AI-Powered Medical Coding and Value-Based Care with Hamid Tabatabaie, co-founder and CEO of CodaMetrix

[00:00:02] [SPEAKER_00]: Hey everybody, welcome back to the Outcomes Rocket HFMA 2024 National Insights Series. I'm so excited that you tuned back in with us to capture some of the most insightful ideas from the biggest innovators that I met at the conference.

[00:00:19] [SPEAKER_00]: Today, as part of this series, I have the privilege of hosting the amazing Hamid Tabatabaie. He was on the podcast before, so you guys remember him.

[00:00:29] [SPEAKER_00]: And Hamid, I'll just do a quick intro on you. He is the co-founder and CEO of CodaMetrix. They are exploring the latest and greatest pathways with the use of technology and the rev cycle to make coding automation a possibility.

[00:00:45] [SPEAKER_00]: Hamid, such a pleasure to have you with us again.

[00:00:48] [SPEAKER_01]: Thank you very much. Good to see you, Gantel.

[00:00:51] [SPEAKER_01]: Likewise.

[00:00:51] [SPEAKER_01]: Thanks for inviting me.

[00:00:52] [SPEAKER_00]: It's such a pleasure. It was great to connect with you at the conference and really excited.

[00:00:57] [SPEAKER_00]: Look, you already shared with us what your business is all about, but talk to us a little bit more about the rev cycle focus area that you're in and what you're doing.

[00:01:08] [SPEAKER_01]: You bet. The company got its start from Mass General Brigham, so it's provider-owned.

[00:01:13] [SPEAKER_01]: From the very beginning, it had a mission. Its mission was, A, to improve accuracy and proficiency of coding, medical coding.

[00:01:25] [SPEAKER_01]: And secondly, to use AI and make claims data much more reliable when it comes to clinical use.

[00:01:35] [SPEAKER_01]: And from the onset, this was a strange ask because the revenue cycle side, you have fee-for-service as a great majority of everybody's volume.

[00:01:51] [SPEAKER_01]: And as a result, coding for medical necessity suffices.

[00:01:57] [SPEAKER_01]: As in, you don't need to clinically be going into too much detail about the procedure and or the diagnosis.

[00:02:06] [SPEAKER_01]: You need to have sufficient evidence to explain to the payer why you did something.

[00:02:12] [SPEAKER_01]: That was justifiable because coders cost too much money if you slow them down to be more specific.

[00:02:20] [SPEAKER_01]: But with AI, there is really no excuse for that.

[00:02:24] [SPEAKER_01]: You can become, A, the avenue for addressing the gap between clinical specificity and claims data.

[00:02:32] [SPEAKER_01]: And B, you can become a partner, or as nowadays it's referred to a co-pilot, to coders, even for things that you can't automate.

[00:02:42] [SPEAKER_01]: So this, over the last couple of years, has become a lot more in focus.

[00:02:49] [SPEAKER_01]: Partly because there is a gap in finding coding resources.

[00:02:57] [SPEAKER_01]: There's just a shortage.

[00:02:58] [SPEAKER_01]: And partly because the amount of information that has been added inside the EHR has grown to the point that now you need to be coding a lot more.

[00:03:10] [SPEAKER_01]: So on one side, it brings the volume down, it brings the cost down.

[00:03:15] [SPEAKER_01]: And on the other side, it becomes much more relevant on the clinical side.

[00:03:20] [SPEAKER_00]: Yeah, that's fascinating.

[00:03:22] [SPEAKER_00]: The clinical specificity is key.

[00:03:24] [SPEAKER_00]: And now with this technology, health systems are able to better capture all of it.

[00:03:30] [SPEAKER_00]: And, Hamid, when we were together, you mentioned a term.

[00:03:33] [SPEAKER_00]: And I want to bring it back up here.

[00:03:34] [SPEAKER_00]: Was it value-based coding?

[00:03:36] [SPEAKER_00]: Is that what I heard you say?

[00:03:38] [SPEAKER_00]: Can you unpack that for us a bit?

[00:03:39] [SPEAKER_01]: Very much.

[00:03:41] [SPEAKER_01]: It's a play from value-based care and or population health, accountable care, and so forth.

