Transcript:
0:0:0.0 –> 0:0:10.870
Jordan Cooper
We’re here today with Sam Garris of Humana, he’s associate vice president of Information and Technology at of Healthcare Services at Humana. Sam, thank you so much for joining us today.
0:0:11.490 –> 0:0:15.780
Sam Garas
Thanks, Jordan. It’s really a pleasure for you to invite me and it looking for a great discussion.
0:0:16.640 –> 0:0:48.950
Jordan Cooper
So for as background for our listeners, Humana is a health insurer based in Louisville, KY with 11 million commercial and Medicare Advantage Health plan members in the United States. Now today we’ll be talking about particularly about humanized response to kovid and how that those innovations have been leveraged and a post COVID world particularly, we’ll be speaking about the use of artificial intelligence and machine learning driven data analytics platforms that were used at Humana to create cohorts for member outreach during Humana as initial response to COVID.
0:0:49.120 –> 0:1:1.530
Jordan Cooper
And subsequently to reduce readmission rates post discharge, Sam would like to kick off the conversation by asking how those products that Humana developed in order to respond to COVID are being leveraged today to improve outcomes.
0:1:2.0 –> 0:1:22.800
Sam Garas
Yeah, Jordan. So as as everyone knows, it’s been three years since that time in March of 2020 when the world, the world did change and we all had to go through it adopt, you know, really basically overnight to what was happening in real time throughout the world, so.
0:1:23.480 –> 0:1:53.590
Sam Garas
And here Humana we we you know this aside from getting making sure all our our our employees and socialites were were safe and moving everybody to work from home and getting all that set up. We also have to focus on our Members and more specifically our Medicare Advantage members who are seniors in their 60s, seventies, 80s some in their 90s obviously and really we’re struggling with the lockdowns and the.
0:1:53.870 –> 0:2:24.370
Sam Garas
The lack of being able to go out, get their medications, getting their food delivered to them. So what we did, I won’t say overnight, but it seemed like it was overnight within a within a few days we were able to get all the all the information that we had on our Members, all the analytics that we had and really retooled that to really target those Members that were at a high risk. You know, like you referenced in the beginning, we have several million members for Medicare Advantage.
0:2:24.470 –> 0:2:56.400
Sam Garas
So we just couldn’t all at once try to contact every single member at the same time really ID to make sure that we were targeting the most at risk. So our data scientists are our IT teams and everybody just really worked around the clock basically for several days and we’re able to identify quickly who those members were that we needed to target that were lower at risk. So and be able to put them into those categories. And so once we did that, we probably identified around.
0:2:56.470 –> 0:3:1.520
Sam Garas
5th or so members that were in that super high risk category they were had, uh.
0:3:2.210 –> 0:3:32.140
Sam Garas
You know, comorbidities, multiple chronic conditions where we’re living at home by themselves. Uh, we’re in a area that was highly impacted at at the time when COVID started. I think you’ll remember that the New York area in the northeast was being hit really hard at first. So we targeted put built that into the model. As far as who we were going to target. And we just really pivoted our teams and had to implement a lot of changes within the processes that we run within humanity of the final, those Members right there and so.
0:3:32.440 –> 0:3:35.440
Sam Garas
So continue to use that, that, that.
0:3:37.360 –> 0:3:59.810
Sam Garas
Analysis As far as who we would target. And so then once we got through that, the high risk and then just kept working through it. So it was really out of that. We were able to adopt A lot of those practices and as far as how we now do our outreaches today and how we try to really target who those at risk members are and who we really need to focus on instead of just doing a broad overall campaigns towards those Members.
0:4:0.490 –> 0:4:5.590
Jordan Cooper
Would you be able to elaborate upon the nature of the messaging that was sent to those 50,000 people?
