How to Use Analytics to Drive ROI in Value-Based Care with Summit Health

This transcription is provided by artificial intelligence. We believe in technology but understand that even the smartest robots can sometimes get speech recognition wrong.

Dr. Jamie Reedy: While there are many aspects of data quality that can and should be assessed, there's a few key signs that lead to confidence by end users in the quality of data. And we have really found those to be the correctness, the completeness, the integrity, the validity, and the relevance of the data. And addressing all of those are just, just incredibly important.

Special thanks to our influence show sponsors Sirius Healthcare and Health Lyrics for choosing to invest in our mission to develop the next generation of health IT leaders.

We're excited about where the community will take this channel. The Academy is about training. It's about training the next generation of health leaders. Here's where we're going to be launching our new show. It's called Insights and the show will actually take highlights from our last five years and break them into 10 minute episodes for your team and perhaps people who are new to health IT to come up to speed.

We have Dr. Jamie Reedy Chief of Population Health at Summit Health and Dr. Ashish Parikh who is also at Summit Health. And I'm looking forward to this because this was a presentation that you did at HIMSS. And I, I was fascinated by the analytics around the ACO, analytics around population health. So thank you very much for coming on the show.

Dr. Ashish Parikh: Thank you for having us.

Bill Russell: So Oregon, New Jersey and New York, that seems like a, not necessarily a geographic strategy. Seems like it's a strategy based on something else. Am I missing something there?

Bill Russell: So number of, of managed lives that the ACO covers?

Dr. Ashish Parikh: So we have about over 165,000 managed lives in both Medicare, Medicare Advantage, as well as commercial value-based contracts in New York and New Jersey. And then about 25,000 out in oregon.

Bill Russell: Fantastic. All right. So that, that gives us the scale. Tell us about the ACO vision and the journey that you've been on.

Dr. Jamie Reedy: Sure. So I can take that one. So in order to be successful in an ACO or any value-based program, we've learned that you need to have access to lots of data and you need to make that data work for you.

So those are a few of the foundational capabilities that we felt were really needed, where we needed data to support.

Bill Russell: So that's I mean, that's fascinating. So you have foundational dataset, you have workflow support, and then you have enhanced analytics is the foundation. So, talk to me about where the data comes from. We heard some of it's coming from the EHR. Some of it's coming from claims. Are there other sources of the data?

And now there are more and more sources of socio-demographic data as well that help us understand the potential social needs of our patients, which are incredibly important influencers of health outcomes. So that the data sources we integrate are really driven by our business needs and prioritized by considering technical challenges for accessing the data with each of these data sources.

Bill Russell: On the show, we've talked about a whole patient profile, building a whole patient profile and it sounds like you guys are getting pretty, pretty close to that.

Can you touch on the social determinants data real quick. Where are we getting that? Are we getting that from surveys and that kind of thing? Are we actually connecting into I don't know, some partners who are bringing that data into us?

Bill Russell: Talk a little bit about the value and the role of the claim.

And so complete claims data provide us that full visibility. In the early days of our value based care journey, our health plans were actually reluctant to provide this level of data, but we're finding now that most health plans are open to providing the data limited only by regulatory and privacy concerns.

Bill Russell: I'm looking at your presentation and you go into data quality and I'm going to ask the question, which sounds like I'm not a former CIO for a health system, but how important is data quality and which I think I know the answer to, but how do you know that your data is high quality and that it is trustworthy?

So incredibly important to get it right from the beginning. We desired strong engagement with our data. We were hyper-focused on data-driven workflows. And so when we built out our analytics platform from the beginning, we built in strong data quality review processes. Every data set that, that we integrate is put through rigorous review and testing before it's incorporated into productionized analytics.

So for instance, lack of confidence in our unreliability of the data, data that's new and unfamiliar may not be understood and hence not engender confidence in end users. Or uncertainty is common when clinicians are unfamiliar with a source or completeness of data. And so provider education about the data sources and how we validate them is, is absolutely critical to successful use of the data.

Bill Russell: Now you're getting the claims data from trusted sources. But I noticed in your presentation, you talked about some of the issues with claims data. What are some of those issues?

Dr. Jamie Reedy: Yeah, I think there's a number of things that we can highlight. As we've mentioned, claims data is critical to managing financial risk and outcomes.

Bill Russell: Is it easy to address the claims data quality issues?

But, as part of our ongoing partnerships with our health plans, we educate them regularly about how we're using the data to guide better care for their members. And when the health plan understands the importance of the data to the care of their patients, There's so much better cooperation with us.

And then we go back to the health plan and we quickly, work through those issues. And then all of our files are reviewed with respect to EMPI mismatches, and other standard data quality checks, which really helps us discover very consequential changes well in advance before that data gets incorporated into downstream reporting that's used to inform workflows.

First we wanted our data to inform the daily workflows and to immediately impact patient outcomes as I, as I mentioned earlier. Secondly, we really desired to use our expanded data set to give our providers real-time visibility into their performance and opportunities to improve care in the short term, as this would then inform the incentive programs that we put in place for not just our physicians, but for care team members as well.

