September 24, 2020

How ProMedica prioritizes data to address social determinants of health

Daily Briefing

    Addressing the social determinants of health (SDOH) within a community is a critical part of any population health strategy—and at ProMedica, it's a priority. In this episode of Radio Advisory, host Rachel (Rae) Woods speaks with ProMedica's Kate Sommerfeld, president of SDOH, and Brian Miller, CMIO, about how their system addresses SDOH, what the business case is for tackling this issue, and why data and analytics are critical to focusing on SDOH.

    Cheat sheet: Social determinants of health data 101

    Read an excerpt of the interview below, and download the episode to hear the full conversation.

    Rachel Woods: I know ProMedica is a very data rich organization, so I'm curious: How are you measuring the ROI of a SDOH strategy?

    Brian Miller: We began asking patients to complete a screening on social determinants of health before meeting their provider for care in 2018, which helped boost our rates of screening completion—and by 2019, our completed screenings hit nearly 40,000.

    So, we partnered with a data analytics group to start to crunch those numbers. Now, we have two methods to do that, because while we have clinically validated methods around the quality of the care and health outcomes, it's harder to get to the cost reduction and whether we're seeing reduced ED utilization, better primary care engagement, and lowered readmission risk.

    ###PLACEHOLDER###

    Now, we're lucky enough to have a fairly large payer piece as one of our business line arms, so we can also do health care payment research. We do deep analytics work for patient population based on screenings and interventions and the patients we can take downstream from that intervention, and then we can bump that up against their internal cost from our payment side, looking at claims and actually tracking the claims reductions stemming from those interventions.

    We also have an analytics tool that we can expand beyond that via AI to essentially create, with clinical data, a cost proxy around clinical burden. And clinical burden corresponds to an average range of what that cost is, so even where we don't necessarily have direct claims, we can get a sense of cost reduction in our downstream interventions—and that in turn helps us understand what interventions work and which ones don't and what message we need to get to our clinicians and third-party payers as we contract around value-based care.

    Woods: Let me ask, to make sure I'm getting that right: Is that AI-based product predicting what the cost reduction for a prospective patient or patient group if they get the interventions that ProMedica is offering?

    Miller: Yes. That is exactly what we want to do with that tool. And going forward, we're going to expand that role.

    You see, from a resource standpoint, SDOH interventions are always challenging, because there's always going to be more need than the ability to deliver to that need. And post-Covid-19, there's a bit of delay on this, right? Obviously huge job losses, evictions that are coming now—a tsunami of SDOH need that we're probably only seeing the very start of, and that we'll be dealing with for months and likely years to come. So it's really important to identify and triage the biggest impact for what that limited resource can deliver to.

    And that's where the AI comes in—the tool works backwards from clinical burden, right, because clinical burden is where most of the cost applies, so we can look at patients with very specific disease burden and then apply a filter backwards around their SDOH need to improve overall costs. So, for example, the analytics tool might tell us that while we're not able to improve their cancer care, we can improve the cost of their cancer care by improving their housing, which snowballs into improving a housing risk for a subset of patients with this disease burden, and from all of that, we could see a $5 or $8 or $10 million reduction in cost.

    And maybe that big bang for the buck is only 300 or 400 patients out of the several thousand might have a housing need, but then we have the capacity to pursue those patients and get them into that intervention and deliver that bigger bang for the buck from a cost reduction on a triage basis. And then downstream, six months later or so, we can look at that to see whether we delivered it and to what percentage? And really hard-metric ourselves on delivering to that KPI.

    Woods: I think a lot of folks would be interested in how you would think about the inputs of something like this, especially now because—as you noted earlier—we're just at the beginning of a tsunami of the aftermath effects of dealing with the pandemic. So from the data perspective, how has Covid-19 changed some of the inputs of this AI platform? Are you able to capture new information that we might not have put at the top of the list six months ago, but because of Covid-19, are new or significantly worse?

    Miller: So we're seeing everything go up around food and housing. And, interestingly but perhaps not immediately intuitively, we are seeing issues around mental health and depression. The anxiety around Covid-19 and the stress and hopelessness is really having an impact, and we're seeing a growing burden climbing from a depression standpoint and that's difficult to address, because providing behavioral health, from a clinical standpoint, is a much tougher, longer lift. Resource to deliver to that need are scares, and like most organizations, we struggle with that.

    Kate Sommerfeld: The one other area that wasn't on our radar before Covid-19 is the digital divide and internet access. We had done a little bit of work there, but the digital divide—access to internet and broadband—certainly an issue for patients and employees as well.

    Woods: Let's talk about each of those, maybe start with the digital divide. What all are you doing to address that challenge in your market?

    Miller: Yea, that one is a big challenge from a technological standpoint, and, as a health care organization, it's really tough for us to deliver to. We do have a fairly large grant that's in process right now to help us identify a digital divide in our clinics and in our ED settings.

    In fact, we have a tool that identifies patients who use the ED very heavily—three or more usages within a 12-month period or less—and then we go screen them for digital need. Our goal is to connect them to tools and to cellular-based broadband and ultimately to a health care app, because Kate's point is a great one: From a digital divide standpoint, the Pandora's box of telehealth is open and will never be closed again. It will become part of the standard of care and it will help us deliver better care to patients who struggle with access. And if they do not have that capacity to have broadband, it's just another way in which we've delivered inequity, so that is another big point for us and we're looking to potentially get that grant within the next month.

    Woods: And that's part of your screening process. I think I heard you say that perhaps one of those 14 points is screening for digital access, is that right?

    Miller: Yes, that's a new one since Covid-19, Kate's exactly right.

    Woods: I think that Covid-19 and even just the kind of national focus on racial injustice has created this moment in which many organizations are waking up to the idea that they should focus on health equity and SDOH. So let's say you're talking to somebody who really is at the very beginning stages, how do you suggest they, with a complete blank slate, prioritize their efforts and figure out where's the right place for them to begin their journey?

    Sommerfeld: Probably one of the easiest things they can do is pull out the community health needs assessments (CHNA) that we're all mandated to do. There's someone in your organization tasked with completing those IRS requirements, so I'd pull out the CHNA, find the individual in your organization who has that responsibility, and really start to look at that data.

    There have been priorities that have been set not only by the health system, but in partnership with your local community partners as well, so take a look at that. I would say bump that up against your strategy, so, for example, if as an organization you're focused on behavioral health, you're going to see that need across the community as well. Figure out what it is as an organization that you're focused on and then go out to your community partners and engage in a conversation.

    Woods: And maybe the other thing I'd add, which is actually something that you said, Brian, is to look whether you have someone internally who is not only tasked with addressing SDOH, but also has the power and the resources to be able to do that. And I'm happy that ProMedica has instituted that, because it's rare that that's actually the case.

    Sommerfeld: Yeah, and I think CEO buy in is also really important—our CEO Randy Ostra is incredibly committed to this work, and that makes a difference. As we talk with other health care systems around the country, the level of engagement from your C-suite and from your executive leader is important. Randy ends up talking to a lot of his peers around what we've been able to see with the cost savings around this. So being able to make the case from both the business perspective to your CFO but also to connect that to the organizational strategy and get CEO buy-in is really important.

    Miller: Kate's right that engagement from the C-suite out drives it internally as well. I think in our organization—to a person—we know that there is a cultural commitment to this. And that really helps to engage the clinical frontend, where we get that initial connection to the patient and where we start to do that screening to get that process going.

    X
    Cookies help us improve your website experience. By using our website, you agree to our use of cookies.