Social Determinants Meet the Business Case

An analytics firm tests whether SDOH data will improve outcomes and save costs in Baltimore.

This article appears in the November/December 2019 edition of HealthLeaders magazine. 

Social determinants of health are the topic that forward-thinking health systems like to discuss in industry forums, but on the ground level, trying to wrap a clinical mission around SDOH has been like trying to hug a cloud of steam. Data analytics may finally be the answer to moving SDOH out of the realm of community altruism into real business strategy.

One firm is trying to push the industry in that direction. Washington, D.C.–based analytics firm Socially Determined blends together clinical, financial, and social data to give health systems a more accurate view of where opportunities are to improve the health of their communities and have a positive impact on the healthcare cost curve.

Earlier this summer, Socially Determined was the winner of the inaugural LifeBridge Health and CareFirst BlueCross BlueShield "Innovation Challenge," receiving $50,000 and the opportunity to partner with LifeBridge Health and CareFirst to use its platform to look at SDOH risk and opportunities in the Baltimore area.

SDOH is not just a buzzword. For many health systems, finally having the tools to spot opportunities could help them in populations where they hold risk. In some cases, "upstream" interventions may be a more effective and less expensive intervention than the clinical interventions that many health systems have tried with inconsistent results.

HealthLeaders recently talked with Socially Determined CEO Trenor Williams, MD, to learn more.

HealthLeaders: Health system leaders are trying to understand social determinants in an actionable way. There is a lot of data coming at them. How do you really translate it to the point where they can align the clinical enterprise in an effective way?

Trenor Williams: Access to data is not the problem. Healthcare organizations are so awash in data that it’s hard to glean insights from it. In most health systems, the internal analytics team is trying to help solve and answer questions for every part of the organization. And while the leadership team may have interest in social determinants, it is rarely somebody’s full-time day job and certainly not a team’s full-time job. We aggregate more than 130 open datasets of commercial data from multiple vendors and combine that with clinical and claims data. It takes domain knowledge and focus in this area, and we aim to become that de facto analytic partner for the organization.

HL: So, you get to scan multiple health systems looking for areas of risk. Any patterns emerge?

Williams: We would say three populations have garnered the most interest. One is managed Medicaid. Depending on the state you're in and the contracts you have for that population, there's a lot of opportunity to thoughtfully evaluate the patients and invest in your organization as well as your community partners. Next would be Medicare Advantage, as it relates to identifying and closing care gaps. And third, those folks who self-insure their own employees have an amazing opportunity to step back and look. Many of those employees may be fragile from a socioeconomic standpoint. They'd have real risks and were just not aware.

HL: I think one of the challenges for a provider executive is how far upstream to go.

Williams: Exactly. A great example from one of our client partners, ProMedica, is their Market on the Green. Before they built the grocery store, the uptown neighborhood of Toledo was considered a food desert. They looked at the financial opportunity of addressing food insecurity for at-risk individuals and found that there was a significant per member per month cost differential. Looking at subsegments of the populations with real risk, it’s in the provider’s best interest—from both a business standpoint and an altruistic perspective—to look at those levers that drive costs and find a way to intervene further upstream. And by building a grocery store and referring at-risk patients, they are making a big impact on the health of the pediatric diabetic population.

HL: How granular do you have to go?

Williams: Looking at the work we do with ProMedica and their managed Medicaid population, we assessed the impact of housing instability in children who had asthma. We found a fourfold increase in costs for children that live in areas of high housing instability versus those that live in low-risk areas. And that intervention strategy did not include finding new housing. Instead it was remediation for upper respiratory disease, which included replacing things like carpets and air conditioning. That's the kind of manageable referrals that work in medical/legal partnerships. We see many similar examples around the country that are right in the scope of work health systems can do. It has a demonstrable impact to the clinical outcomes for those kids as well as a financial impact for the health system.

HL: How does a health system get down to that level of the data?

Williams: We start with an opportunity analysis where we're combining open public data, commercial data, and the health system’s own data. We do a top 20 list of opportunities, so think of housing instability and asthma, or food insecurity and diabetes, or transportation or anxiety. We build this top 20 opportunities [list] across the entire population, and then do an opportunity analysis of different cohorts of patients a health system can act on. And then that becomes a decision the organization makes, based on the other priorities that they are looking at. Maybe the existing programs that they have will work or we’ll need to look at new interventions. Everybody has limited bandwidth, so this now becomes the new data-driven way to look at the entire population and narrow it to a few places.

HL: And you can then narrow it down to a patient level?

Williams: We can feed both the community-risk information and the person-level information into their systems. If you're in an Epic shop, like some of our customer base is, we can feed that into a care management system. So, if you're a care manager now looking at a patient that is in front of you today, having some of that background information about what's going on in their community may help support the conversation.

HL: When you map the areas of risk in a community, how often are the health system leaders surprised by what they see?

Williams: I do think that just visualizing the geospatial maps, it's amazing to me how much that resonates with people because it's sometimes the first time they've ever thought about it. When we lay our patients or members on top of the areas of risk, it's unbelievable to see. We were showing economic risk for a client in South Carolina and then we layered their existing clinics on the map. There were large pockets where their members lived in areas of high economic risks but were more than five miles from clinics. And if they built two new clinics in areas we recommended, almost 100% of that population would be within one to three miles of one of the clinics. That visualization creates the connection.

HL: Are we seeing a shift in how leaders approach social determinants?

Williams: Yes. We are seeing a migration from believing that this is the altruistic part of what the foundation of a health system does to becoming an integral part of how they think about their business. Again, we have more data than we can use today. It is that idea that data is insightful and actionable—it’s what every healthcare organization should expect.