And this question really led to an assessment of whether we desired to build our own homegrown solution and build the team to manage and maintain the solution versus buying a technology solution that was really custom made for the use cases and priorities that I just mentioned and that we were trying to solve for.

And I know Dr. Parikh's prepared today with a number of examples to show you what we did with our data in those four priority areas.

Bill Russell: Yeah, and I'm looking forward to getting there, but I, if you don't answer the question, I'm going to get a bunch of emails, which is who did you use?

Bill Russell: Fantastic. Okay. That saved me a bunch of emails. And and I appreciate you for the detail that you gave us around the program. And so let's, let's get to. We have this in place. We have this platform in place but there's a lot of different areas we can start going after along that chain to create value. How do you identify the areas to prioritize first?

And as a independent medical group we also have found that, that the more we spend on a patient in the ambulatory setting and the less we spend on the inexpensive care settings, like hospitals and emergency rooms and post acute care facilities, the better the outcomes and lower the total cost of care.

Improve outcomes, which is basically better quality, better patient experience. Appropriately capture disease burden so you have the appropriate cost benchmarks and then reduce costs. But again, if you do that, the outcomes part, the costs will follow. So then we looked at where are our costs opportunities. For example, pharmacy, we know it's one of the largest and fastest growing segments of healthcare expenditure, particularly biologics, infusions, and especially pharmacy.

So we looked at pharmacy. And then we moved on to more ambulatory level types of analytics in terms of annual wellness visits and things like that.

Dr. Ashish Parikh: Absolutely. And, and again, we always lead with our providers and our clinical teams in particular. We lead with the fact that what you're doing is better for your patients and better to get outcomes. But as you said, it never hurts to have your incentives aligned with that.

And so we have this perfect every single specialty, whether you're a dermatologist or pediatrician or behavioral health specialist, but in particular for our primary care physicians, we wanted to really move that one step further and tie a panel-based outcomes and performance and its impact on our value based contracts.

And we took their panels and what we saw, what was the impact of their performance on each of these quality measures on our value-based contracts. So it, it took into account the size of their panel, as well as the difference in performance for their panel compared to the summits overall performance.

Bill Russell: So you guys are doing this across a single EMR, is that correct?

Dr. Ashish Parikh: Yep. Absolutely.

Bill Russell: Yeah. In Southern California, I was asked to do this across a hundred different EMRs and man, what you're describing is really elegant compared to what we were able to do, given the complexity of pulling, pulling all that in and then delivering it back into the workflow. But let's talk about that workflow a little bit.

Dr. Ashish Parikh: Your absolutely right, it's much easier to do this with a single EHR platform, which we're, fortunate to have and was, was the strategy from the beginning, right.

So for our disease burden accuracy, we already knew the codes that were building our EHR. And we could always put a surface those for our providers to, to make sure that they're captured every year and consequently addressed. Right? These are clincial conditions that need to be addressed. If you have a bill from you have to address them.

There's a chance that they're diabetic even going code. So our, our coding compliance team is able to take this combined EHR and claims data, find additional conditions that the patient may have, and then surface them within the workflow of the EHR so that the clinician can then decide, yes, this is a true condition I haven't addressed yet.

Bill Russell: One of the areas in your presentation I found interesting was a skilled nursing facility performance. And you sought to improve that performance. Talk about the role of data in that process.

And with our patients going to, to over a hundred skilled nursing facilities across the state, and even outside of our state, we have many snowbirds from New Jersey and New York that will end up in Florida and Arizona and other places we really wanted to figure out how can we help positively impact that care and those outcomes.

And say look, here's your length of stay for the same DRG as our other partners. And how can we help you improve that so that we can get better outcomes for our patients. And so we were able to develop this, a skilled nursing facility dashboard and it's been so successful that our Katie has actually taken this and made a part of their product that is now available to all of their clients across the network.

One is the site of care, right? So, the same infusion done in a hospital-based ambulatory facility is going to have a significant greater costs than an ambulatory infusion center. And then secondly, we have our own infusion center. So if we're able to identify those patients that are getting infusions outside, getting particularly in hospital-based facilities and move them to our facilities, one, you reduce a cost of care.

There was a lot of overdosing or wastage and so being able to then pull that data in, we were able to identify those opportunities, educate the providers to, to make sure we use the optimal drug as well as the optimal.

Is it just that people are doing AWVs or is it making an impact? And with our analytics platform, we were able to show that patients who got annual wellness visits had multiple benefits. One is the obvious ones, things like they had far better quality gap closure site or care gap closures in some measures up to 40% greater than people that didn't have any wellness visits.

Bill Russell: So I've, I've gotten into a habit of closing my podcast in a weird way. And that's letting you guys have the last word and say, you know what question didn't I ask? What's the close for this? The takeaway that people should have?

Analytics can inform when and how an organization takes on upside and downside risk and provide data that's really critical to effective negotiations. So that's, that's one. Secondly, analytics has allowed us to maximize our organizational investments by identifying the very best opportunities to create new population health initiatives.

So, so none of that would've been possible without the analytics capabilities that we build. So I'll leave you with that.

Dr. Jamie Reedy: They're welcome to contact us directly. I think we're both on LinkedIn with our cell phones and email addresses and we would be happy to talk to folks about this work and share a share of war stories with others.