Skip to main content

Advanced Search

Advanced Search

Current Filters

Filter your query

Publication Types



Issue Briefs


How the CMS Innovation Center’s Payment and Delivery Reform Models Seek to Address the Drivers of Health

Two women in grocery store look at bagged lettuce

Adrienne Dove (L) and her mother Joanne Dove (R) shop for produce at the Giant food store in Washington, D.C., on May 3, 2019. Screening for food, housing, and transportation issues is important to identifying needs related to the drivers of health (DOH). Photo: Michael S. Williamson/The Washington Post via Getty Images

Adrienne Dove (L) and her mother Joanne Dove (R) shop for produce at the Giant food store in Washington, D.C., on May 3, 2019. Screening for food, housing, and transportation issues is important to identifying needs related to the drivers of health (DOH). Photo: Michael S. Williamson/The Washington Post via Getty Images

  • Participants in CMMI payment and delivery reform models are implementing a variety of approaches to support beneficiaries with social, economic, and location-based needs that drive health

  • Standardized screening and performance measures, along with financial incentives and technical support, could help participants address drivers of health

  • Participants in CMMI payment and delivery reform models are implementing a variety of approaches to support beneficiaries with social, economic, and location-based needs that drive health

  • Standardized screening and performance measures, along with financial incentives and technical support, could help participants address drivers of health



  • Issue: Social, economic, and location-based drivers of health (DOH) have an outsize impact on health — especially for Medicare and Medicaid beneficiaries, who are more likely to be elderly, have low income, or have complex medical needs. The Center for Medicare and Medicaid Innovation (CMMI) addresses DOH in many of their health care payment and care delivery models, but lessons from them have not been synthesized.
  • Goals: Describe how CMMI models have addressed DOH-related needs and identify factors facilitating or hindering participants from addressing these needs.
  • Methods: Review of evaluations, applications, and memos for 40 CMMI models.
  • Key Findings and Conclusions: In 23 of 40 models, participants addressed DOH-related needs to some degree. For some models, CMMI required these activities; in others, participants voluntarily pursued them. Common strategies included screening beneficiaries for nonmedical, social needs and/or referring beneficiaries to social services and community-based organizations. Participants were rarely required to assess or address DOH needs, however, and evaluations, when conducted, were not standardized. Participants noted that greater financial support, financial incentives, and technical assistance would enable them to better address patients’ DOH needs. CMMI might also consider requiring standardized DOH screening and incorporating performance measures in evaluations.


While physical health has biological and environmental determinants, social, economic, and location-based factors account for about 80 percent of health outcomes.1 Known collectively as drivers of health (DOH), these include access to safe housing, reliable transportation, and nutritious meals, among many other influences.2 Because unmet needs related to DOH have been linked to increased hospitalizations, emergency department visits, and overall costs to the health care system, health care providers and payers have become increasingly interested in identifying and addressing the DOH-related needs of patients.3 It has been further demonstrated that investments in programs designed to meet these needs can produce significant returns.4

For more than a decade, the Center for Medicare and Medicaid Innovation (CMMI) — part of the Centers for Medicare and Medicaid Services (CMS) — has tested new, potentially transformative models of paying for and delivering health care. Across models, participants have addressed DOH to varying degrees, with the intent to better support Medicare and Medicaid beneficiaries, who are more likely to be low income, older, and/or have complex medical conditions, placing them at risk for unmet DOH needs, poor clinical outcomes, and high-cost care.5

To date, no studies have systematically assessed these models to glean lessons that might inform future CMMI efforts. Having such information will be particularly important as CMMI implements its strategic vision for the next decade of delivery system reform, which is expected to place a heavy emphasis on addressing social needs as a means of advancing health equity.6

To help fill this knowledge gap, we reviewed evaluation reports and requirements for 40 CMMI care delivery models (see the appendix for details on each one).7 We searched for information on the extent to which CMMI required or encouraged participants to address DOH needs of patients, examples of how model participants addressed DOH needs, and barriers and incentives for doing so.

Key Findings

Of the 40 models we investigated, 23 incorporated DOH activities to some degree. Within eight models, CMMI required participants, which could include hospitals, physicians, health organizations, etc., depending on the model, to screen for or address patients’ DOH needs. We found evidence in an additional 15 models that participants voluntarily addressed DOH to some degree. These participants engaged in efforts to address DOH for many reasons — from improving the overall patient experience to seeking specific outcomes, such as decreasing hospitalizations.

