‘Win-Win’ Payer Strategies for Predicting Risk and Supporting Health Equity

The widespread cancellation or postponement of elective procedures and medical visits has led to long-term risks associated with Star Ratings and the accuracy of predictive risk adjustment models. Because fewer encounters mean less population data, inaccurate predictions of member risks could drive negative financial and clinical impacts when utilization returns to typical levels.

Effective capture and analysis of member data is highly relevant for payers focused on meeting equity-related member needs.

Notably, these pandemic-related complications with member data co-occur at a time of unparalleled need to address social determinants of health (SDoH). Gaps in SDoH due to the pandemic and social distancing measures have gravely impacted diverse populations, including:

• Low to moderate-income and under-insured essential workers
• Medicare beneficiaries, especially those with chronic conditions
• Individuals who are socially isolated and need regular care
• People who live in overcrowded housing environments, which is common in low income areas

Low income individuals are particularly at risk of SDoH-related burdens, posing high member costs that impact both Medicaid and Medicare populations. For example, Medicare Advantage payers face unique financial challenges, as most states do not cover members beyond 135% of the federal poverty line, yet roughly 32% of Medicare Advantage members are <200% of the federal poverty line.

Payers are struggling to target wide-ranging and diverse needs—underscoring the criticality of data analysis to formulate a SDoH strategy.

Whether it is due to SDoH or the underutilization of care during the pandemic, inefficient data on members through diagnosis codes, procedure codes, revenue codes, or claims information may project a member to be less risky than they are. The resulting half-baked risk adjustment model leaves payers with a lack of confidence that they have fully captured and are being accurately compensated for the full burden of risk for their panel.

To optimize federal reimbursement, payers should include social risk factors in their risk adjustment models.

There are several win-win strategies to better position payers to accurately predict risk and meet their members’ holistic needs—both critical to maintaining or improving a program’s Star Ratings.

The Centers for Medicare & Medicaid Services (CMS) is increasingly pushing plans to holistically address the needs of low income and vulnerable members. Plans that do not take steps to better address SDoH in risk adjustment and clinical response place CMS Star Ratings at risk, potentially impacting membership retention and growth, as well as their quality bonus.

While payers may not yet be positioned to harness social data, there are several datasets that can help to incorporate social risks into risk adjustment, stratification, clinical response, and total cost of care, including:

• Patient claims and utilization data
• Community health needs assessments (often performed by health systems or public health agencies who may be able to offer raw data)
• Public health/municipal government studies
• Insights from historical Affordable Care Act-required community benefit spend (nonprofit hospitals)
• Stakeholder input/patient satisfaction surveys
• National data repositories, including the Census, American Community Survey, and other third-party data sources

Three Approaches to Integrating Social Risk Factors into Risk Adjustment

1. Repurpose Existing Administrative Data on SDoH
   
   Minnesota updated its state Medicaid program’s accountable care organization model by using beneficiary-level claims and administrative records that adjusted payments based on social risk factors. Payments are adjusted based on social risk factors identified through Medicaid eligibility data, self-reported information, or analyses of members’ data (such as using addresses to identify homelessness). Although Medicaid-focused, this approach may be applied to Medicare plans as well.
   
   
2. Use Survey Data to Identify Social Risk Factors
   
   The nonprofit organization MN Community Measurement (MNCM) created a risk adjustment model by pairing Z-code data from member records to data from the US Census Bureau’s American Community Survey (ACS). MNCM identified social risk factors associated with Z-codes through the ACS. This methodology could also be duplicated by Medicare plans because it uses public data.
   
   
3. Collect New Data on Social Risk Factors through Billing Z-Codes
   
   Massachusetts used various data elements from diverse datasets, including claims data, administrative data from their state Medicaid agency (e.g., data on housing instability or disability status), and the ACS to update their risk adjustment model. However, Massachusetts also began collecting new data on social risk factors by creating ICD-10 Z-codes for homelessness. Additionally, United Healthcare and the American Medical Association jointly developed billing codes for SDoH. Together, they created more than 24 new ICD-10 codes to better characterize SDoH barriers members face, including transportation, social isolation, and food insecurity.

Holistic risk adjustment will improve the accuracy of risk scores to promote higher CMS Star Ratings and optimal reimbursement.

In the pandemic-era, payers have a prime opportunity to strategically expand benefits in a data-driven way to seize long-term opportunities that more comprehensively serve members’ whole-person needs.

Lessons learned from this strategic expansion can build momentum that Medicare Advantage programs are already able to capitalize on, due to CHRONIC Care Act provisions allowing nonclinical services for community-based care. These changes are critical to address consumer demand, can be a market differentiator for Medicare Advantage plans, and position Medicare to better support lagging state budgets that could lead to Medicaid benefit cuts—impacting dual eligibles.

It's time to fold social determinants into risk-adjustment strategies and member assessment practices to better meet individual needs.

Doing so will mutually benefit payers, federal and state government, and most importantly—our nation’s Medicare and Medicaid beneficiaries who have been disproportionately impacted by COVID-19 and a range of inequities for far too long.

Co-authored by Jason Gerling and Brinda Gupta