Social Determinants of Health and Patient Identification
In December 2017, the National Quality Foundation (NQF) published a report, funded by the Centers for Medicare and Medicaid Services (CMS) titled: “A Framework for Medicaid Programs to Address Social Determinants of Health: Food Insecurity and Housing Instability.” In publishing this report, NQF convened a panel of experts and conducted a literature review. They ultimately determined that Medicaid policymakers should prioritize addressing certain social determinants of health (SDOH) to better optimize health outcomes for the Medicaid population.
The report’s focus was on housing stability and food insecurity, with Medicaid as the primary driver. The report states: “Food insecurity and housing instability were selected as key areas for which Medicaid programs can support data collection efforts in the short term.” It was not clear in the report if food insecurity and housing stability were selected prior to convening the experts panel and literature review, but it provokes the question: Were other social determinants of health considered for evaluation, too? Intuitively, it seems that meeting these foundational needs would have a bigger impact on health outcomes. A more empirical evaluation of other levers — such as transportation, employment, or public health interventions — may have yielded different priorities.
Nonetheless, the effort to examine food insecurity and housing instability is a great start. The report covers several recommendations for administrative and clinical data sources that could be combined and analyzed to increase the likelihood for better health outcomes. What the report does not cover in any detail is how these sources get linked to make conclusions. For example, some clinical and administrative data may be identified and some may be anonymized or aggregated, but drawing conclusions on individuals’ experiences means creating probable linkages — like a unique identifier or a series of common data points — between the data sets. This is a hard problem.
Some states are pursuing Master Patient Index (MPI) systems to track individuals or families through multiple systems with more certainty. SDOH is just one area of potential benefit from combining data sets with Medicaid data. Other examples include patient/provider attribution for outcomes-based reimbursement models, predictive financial modeling, and eligibility churn between different programs and payers. (Note: “Patient” is not always the preferred term when referring to Medicaid beneficiaries, since some people are not actively engaged in care that would make them a patient, but their data is still important.)
MPI systems involve many complex issues that make them hard to build and use. It is necessary to have a steward and a governance process that reflect the goals and priorities of all entities with data. For example, if a state Medicaid agency is the steward, then Medicaid’s goals and priorities should be prioritized. However, data from other agencies — such as housing, public health, and nutrition programs — is also needed to make meaningful conclusions. With so many competing priorities, it can be a challenge to negotiate stewardship and governance. Related, these systems cost money — a lot of money. Establishing an upfront investment, as well as a financial sustainability model, can be a challenge.
States interested in pursuing a strategy to uniquely identify individuals, analyze their SDOH, and track their health status or outcomes may wish to evaluate the state’s existing systems that accomplish some or all of these goals. The state should conduct a feasibility analysis for the governance, design, development, and subsequent maintenance of a system that meets these goals. Finally, an important feature to the feasibility study would be an analysis of funding opportunities.