In the 23 models that addressed DOH to some degree, participants implemented various approaches, including screening for unmet DOH needs, and referring beneficiaries to relevant services. Only one model included DOH screening and referral measures in the formal model evaluation, and none included DOH activities in financial incentives or as performance measures.

Our mission is to enhance the health and wellbeing of the people in the communities that we serve ... and we’re particularly interested in any way that we can identify social determinants of health that might create barriers to care.
Medicare Advantage Value-Based Insurance Design Model Participant
(CMS, First Annual Evaluation Report of the Medicare Advantage Value-Based Insurance Design Model Test, 2017, p. 22)

Participant Efforts to Screen for Drivers of Health Needs

We found evidence in 21 models of participants screening beneficiaries to identify their DOH needs, including six models that required participants to do so. In evaluation reports, participants often described the importance of screening as a first step to improving beneficiary health and access to care.

Participants commonly focused on screening for food, housing, and transportation issues. Screening was conducted by physicians, nurses, or social workers, who used questionnaires or held informal conversations with beneficiaries. When using questionnaires, participants occasionally leveraged standardized, validated screening tools, such as the Protocol for Responding to and Assessing Patients’ Assets, Risks, and Experiences (PRAPARE). In most cases, however, screening was unstructured and unstandardized. This inconsistency meant model evaluators and CMMI were generally unable to understand the scale and scope of beneficiary DOH needs — and the impact of these needs on patients and participants — within or across models. Lack of standardized screening by participants is likely because of lack of clear screening requirements by CMMI or inclusion of screening results in evaluation design.

Participants used screening results in various ways. Some participants leveraged results when developing patient care plans (for example, specifying referral to social services or addressing DOH as care objectives) while others incorporated the results into risk-stratification approaches to determine the appropriate level of care for beneficiaries. Only in one model, Accountable Health Communities (AHC), did participants formally track the DOH needs of their patient population over time, which CMMI required (see box for the AHC model description).

Lessons from CMMI’s Accountable Health Communities Model

Accountable Health Communities was CMMI’s only model designed to test whether addressing DOH-related needs could improve the health of Medicare and Medicaid beneficiaries while reducing health care utilization and costs. CMMI provided technical assistance in the form of written guides and meetings to assist organizations in addressing DOH needs.8

Participants in AHC screened beneficiaries before and after visits using a standardized tool that could be tailored to each participant’s population. In the first year, evaluators found that among the nearly half a million beneficiaries who were screened through the model, 34 percent had at least one DOH need. The year 1 evaluation also revealed that location of screening — whether an emergency room, primary care clinic, or other setting — could impact screening completion and how likely a beneficiary was to report any DOH needs. After screening, beneficiaries with unmet needs were referred to community-based organizations (CBOs) and social services. Staff worked with beneficiaries to create care plans based on their specific needs and set up introductory meetings between beneficiaries and CBOs.

AHC presented valuable lessons on addressing DOH needs and challenges. Staff experienced higher-than-expected caseloads; nearly twice as many beneficiaries as expected used the navigation services. Evaluators found that 14 percent of beneficiaries who were eligible for and opted to receive navigation services had at least one need resolved at the end of the first year, but they noted that lack of community resources was a significant barrier to resolving needs. In addition, screening results were not systematically shared with physicians, impeding their ability to participate in monitoring and addressing DOH needs.

We need to understand [which] community-based organizations are willing to help patients in the right way.
Bundled Payments for Care Improvement Advanced Model Participant
(CMS, CMS Bundled Payments for Care Improvement Advanced Model: Year 2 Evaluation Report, 2021, p. 36)

Referral to Social Services to Address DOH Needs

We found evidence of participants in 20 models connecting beneficiaries to social services and CBOs to address identified DOH needs, five of which were required to do so. Participants commonly engaged multidisciplinary teams, including peer counselors, community health workers, and social workers to identify available services and facilitate referrals for beneficiaries.

Participants often developed and maintained directories of social supports in the community. In two state initiatives, the Vermont All-Payer and the Maryland Total Cost of Care models, participants used digital care coordination and navigation platforms shared across hospitals, primary care providers, and CBOs to support and track social service referrals. To improve referral efforts, some participants convened learning collaboratives with CBOs and other stakeholders, while others leveraged existing community relationships or built new relationships.

Referring beneficiaries to social services did not come without challenges. Digital care coordination tools could be limited or difficult to use. Participants using the Vermont All-Payer online platform, for instance, reported that it lacked interoperability with their existing electronic health record (EHR) systems.