0:4:6.540 –> 0:4:39.160
Sam Garas
So it was really more of a of a of a of a personal touch. We were so back up a little bit. So with within the Humana organization as at a whole we and as both just about every other healthcare company has, we have our care management teams and are really focused on you know when you get discharged from the hospital or you maybe have certain circumstances that meet that higher level of care. And so we had those teams in place, but we really pivoted them towards those at risk members so.
0:4:39.250 –> 0:5:9.780
Sam Garas
To really just kind of separate it in ourselves from our traditional ways of operating and and really going after those Members, so we also partnered with a lot of companies that were specific to the area. So like in New York, there was a lot of meal service companies that we were able to contact with. So it may not have been if we couldn’t get in touch with the member directly or if we did. And they said, well, I’m fine, I’m OK. But I’m really having a hard time going out and getting just some food delivered to me either they maybe.
0:5:9.960 –> 0:5:23.830
Sam Garas
From a physical standpoint, they couldn’t do it or just because of the the lockdowns that were in place couldn’t do it. So we contracted with a lot of and found the best meal service deliveries for seniors that we could partner with and we would make the call for them. We would say, hey, we’ve got.
0:5:24.330 –> 0:5:50.300
Sam Garas
Uh, you know this person that’s at this address in New York and we want to have this, umm, this food delivered to them. So we really were able to do that quickly. So. So I was kind of how we handle it and then as we started to move more into a broader population and may have been just an automated call out if everything was going OK. If you need more assistance, here’s where you can contact some e-mail campaigns. More of those types of more traditional outreaches that we would do.
0:5:51.330 –> 0:6:22.80
Jordan Cooper
So I’m interested. Obviously 40 years ago, you wouldn’t see a health insurer paying for groceries for their members. I think to some extent that probably reflects the increasing prevalence of value based payment models. But then again, you are the payer. I’m wondering, is Humana fully at risk, especially with these publicly funded programs, what what would you speak to the evolution of how?
0:6:22.340 –> 0:6:33.370
Jordan Cooper
Uh Humana has been seeing its role in caring for aging members, promoting health and, and in particular what would lead organization like Humana to begin coordinating with mail service companies.
0:6:33.680 –> 0:6:58.920
Sam Garas
Well, I think it’s all, you know, this is across everyone and it’s really about how do you continue to engage with the Members you you know you talked about value based cares you know bringing those incentives to the providers to actually be vested in the outcomes for their members, for their patients. And you know I think we’re as an industry learning through that.
0:6:59.790 –> 0:7:20.730
Sam Garas
You know more in moving in for, you know, the traditional fee for service. You go to the doctor, you you get your, you know, your procedure, you do whatever you’re doing, and then you get, you know, you charge a fee and you get paid. So that’s all still a big part, obviously, of how with Medicare and Medicare Advantage and trying to drive more towards the value based.
0:7:21.900 –> 0:7:51.140
Sam Garas
Outcomes or plans and going that way, but you know we Humana at least as an has always been focused on on on the Member and those outcomes and there’s you know the overall well-being and that’s an easy thing to say and it’s very broad but we able to impact their overall health by how the programs we do for for example like again how some of that work that we did related to COVID now is.
0:7:51.220 –> 0:7:56.130
Sam Garas
Is kind of advanced three years later into a lot of it’s just normal day-to-day practices.
0:8:23.50 –> 0:8:23.570
Jordan Cooper
Yeah.
0:7:57.410 –> 0:8:26.280
Sam Garas
As soon as our our Members are are discharged from the hospital, you know we that kicks in our our processes to go through the algorithms that we have to determine what Members are at higher risk for readmission within that 30 day 30 day window. So we know that if you can get to the member, the Members can do follow up within a certain period of time if they can get their medications timely that they’re going to, they’re not going to be readmitted within that 30 days.
0:8:26.830 –> 0:8:57.620
Jordan Cooper
So, Sam, you you are the the associate vice president of information technology. So I’d like to delve into, well, your analytics, but also first that the tech infrastructure that’s required to create those cohorts to drive that new outreach. Would you speak about what solutions you’re you’ve been leveraging in order to identify those cohorts and then subsequently to manage all the different connection points and different data interoperability that’s required to facilitate?