[00:03:49] [SPEAKER_01]: At the moment, when you talk to large health systems, or even smaller ones, they tend to be recoding things for a variety of use cases.

[00:04:01] [SPEAKER_01]: As we've talked about, fee-for-service does it for reimbursement of encounters and hospitalization.

[00:04:09] [SPEAKER_01]: But then when you go to the population health group, they pull charts and have nurses recode everything to achieve that level of clinical specificity.

[00:04:20] [SPEAKER_01]: When you go to the department that is in charge of clinical trials and recruitment, they pull charts and coded yet one more time.

[00:04:31] [SPEAKER_01]: Because now you need to know not only what the patient's history plays a role.

[00:04:39] [SPEAKER_01]: Was this patient suffering headaches and or dizziness a year and a half ago before they started this medication compared to right now, whatever their condition is.

[00:04:52] [SPEAKER_01]: So you need to go further than just the present acuity of the patient.

[00:04:56] [SPEAKER_01]: Then when you go to different departments for clinical registries, they yet pull charts and code again.

[00:05:05] [SPEAKER_01]: So value-based coding is a reference to code somehow that the entire enterprise is valuing the code.

[00:05:14] [SPEAKER_01]: Today is harder to convince people of that because they're very departmentalized.

[00:05:20] [SPEAKER_01]: One has this budget for coding and doesn't really get involved in the budget for coding with the next and so forth.

[00:05:27] [SPEAKER_01]: But as things are getting more integrated, this value-based coding and as AI becomes more pervasive, this value-based coding really is what will happen.

[00:05:39] [SPEAKER_01]: Whatever the terminology for it will be in the industry, we have elected to call it value-based coding.

[00:05:45] [SPEAKER_00]: Yeah, no, that's super interesting.

[00:05:47] [SPEAKER_00]: And I'm glad we're double-clicking on that today, Hamid.

[00:05:51] [SPEAKER_00]: And I just envision, just with what you shared, an opportunity for a health system to modularize their needs.

[00:05:58] [SPEAKER_00]: So Legos, right?

[00:06:00] [SPEAKER_00]: You got a value-based care Lego, you got a clinical billing Lego, and you could decide what you want.

[00:06:05] [SPEAKER_00]: And then when the AI does its work, it could give you all that or just one of them.

[00:06:10] [SPEAKER_00]: So it's pretty customizable from what I hear you saying.

[00:06:13] [SPEAKER_01]: Very much so.

[00:06:15] [SPEAKER_01]: Particularly that, in our opinion, there shouldn't be a book that decides what coding quality is.

[00:06:23] [SPEAKER_01]: Each provider should decide that because for all practical purposes, it's the provider's performance that is being documented.

[00:06:32] [SPEAKER_01]: And it's provider's decisions that are being documented.

[00:06:36] [SPEAKER_01]: And we've created this rubric of one star to five star.

[00:06:41] [SPEAKER_01]: And our customers get to decide what procedures they want coded with what quality.

[00:06:50] [SPEAKER_01]: One and two stars are, one way or another, incorrect.

[00:06:55] [SPEAKER_01]: Three stars are perfectly great for fee-for-service.

[00:06:58] [SPEAKER_01]: Four stars are clinically appropriate.

[00:07:01] [SPEAKER_01]: And five stars are all the way to research.

[00:07:04] [SPEAKER_01]: And for every exam that goes through us, we do three, four, and five.

[00:07:09] [SPEAKER_01]: And whatever use case is appropriate, we hand that use case the appropriate code sets.

[00:07:15] [SPEAKER_00]: That's so awesome.

[00:07:16] [SPEAKER_00]: I love that gradation of need.

[00:07:19] [SPEAKER_00]: And back to cementing your point on clinical specificity, when doing trials, you have the most.

[00:07:24] [SPEAKER_00]: So that's a five.

[00:07:25] [SPEAKER_00]: Whereas billing, you still want to do a good job.

[00:07:28] [SPEAKER_00]: So it's a three.

[00:07:29] [SPEAKER_00]: That's exactly right.

[00:07:31] [SPEAKER_00]: So cool.