In addition, rarely was there either a formal or informal feedback loop in place, whether EHR-based, email-based, or phone-based, to ensure that CBOs delivered services and that beneficiaries’ needs were met. Without a feedback loop built into navigation efforts, participants may be unaware of which social services are no longer able to adequately address beneficiaries’ needs. AHC participants also reported that feedback loops promoted morale by letting staff know about referral “success stories.”

It is easier to pay for a $50 cab ride than an expensive hospital stay.
Comprehensive End-Stage Renal Disease (ESRD) Care (CEC) Model Participant
(CMS, Comprehensive End­Stage Renal Disease Care (CEC) Model: Performance Year 1 Annual Evaluation Report, 2017, p. 25)

A further challenge was the substantial staff time it took to identify social services that could appropriately address DOH needs. And social service organizations often had limited capacity to serve beneficiaries with unmet needs.

In some cases, the model’s design enabled participants to deliver services directly to beneficiaries instead of connecting to CBOs. For example, the Comprehensive End-Stage Renal Disease (ESRD) Care (CEC) model allowed participants to use waivers to cover patients’ transportation and nutrition services related to dialysis.

Models That Did Not Address DOH-Related Needs

Not all models required participants to address DOH, and many participants did not have the capacity to do so. In 17 of the 40 models we reviewed, we found no evidence of participants addressing DOH needs. In evaluation reports, participants often cited their inability to address patients’ DOH needs as a barrier to achieving model goals like improving patients’ access and quality of care.

In these models, participants described the lack of direct financial support from CMMI as a barrier to addressing DOH. Some participants could not afford to hire the staff necessary to support screening or navigation efforts, or they lacked adequate financial resources to cover the expense of directly offering services like providing transportation or covering an appointment with a nutritionist.

Our biggest concerns [with the model] were the social barriers for our patients. It’s easier when people you’re sending home have a family member who can take a week off work to be a caretaker, or even just have running water. Having to look at it from the perspective of our patients who don’t necessarily have those things makes everything look different. Social determinants of health need to be recognized in these models.
Comprehensive Care for Joint Replacement Initiative Participant
(CMS, CMS Comprehensive Care for Joint Replacement Model: Performance Year 1 Evaluation Report, 2018, p. 60)

Further, in many models, financial incentives and other requirements focused on a limited set of quality measures that did not include DOH. Without incentives, participants would be less likely to prioritize these needs, particularly while they implemented new care delivery approaches, potentially within a short period. For example, participants in the Federally Qualified Health Centers (FQHC) Advanced Primary Care Demonstration model said they had to rapidly learn new ways of working and delivering care, leaving even less time and capacity to address patients’ DOH needs. Many participants noted that additional CMMI learning collaboratives, trainings, or other technical assistance would have helped them design and implement DOH-related activities.

In some cases, participants expressed concern about unintended consequences for beneficiaries with unmet DOH needs. For example, participants in the Comprehensive Care for Joint Replacement Initiative felt that the model encouraged earlier discharge to the home, even for those who lacked safe and reliable housing or caregiver support. Providers often felt unequipped to manage such issues.

Policy Implications

As CMMI implements its 10-year strategic vision and seeks to advance health equity, it will be critical to address beneficiaries’ DOH-related needs within payment and care delivery reform models. Based on lessons from the past decade of experimentation, CMMI can consider the following steps to enable model participants to better address these needs:

Standardize DOH screening and measures. CMMI has stated that some, but not all, future model participants will be required to collect data on beneficiaries’ health-related social needs.9 Ensuring that data collection approaches, screening tools, and measurement are standardized and consistently applied within and across models over time could improve identification of DOH needs as well as model evaluation. CMS is currently considering including two DOH measures derived from the Accountable Health Communities model across Medicare programs to enable this standardization.10 CMMI has stated it will encourage participants in the Accountable Care Organization Realizing Equity, Access, and Community Health (ACO REACH) model to collect beneficiary DOH data; this is an opportunity to leverage the new screening measures and test new methods of screening and measurement.11

Build DOH into model evaluations. Standardized data collection and measurement can be used in model evaluation to compare how models affect DOH for different populations, identify the model’s unintended consequences for beneficiaries’ needs, determine how beneficiaries’ unmet needs impact model performance, and design efforts that better meet these needs. However, this will require building DOH objectives into model design and incorporating DOH into evaluation designs from the start.