0:8:57.40 –> 0:8:58.490
Sam Garas
Yes. Yeah.
0:8:58.110 –> 0:8:58.830
Jordan Cooper
Those players.
0:8:59.630 –> 0:9:0.210
Sam Garas
Yeah.
0:9:0.920 –> 0:9:5.540
Sam Garas
There’s a lot there, a lot there, but and I’ll I’ll try to keep it as.
0:9:6.330 –> 0:9:27.680
Sam Garas
You know, as as high level or at least from my understanding, but you know we all we all talk about you know the cloud and we’re we’re moving to this and you know a lot of for healthcare companies what is that, what does that really mean? But what does that really mean? But for me it’s and it’s where we’re at is that leveraging the cloud capabilities and that can be with multiple partners it could be.
0:9:29.510 –> 0:9:57.80
Sam Garas
You know, with with Microsoft or or or or Google or Snowflake or and and other large scale data platforms that are that are either either SAS or cloud based. That or leverage certain cloud providers that we can utilize to house the more an enormous amount of data that we’re that we’re dealing with. I mean we’re talking about you know again as you referenced earlier, we have millions of Members millions of claims.
0:9:58.90 –> 0:10:18.940
Sam Garas
You know all that millions of medical records, pharmacy information, you know, all that and pulling that all together and driving to what the best outcomes would be and how we enable and or engage the right members for those outreaches that we’re doing. So so again just using that example of if you’re.
0:10:19.190 –> 0:10:52.530
Sam Garas
If you’re, you know if, say, you’re discharged for whatever the procedure was, and let’s let’s say it was a, you know, knee replacement. So you’re you go in for your knee replacement, you’re in the hospital for a few days and you get discharged. So we get that discharge event. Now, you may be a little bit different than your normal knee replacement. Maybe you had some other conditions that put you in a little bit of higher risk, higher risk of infection, a higher risk of maybe you’re older or so your rehab may not be as great or you have a higher risk of.
0:10:52.620 –> 0:11:1.740
Sam Garas
You know, you know, getting another potential injury as a result of the new replacement you just had within within that time window, so.
0:11:2.560 –> 0:11:3.130
Sam Garas
Umm.
0:11:4.480 –> 0:11:16.220
Sam Garas
The data that we have and what kind of what some of the prior experiences have been, we’ll all be used as far as formulating a method about who would who would you fall in that category? Would we want to have a heightened level of of outreach to you, so.
0:11:19.480 –> 0:11:19.830
Sam Garas
Mm-hmm.
0:11:24.590 –> 0:11:24.900
Sam Garas
Yep.
0:11:16.80 –> 0:11:45.540
Jordan Cooper
Just suppose we have Sam Smith. He lives in a suburb of Louisville, KY. He’s maybe 50 miles away, kind of suburban rural area. And let’s say he’s 73 years old. And the thing about Sam is, you know, he visits a he got an MRI at a standalone center. He visited a PCP at an independent practice.
0:11:51.270 –> 0:11:51.670
Sam Garas
Yeah.
0:11:52.750 –> 0:11:53.120
Sam Garas
Yeah.
0:12:1.20 –> 0:12:1.450
Sam Garas
Yep.
0:11:46.320 –> 0:12:16.90
Jordan Cooper
He was able to get orthopedic care from a specialty orthopedic practice. He had an inpatient stay, right. How do you reconcile all those different data sources? EHR is gonna be living in epic and other and Cerner, there’s gonna be all scripts how and there’s a PBM that mails him his drugs. How do you reconcile all these different data sources, data formats and pull them all together in such a way that you’re able to compare apples to apples with one source of truth?
0:12:16.180 –> 0:12:22.0
Jordan Cooper
And create that cohort among all your 11 million customers paying covered a lot.
