← Becker’s Healthcare Podcast
Becker’s Healthcare Podcast
Real-Time Prior Authorization at the Point of Care: Abridge CEO Dr. Shiv Rao
00:11:02
Loading summary
Transcript11 lines
- [00:00]
A
Hello and welcome to the Becker's Healthcare Podcast. I'm Molly Gamble and I'm excited to share a special episode today in conversation with Dr. Shiv Rao, founder and CEO of Abridge. Abridge is one of the ambient AI companies we've been following closely at Becker's. And in full transparency, they're also a valued partner. Their support helps keep Becker's accessible and allows us to bring you conversations like this one at no cost. This company's been moving quickly. This year it expanded from 100 health system partners at the start to more than 150 today. And in this episode, Shiv and I are talking about their newest capability. It's this Real Time prior authorization solution launched in partnership with Highmark and Allegheny Health Network in Pittsburgh. This interview is conducted as part of my reporting on Abridge, but rather than keep it behind the scenes, we wanted to share it directly with you, our listeners, letting you see some of my homework and also hear directly from Shiv. So without further ado, here's my conversation with Dr. Shiv Rao. Thanks for listening. Let's get into it. I mean, this partnership, in the next few days, Abridge will be announcing is with Highmark and AHN in Pittsburgh, where you have some deep ties as a bridge. And it's not just a new partnership for the clinical note, but also with this launch of the Real Time prior Auth solution activated at the point of care, this feels like a breakthrough in some ways. I was curious, what about this moment feels most significant to you?
- [01:21]
B
Yeah, yeah. I think that there are in some ways there have been like three large strategic pushes for the company over these last several months that you've done the best job of being able to, like, truly break down at the right level of detail. One is just being enterprise grade. And I think you've, you've, you've written to that. And that's, you know, us being able to serve all the different clinicians and all the different specialties and all the different settings, like emergency departments, for example. And now we're in the inpatient setting. Another piece is about that contextual reasoning engine. And we've talked about how Abridge is now pulling information from multiple sources, sometimes seemingly disparate. They could be from data from the medical record. It could also be information or rules or clinical pathways that are coming from, you know, medical textbooks or from coding manuals. But then the third push for us that this is really all about, and I think to your point, frankly, for, for us, it's like a very, very important milestone, maybe like, the biggest milestone for us in years, really, is that we're starting to demonstrate to the market that a bridge is moving from largely being a passive experience where it's, you know, we talk about good air conditioning, where it's set right, you're not aware of the air conditioning. You're just fully focused on all the other things that should be taking your attention. But moving from that passive world where we'll always want to anchor, you know, that will. That will always be our first objective to being active when it makes sense, being in being more assistive when it makes sense. And so in the case of prior authorization, we are making the system smarter by, to your point, shifting left, moving a lot of workflows upstream, and thinking about what can we do to bend the trajectory for the quality of care and, and the quality of all those different downstream processes. At the point of conversation itself, what can we do in the conversation itself while the most important people in healthcare, the clinician and the patient, while they're together, what can we do to lighten the load on everybody, including them?
- [03:31]
A
Right, right. The shifting left piece, I'm gonna circle back to that in a bit, but I wanted to acknowledge, I mean, this has been such a persistent problem and it's getting worse in recent years. The prior auth piece, especially under MA plans. I know. And you've seen, you know, state lawmakers have passed legislation, or at least considered it in dozens of states. You've got both Biden and Trump. It's been a priority with the Trump administration. Just recently in June, HHS had a pledge with about 52 health insurers committing to some improvements they were going to make to this process. At the same time, we've heard from health system CEOs. You know, it starts to feel a little like lip service some of the conversation on this topic. There's been promises made about improvements that they haven't yet seen. Can you talk a little bit about where you see a bridge fitting into this landscape, what a technology partner like you can bring that maybe legislation alone can just simply not accomplish?
