Algorex and Healthify discuss leveraging predictive analytics and community-level data to inform SDoH interventions
The COVID-19 pandemic has highlighted an acute struggle for community-based organizations (CBOs). The capacity to deliver social services is low, while the demand for services is increasing. Given the current fragility of CBOs, it’s now a necessity for health plans to rework reimbursement models to build sustainable social determinants of health (SDoH) interventions with their community partners.
Predictive analytics can help build that sustainability, according to Manik Bhat, Founder and Chief Strategy Officer of Healthify. Joined by Jacob Luria, Managing Partner at Algorex Healthcare Technologies, Healthify recently hosted a webinar to inform health plans on how to use data to build scalable SDoH intervention strategies.
“One of the biggest things we see health plans and risk-bearing providers struggle with is that it’s really hard to know where to start when thinking of social determinants interventions,” Bhat said. “It’s also hard to understand the long-term financial cases and ROI for addressing these needs.”
Understanding a community’s scope-of-need
When building a robust SDoH program, health plans need to determine the scope of the intervention. To do this, plans must define their population to understand what interventions would be valuable to them.
Traditionally, payors have focused on high-acuity members and fallen short in identifying the nuanced social needs in a community, according to Luria. “To understand the true scope is to move beyond the traditional population health playbook of identifying the top 1% by a clinical risk score and putting those members in a care management program,” he said.
To go beyond the 1% risk score, Luria said to use claims data and existing community-level data sets to build digital profiles. Community-level data can be used to evaluate social needs. These data sets come in many forms, including non-clinical data. Geographical data from the American Community Survey or Census.gov can provide context to broadly understand the community. However, while Luria acknowledged while these data sets help narrow the scope-of-need, the information is only directional.
“It’s one thing to say ‘x is really important,’ but it's another to say with a level of confidence that ‘there are 3,600 individuals that have this need,’” Luria said. Predictive analytics can help health plans get even more granular in their digital profiling by surfacing SDoH needs. By rolling profiles up into specific cohorts, health plans can start to plot programs and identify interested community partners.
From the varied data points, a plan emerges to fit a community’s true scope of need. For example, Algorex identified and targeted more than 4,700 plan members out of 200,000 enrolled members for a BlueCross BlueShield plan when building out an SDoH program using data analytics.
Building the financial case to justify SDoH interventions
Once a population is defined, it’s critical to understand the lifetime financial value of an SDoH intervention. The lifetime financial value is essential for building sustainable SDoH intervention models. The metric also justifies the expense when trying to get organizational buy-in as plans build out these programs.
“The key to long-term financial sustainability is the evaluation of impactable cost,” Luria said. Analytics can be used to build out a potential model. Organizations can place a proxy-value for resource allocation and use profile differentials to build a model. In the webinar, Luria shared a theoretical revenue model built for a food insecure population:
In this hypothetical example, more than 22,000 members out of 100,000 were targeted for a food insecurity intervention. Armed with a figure of $1.6 million in potential revenue opportunity, a health plan would have the justification to launch a food program for these members.
Stream the whole webinar and learn more about how data can be leveraged to build targeted SDoH interventions.