0:12:21.530 –> 0:12:27.480
Sam Garas
Yeah, I mean that’s the challenge and that’s what we’ve been really that’s our whole integration strategy that we’re doing so.
0:12:28.540 –> 0:12:57.490
Sam Garas
Where at we’re just we’re at the the beginning of it. I think all of us are and I think we’ve done with Humana. I think we really advanced it significantly. I mean you know, all all of which you just mentioned, we’re gathering pieces of that, integrating it all. We really probably simplistic method. We want our when our care managers get that I think you said Sam Smith Ducky when we get that acknowledgement that hey we got a referral.
0:12:57.570 –> 0:13:29.100
Sam Garas
That Sam Smith or we would call it referral. But you know, we need to reach out to Sam. We maybe we need to do an initial assessment kind of get where they’re at, how they’re doing. We want that care manager to have all of that information right there. You know, basically it’s on that one screen. If you wanna simplify it to that, they can have all that pulled in. Now is it all going to be there? It’s trying to get everything that at least we have access to. You mentioned epic. We have a lot of we work very closely with EPIC obviously and try to do with all that integration and.
0:13:29.370 –> 0:14:0.480
Sam Garas
And hot. So how were those other sources we can get out of the third party vendors that we can use that leveraging and pulling that information and complete some of that integration pieces and pull those out together. So but yes, that’s the challenge is making sure that we have and we also want to make sure it’s really difficult even within Humana, we may have another group that’s focused on closing gaps in hair. So say you’re a certain age and you haven’t, you just turned 50 and you never had a colonoscopy. So we want to make sure you get that done.
0:14:0.910 –> 0:14:33.300
Sam Garas
Uh, maybe you’ve never. You haven’t had a mammogram, or you do for one or whatever your situation is. So there’s other teams that may be doing that outreach. While Sam Smith he just got released for that knee replacement. We want to make sure we’re not having multiple people contacting and just really overwhelming the Member as well. So integrating that all together and what those things are and we want to test saying that just got out of the hospital with the new replacement, that they should also look at a colonoscopy. And that’s not a great member experience.
0:14:33.390 –> 0:14:39.160
Sam Garas
Uh, so really, how do we get that all there? So, but that’s a challenge. I mean we.
0:14:40.510 –> 0:14:52.460
Sam Garas
But it’s all member focused. It’s all outcomes based. It’s all. How do we eventually enable and and get to the point where we’re influencing the overall health of the Member and giving them what they need?
0:14:53.230 –> 0:15:1.700
Jordan Cooper
And now that you’ve been operating with these programs for a few years, have you gotten any feedback from patients, for instance, any sort of HEDIS scores?
0:15:10.30 –> 0:15:10.480
Jordan Cooper
Mm-hmm.
0:15:2.720 –> 0:15:15.970
Sam Garas
Yeah, I think you know that’s that’s very well known. I mean Humana is, you know done great the past couple of years with our scores and you know a lot of a lot of the plans that have been there. So that’s that’s always.
0:15:23.280 –> 0:15:23.680
Jordan Cooper
Mm-hmm.
0:15:34.690 –> 0:15:35.370
Jordan Cooper
Hmm.
0:15:44.80 –> 0:15:44.510
Jordan Cooper
Umm.
0:15:16.990 –> 0:15:46.50
Sam Garas
You know, a major focus across the organization, the across all of all of our, all the healthcare payers and all within the healthcare system. I mean that’s and so yes, I think that’s kind of a great benefit of it as we’re really focused on the member by implementing those program which is going to improve those scores and it just continues down that chain. So it’s not we’re not doing it just because we want to improve our scores, but we we’re also doing it because.
0:15:46.130 –> 0:15:50.140
Sam Garas
Wait, it’s going to impact that the the health of the members so.
0:16:7.770 –> 0:16:8.200
Sam Garas
Yeah.