- [04:30]
B
Well, one aspect of these last few years that we certainly don't take for granted is that we've been able to earn trust and scale across the entire country. You know, as we've talked about before, we're live in well over 150 of the largest health systems in the country. And that means, you know, as we process these millions of encounters every week, we have the opportunity in the conversation to think about what other kinds of value can we create and I think sometimes technology's role is to, is to trailblaze, is to actually build those solutions that, you know, solve for all the different pain points that all of us in the industry and all of us as people are acutely aware of. I recently saw patients while I was rounding in the hospital and I wanted to order a cardiac MRI for a patient. And just that cardiac MRI led to any number of paper cuts, I would say, over the subsequent days and even weeks around whether or not this patient was a candidate for the cardiac mri. Did I think, through all the clinical pathways that this specific payer had put in place as friction for me to get that cardiac mri? Had I tried a CT scan before I jumped to the cardiac mri? And should I talk to another clinician on the insurance side to make my case? And I think to some extent there are absolutely good reasons to have there be some friction such that clinicians are thinking through, is this really the right thing to do? And that's where this technology comes in. Because we're, we're simply trying to improve the quality of the conversation itself. We're trying to give the clinician a cue to say, hey, have you thought about the CT and maybe you should address that? Or have you asked this other very specific question that their specific insurance plan will require from you to be in your documentation? And if you can think through those things, it's not as if you're gaming the system at all. You're just trying to help all these clinicians be omniscient, something closer to omniscient, so that they can deliver the best possible experience and hopefully the best possible outcome.
- [06:46]
A
And everything you just described there, that was for one patient, that was for one episode of care.
- [06:50]
B
Yeah, it's an incredible amount of friction. I mean, I think this is, we think about our mission. Of course, it's CRISPR into our DNA, but we're trying to unburden clinicians from all this clerical work and trying to bring people closer together. And clerical work, like clinical documentation, is such a great example of that. But clerical work, like prior authorization, is another great example. And to the points you were making earlier, it's top of mind for everyone. It's top of mind for patients, for all of us as people. It's top of mind for clinicians, and it's top of mind for insurance companies right now. I think that there are a lot of incredible tailwinds and a lot of, like, stars are aligned. You know, everybody, I think, is trying to figure out the most responsible way to solve for a part of this problem. And so that's where we fit in. So our models need to navigate through highmark specific prior authorization policies and need to determine, you know, approval eligibility basically before the patients leave the exam room. And that means that our models have to detect in real time that the clinician just recommended a certain diagnostic or therapeutic that needs to be authorized in this way, shape or form that requires these models to not just detect it, but to map that diagnostic or therapeutic to a structured code, a CBT code. In this case, the AI needs to do real time policy lookup and then decisioning based on what that code actually ended up being. It then needs to identify information from the conversation between the clinician and the patient so that it can queue up those gaps that are outstanding. So if you asked five out of the six questions, for example, related to a knee repair or a knee revision, then a bridge can kind of help you understand. Before the patient leaves the room, there's one more question that you need to ask. And then, you know, even beyond that, akin to the way we've thought about auditability and this concept of linked evidence with our core notes product, in this case, we also expose the AI's reasoning and we, we map all of the different, you know, boxes that have been checked off to the actual original policy itself so that that can build trust at the point of conversation as well. For, for the clinician that this technology is really has their back and their.
- [09:13]
A
Patient'S back prior auth it can be a dirty word in some circumstances or I worry sometimes that it has is gaining that connotation when it does serve a purpose. It, you know, it's a cost containment tool. It's supposed to prevent unnecessary care overuse. But I was curious about the payer's incentive here. You know, AHN and Highmark have a relationship, but what assumptions were challenged with this partnership? What is in it for the payer? Why would they be invested in a better point of care solution in real time for prior authorization? Can you talk about what you learned through this partnership on that?
- [09:51]
B
Yeah, I can speak at a high level and I'm sure that they'd be able to unpack this in a lot more detail. But I think in the case of hydmarc, there's no question that for them the objective function here is to deliver the best possible care and the best possible outcome to their patients. And I think what we're all recognizing perhaps as an industry is AI represents a tool that allows us to do this in radically different ways that allow us to solve problems earlier, smarter, and faster. We're shifting intelligence to the point of the conversation itself, and that they believe will lead to better outcomes, you know, for their, for their patients. And at the end of the day, that's what they're, that's what they're looking for.
- [10:34]
A
If this feels really important, it feels big, and it feels like a different type of solution to a problem that so many have been raising and talking about. It's been affecting more people every year as claims and prior auths go up. So congratulations, and I appreciate this insight from you today, and, you know, looking forward to what's to come.
- [10:56]
B
Awesome. Thanks so much, Molly. Really appreciate your time.
- [10:58]
A
Thanks, Shiv.