate. And then information gets lost in between them, right?

[00:07:50] Which then could mean that a patient falls through the cracks when they want to get booked or someone forgets to check their insurances before the evaluation. And then they come in and then the psychiatrist can't see them. So, all of these different problems arise if you operate on all of these different point solutions. So, what we do, for example, with one of our customers is where we always give our agents different names. So, Sarah is our AI admissions coordinator.

[00:08:19] Emma is our AI clinical assistant to help the clinicians before, during, and after the session. And then Eric is doing all of the billing work in the background once the note is done. So, when, for example, Sarah picks up the phone, checks your insurance, and basically triages you and then matches you with the right clinician, we then start to collect all of your insurance information, right?

[00:08:42] And then, basically, you flow into our CRM where we have all these automations in place to get you ready for the first appointment. So, it might be reminders for you to then fill out the paperwork. It might be actually a nudge from our agent giving you a call. So, really engaging you and making sure the VOBs is checked in detail and that then the clinician has all of the relevant information actually pre-charted through Emma before they see you.

[00:09:11] So, they have a lot of information about you before they actually are seeing you. And then, obviously, you know, we have already collected the VOB and the insurance information in the beginning for Sarah. Then Eric then also knows how to create the claim, how to basically make sure it's clean, and then submit it with the clearinghouse. And this way, actually, we've built a system where AI workers and the clinicians and the admin team all collaborate in one system. Wow, wow, wow.

[00:09:40] That is, and for our audience who are not familiar, VOB, that's verification of benefits. That's just remarkable. I feel like you're missing one more. You need intern Solomay Timaboo somewhere in there. A hundred percent. Yeah, we'd gladly have you. You said awesome, amazing. You've got like a whole clinic.

[00:09:58] Now, I can also infer some of the benefits of what you're saying when it comes to, instead of just layering on AI onto existing systems, just having everything built on one AI native operating system. Is that fair to say? Yeah, it is absolutely fair to say. And it's a story that we've seen a lot, right? Like incumbents versus new players like ours, right? Who's going to win?

[00:10:27] Obviously, EHRs, they do have the distribution. They do have the relationships with the customers. But I see them over and over really struggling building something truly AI first. Building an AI scribe doesn't mean that the EHR all of a sudden is AI first, right? Or adding one call agent to that as a bolt on solution doesn't make the whole system AI first.

[00:10:56] And obviously, AI is a remarkable technology and it has become easier to build companies from scratch. And that's why our focus actually on behavioral health allowed us to go really deep into this industry, but build a solution that works end to end.

[00:11:12] Really from first patient phone call to final reimbursement and really building a system that is built from the ground up AI first rather than bolting on different external solutions or building poorly built AI scribes potentially into legacy systems. Right. That makes a lot of sense. Thanks so much. That's really helpful to understand.

[00:11:36] And now, Dominik, I wish I could ask so many more questions, but if I could ask you one more, it would be we've got quite a few behavioral health providers in our audience. And some of them might be thinking, like, where do I even get started with all of this? What would you recommend that they have in place first to set them up for success when embracing really advanced AI solutions like yours? Yeah, absolutely. I think there are different ways to approach it.

[00:12:04] And I think maybe for a solo clinician, right, one thing is to make sure you test the tools, right? A demo is always shiny. It might work. It might not work then in production, right? When you're actually seeing your patients. So you really should use a tool that allows you to try it potentially even for free for a while and then really get this adopted into your workflows and then see if it works actually at scale.

[00:12:33] And for larger organizations, I think, obviously, when it comes to call agents or billing agents, it is relatively straightforward to, you know, promise you the world. And I've seen a lot of clinics actually, you know, being over-promised a lot from AI vendors. And that's why we're taking a bit of a different approach. And companies see us more as their AI partner.

[00:13:00] And here it is very important to look beyond the demo and set up a pilot, maybe with demo data or a small scale pilot when we, for example, talk about the admissions coordinator after hours or at peak times to really see this in production and make sure that this actually works. I think this is really, really important in a world where obviously there's a lot of competition and companies out there.

[00:13:28] And then also checking references, right? So, for example, if you work with us, we can provide you with references from clinics that have like a thousand plus clinicians and are using all of these tools in production. And I think this is really, really important to really take one step further than just, you know, listening to the demo. Really great advice for any new solution, candidly. Excellent. Thank you so much, Dominic.

[00:13:52] Really a pleasure to connect with you and share your story and expertise and the great examples you're seeing across MD Hub. Absolutely. Thanks for having me and excited to see you very soon in person again. Thanks for tuning in to Expanding Access. We hope you're feeling inspired by today's discussion. If you've enjoyed the episode, subscribe and share it with your network.

[00:14:16] For more content and opportunities to get involved in transforming care, visit BehavioralHealthTech.com. Until next time, let's keep pushing boundaries and expanding access together.