0:15:50.330 –> 0:16:19.860
Jordan Cooper
Should Sam, as we approach the end of this podcast episode to bring us back to the original topic that we introduced at the top of the of the hour here we and that topic is analytics, particularly machine learning driven analytics and how those algorithms are being applied in a post COVID world, would you speak more about what sort of innovations you’re analytics platform is, is is working on what sort of projects you’re pushing?
0:16:18.420 –> 0:16:48.870
Sam Garas
Yeah, I’ll give. I’ll give one. I’ll give one great example. And actually I believe where it one of the finalists for one of the CIO awards. So we’re hoping to to get that. I think it’s gonna be announced here over the next few months. So hopefully we we win that. But I can I can tell you a little bit about it. It’s so as you know we get the healthcare industry, we still get lots providers still use faxes, right. And believe it or not, that’s still what technology that’s being utilized. So that’s our.
0:16:48.990 –> 0:17:19.70
Sam Garas
Pre authorizations come in. You know we talked about knee replacement. You know, whatever the procedures are, you know there could be another request for the medical records. Uh, you know, different types of things. You know, maybe there’s just more specific requests for those of those pre OPS. So a lot of those are mostly facts. Dan from the provider. So we’ve received those and it could be a fax to mail, fax to e-mail type situation. So we’ll get those in and so we’ve implemented a.
0:17:19.730 –> 0:17:49.900
Sam Garas
More of a natural language processing, but also feeds into the machine learning to kind of get what some of the more common ones that we’re receiving are and so we can really batch those up and then we will funnel those and try to match those back to the authorization or the preauth that was originally requested for that information. So before we would, we would have a team of people that would go in and say OK, we just received you know, balance and faxes and the past hour, let’s see, OK, this one is for authorization.
0:17:49.980 –> 0:18:24.900
Sam Garas
OK, this one’s for authorization B. Let me go make that linkage within our utilization management system and put all those connection points together. So now we can use the natural language processing a lot of the ML that can continue to learn feed off of itself and then be able to start to match up where those should go. Now we still have a level of verification. So we still want to make sure that it did match up. But I think as we continue to advance this and get to it, it’s just those are going to become in and those are just going to be completely followed to it. So you can imagine the, the.
0:18:26.980 –> 0:18:27.540
Jordan Cooper
Umm.
0:18:24.990 –> 0:18:54.830
Sam Garas
Efficiencies that we’re gaining there, it’s reduced our time of maybe taking 10 minutes or so to go through that and trying to match it up with the preop down to just the just the you know a minute or so and eventually that will go down to 0 because we won’t need any user interaction at all. So that’s a great initiative that’s kind of a simple example, but it’s one that’s really paid off significantly and and just to end too, I know you may have seen that Larry Ellison from Google, I’m not Google from Oracle.
0:18:55.0 –> 0:19:15.580
Sam Garas
Mention of on their earnings report a couple of weeks ago, some of the work that they’re doing on reducing readmission rates and some of the AI that they’re doing in any compared it to ChatGPT can write your high school essay, but you know, a lot of things that we’re doing within the healthcare space is reducing readmission rates. And I think that’s that was a powerful statement about.
0:19:16.220 –> 0:19:39.170
Sam Garas
How the AI is not just about, you know, telling you where you want to go for your next you know, restaurant, you know which restaurant you want to book or write helping you write an e-mail. It’s also using that to really target where you can reduce the like. We talked about before the readmission rates and that’s really powerful. It’s it benefits the member and it’s also from a claims experience and expenditure. It can significantly reduce that.
0:20:6.860 –> 0:20:7.390
Sam Garas
Awesome.
0:20:7.850 –> 0:20:19.920
Jordan Cooper
Yeah. For our listeners, this has been Sam Garris, the associate vice, the vice President of Information Technology and Healthcare Services at Humana. Sam, on behalf of all our listeners, I’d like to thank you for joining us today.
0:20:20.560 –> 0:20:21.970
Sam Garas
Thanks, Jordan. Really appreciate it.