[00:07:31] [SPEAKER_00]: I love that.

[00:07:32] [SPEAKER_00]: Folks, super interesting.

[00:07:33] [SPEAKER_00]: This is a type of innovation that Codometrics is bringing to market.

[00:07:37] [SPEAKER_00]: The ideas and thoughts that we were having at the HFMA conference are super fascinating.

[00:07:43] [SPEAKER_00]: So I'm glad we're getting to share this with all of our listeners and viewers today.

[00:07:47] [SPEAKER_00]: I mean, what recent news or insights about the company do you want to share with the viewers and listeners today?

[00:07:53] [SPEAKER_01]: Sure.

[00:07:53] [SPEAKER_01]: I'm happy to report that we're a healthy, growing company.

[00:07:57] [SPEAKER_01]: And we now have over 20 large health system customers.

[00:08:03] [SPEAKER_01]: Many of them, if not a great majority of them, are using Epic as their EHR.

[00:08:10] [SPEAKER_01]: And that caused us to have more dialogue with Epic.

[00:08:14] [SPEAKER_01]: And we have entered into joint Epic's developer program called Toolbox.

[00:08:20] [SPEAKER_01]: They have defined their development programs into four categories.

[00:08:26] [SPEAKER_01]: It starts with having access to anyone that can have access to certain application programming interface, APIs and interfaces.

[00:08:37] [SPEAKER_01]: The second one is getting guidance from them to know how to customize and or make it appropriate for any given customer.

[00:08:45] [SPEAKER_01]: The third one, which is called Toolbox, is what we have started with.

[00:08:50] [SPEAKER_01]: And that is certain joint developments for specific use cases.

[00:08:57] [SPEAKER_01]: And what we have is essentially a crawl, walk, run approach to start with.

[00:09:03] [SPEAKER_01]: It will be a lot easier for our customers to have higher automation, better implementation, faster deployment.

[00:09:12] [SPEAKER_01]: Walk will be all sorts of exchanges of new information pieces.

[00:09:16] [SPEAKER_01]: Because as you would imagine, Epic is going to be doing predictions on codes and or other information.

[00:09:23] [SPEAKER_01]: And right now, there aren't any interfaces like HL7 or others that govern the passing of that data.

[00:09:31] [SPEAKER_01]: What was your logic for predicting such and such?

[00:09:35] [SPEAKER_01]: That doesn't have any HL7 mandated anything.

[00:09:38] [SPEAKER_01]: So that's our walk approach, jointly developing that.

[00:09:44] [SPEAKER_01]: And third is essentially run is when we are incorporating our predictions into Epic's coding workflow hand-on-glove.

[00:09:57] [SPEAKER_01]: We do that today via the existing protocols, but that doesn't seem to be as enriching as one can be with this kind of a program.

[00:10:06] [SPEAKER_01]: We'll be able to serve our Epic customers even with more level of confidence, if that's the right term to use.

[00:10:15] [SPEAKER_01]: That's probably the biggest news that's going on.

[00:10:17] [SPEAKER_01]: Other than that, in the next couple of months, we'll be announcing additional executives that have joined the place.

[00:10:23] [SPEAKER_01]: We're getting ready for scaling the company.

[00:10:26] [SPEAKER_00]: Well, that's certainly exciting, both from a technology interoperability perspective.

[00:10:32] [SPEAKER_00]: Obviously, Epic being the big player, your 20 customers and now a growing leadership team.

[00:10:38] [SPEAKER_00]: Certainly an exciting time to be at Codometrics today.

[00:10:40] [SPEAKER_00]: And for everybody listening, when you hear these things, you hear something that resonates with you.

[00:10:45] [SPEAKER_00]: It's about taking action about that.

[00:10:47] [SPEAKER_00]: And that's why we're serving these insights up for you.

[00:10:50] [SPEAKER_00]: And regarding insights, HFMA, it was a fantastic conference.

[00:10:55] [SPEAKER_00]: What rose to the top for you as an insight that everybody with us today needs to know?

[00:11:00] [SPEAKER_01]: Something about hammer and looking for nails, so to speak.