Strengthen incentives and financial support. CMMI could consider offering direct funding to support participants’ efforts to address DOH needs. This could include funding to hire staff to identify CBO partners and create feedback loops as well as designing explicit financial and nonfinancial incentives. In addition, CMMI could require participants to describe how they will address beneficiaries’ DOH needs at the beginning of model performance periods to understand participants’ capabilities; similar requirements related to health equity are part of the new ACO REACH model.

Offer technical assistance. Given the range of model participants’ abilities to address social needs, technical assistance that includes learning collaboratives and structured trainings could help participants identify and apply promising approaches, such as best practices for collecting and using data on patients’ DOH needs, developing partnerships with CBOs, and implementing interventions. Given the returns that investing in services targeting DOH needs can yield for health and social organizations, CMMI also could share this evidence with participants to support their efforts to engage with CBOs.12


The Center for Medicare and Medicaid Innovation has made it a top priority to address DOH-related needs, recognizing that doing so is critical to patient health and is a lever for advancing health equity. Participants in CMMI care models understand the importance of DOH initiatives but require greater support. By heeding lessons learned during the past 10 years, we hope CMMI can develop new models that ensure all beneficiaries have access to high-quality, affordable, and equitable care.


We first reviewed all Center for Medicare and Medicaid Innovation (CMMI) models and identified those that were implemented, were major reforms, and could reasonably be expected to address drivers of health (DOH) through care delivery by the end of 2021; 40 models met this criterion. For each of the selected models, we reviewed the available evaluation reports and applicable supporting documents, such as applications and case studies, to understand how model participants addressed DOH needs. To streamline our document review, we used the following search terms: social, community, determinants of health, social needs, nonmedical, food, nutrition, housing, transportation, screen, health-related social needs, social worker, and community health worker.

We pulled all relevant information on efforts to address DOH and categorized models based on activities conducted and requirements for addressing DOH into three categories: no evidence of model and/or participants addressing DOH; evidence of participants addressing DOH through screening and/or navigation services; and CMMI formally required DOH be addressed as part of the model. We then summarized DOH-related activities in the appendix.

  1. David Lakey, Elena Marks, and Eileen Nehme, “Finding Effective Ways to Address Social Determinants of Health,” Health Affairs Forefront (blog), Apr. 22, 2021.
  2. John R. Lumpkin et al., “What We Need to Be Healthy — and How to Talk About It,” Health Affairs Forefront (blog), May 3, 2021.
  3. Paula Braveman and Laura Gottlieb, “The Social Determinants of Health: It’s Time to Consider the Causes of the Causes,” Public Health Reports 129, no. 1, suppl. 2 (Jan.–Feb. 2014): 19–31.
  4. Mekdes Tsega et al., ROI Calculator for Partnerships to Address the Social Determinants of Health: Review of Evidence for Health-Related Social Needs Interventions (Commonwealth Fund, July 2019).
  5. Centers for Medicare and Medicaid Services, “CMS Issues New Roadmap for States to Address the Social Determinants of Health to Improve Outcomes, Lower Costs, Support State Value-Based Care Strategies,” press release, Jan. 7, 2021.
  6. Centers for Medicare and Medicaid Services, Innovation Center Strategy Refresh (CMS, 2021).
  7. Centers for Medicare and Medicaid Services, “Innovation Models,” n.d.
  8. Centers for Medicare and Medicaid Services, A Guide to Using the Accountable Health Communities Health-Related Social Needs Screening Tool: Promising Practices and Key Insights (CMS, June 2021).
  9. CMS, Innovation Center Strategy Refresh, 2021.
  10. Debbie I. Chang and Rachel Nuzum, “Now Is the Time for Measuring Social Drivers of Health in Medicare, Medicaid, and the Children’s Health Insurance Program,” To the Point (blog), Commonwealth Fund, Mar. 28, 2022.
  11. Centers for Medicare and Medicaid Services, “ACO Realizing Equity, Access, and Community Health (REACH) Model Health Equity Updates Webinar,” Apr. 5, 2022.
  12. Tsega et al., ROI Calculator for Partnerships, 2019.

Publication Details



Celli Horstman, Senior Research Associate, Delivery System Reform, The Commonwealth Fund

[email protected]


Celli Horstman, Alexandra Bryan, and Corinne Lewis, How the CMS Innovation Center’s Payment and Delivery Reform Models Seek to Address the Drivers of Health (Commonwealth Fund, Aug. 2022).