[00:11:04] [SPEAKER_01]: I've tuned into a lot of the AI conversation.

[00:11:08] [SPEAKER_01]: And I was very happy to see that the market is graduating to a new level of sophistication.

[00:11:17] [SPEAKER_01]: On one side, you can tell when vendors are making bogus claims that we're going to automate 99% of everything and so on and so forth.

[00:11:28] [SPEAKER_01]: And on the other side, people are starting to be more armed with trying certain things to start in a pilot kind of a way,

[00:11:40] [SPEAKER_01]: but yet keep the enterprise in mind.

[00:11:44] [SPEAKER_01]: So start to move away from point-to-point solutions and be more holistic about what you can deploy over time.

[00:11:53] [SPEAKER_01]: So to net it, I think we are entering a new phase of understanding of AI and incremental adoption.

[00:12:04] [SPEAKER_01]: I don't think by any means are we in a place that everybody is going to be AI-enabled in the next 12 months.

[00:12:10] [SPEAKER_01]: But I think everybody is going to try certain aspects.

[00:12:14] [SPEAKER_01]: In my personal life, I do that.

[00:12:16] [SPEAKER_01]: In my personal life, I use AI for certain use cases.

[00:12:20] [SPEAKER_01]: And I saw in our HFMA that certainly AI is neon and everywhere.

[00:12:25] [SPEAKER_01]: But when you talk to users and buyers, there's more sophistication than it was before.

[00:12:32] [SPEAKER_00]: Yeah, that's really interesting.

[00:12:34] [SPEAKER_00]: And I definitely attended a couple sessions too where I did see a lot of engagement, to your point,

[00:12:41] [SPEAKER_00]: and a lot of curiosity, a lot of buying questions.

[00:12:45] [SPEAKER_00]: So it's a hot time to be in this space for sure.

[00:12:49] [SPEAKER_00]: Okay.

[00:12:49] [SPEAKER_00]: I mean, this has been incredible.

[00:12:51] [SPEAKER_00]: I want to, first of all, thank you again for being with us as part of this Insight series.

[00:12:56] [SPEAKER_00]: What call to action would you close us out with?

[00:12:58] [SPEAKER_00]: And where can the listeners and viewers reach out to you and the team?

[00:13:02] [SPEAKER_01]: You can reach out to us on codometrics.com or hello at codometrics.com via email.

[00:13:09] [SPEAKER_01]: My call to action is the observation I had earlier is revenue cycle teams to find the time and the mental focus to think about the enterprise.

[00:13:23] [SPEAKER_01]: Whether you're the fee-for-service department for the professional fees or you're at the hospital coding side of things,

[00:13:32] [SPEAKER_01]: what you produce does more than just get dollars collected.

[00:13:38] [SPEAKER_01]: It has an impact on downstream activities, on research and clinical areas.

[00:13:44] [SPEAKER_01]: And it can have a significant impact on patient care.

[00:13:50] [SPEAKER_01]: So my advice and my ask and my call to action is think enterprise.

[00:13:57] [SPEAKER_00]: I love it.

[00:13:58] [SPEAKER_00]: What a great call to action.

[00:14:00] [SPEAKER_00]: And I love it because most of the thinking is very much focused on a certain part of the funnel, right?

[00:14:07] [SPEAKER_00]: You hear the front of the funnel, the middle of the funnel, the ref cycle funnel, the back.

[00:14:11] [SPEAKER_00]: No, let's think enterprise.

[00:14:12] [SPEAKER_00]: And I love the reset button that we're hitting here on this interview with Hamid.

[00:14:18] [SPEAKER_00]: In the show notes, you're going to find where to get in touch with Hamid and his team.

[00:14:23] [SPEAKER_00]: Thank you all for tuning back into this Insights from HFMA.

[00:14:27] [SPEAKER_00]: Stay tuned for the next one.

[00:14:28] [SPEAKER_00]: And Hamid, we really appreciate you being with us.

[00:14:31] [SPEAKER_01]: Thanks for inviting me.

[00:14:32] [SPEAKER_01]: I appreciate that.

[00:14:33] [SPEAKER_01]: Pleasure.