Report: Accountable Health Communities Led to Savings, Addressed SDOH

An evaluation report released by the Center for Medicare and Medicaid Innovation found that the Accountable Health Communities (AHC) model, launched in 2017, demonstrated that focusing on patient’s needs related to upstream drivers of health can lead to cost savings while maintaining or improving the quality of care beneficiaries receive, as evidenced by reductions in inpatient and emergency department (ED) utilization.

The ACH model was designed to serve beneficiaries with needs related to upstream drivers of health, including housing instability, food insecurity, transportation problems, utility difficulties, and interpersonal violence. The model tested whether connecting these beneficiaries to community resources could reduce healthcare expenditures and utilization.

Twenty-eight participants, known as bridge organizations, collaborated with clinical partners, community-based organizations, state Medicaid agencies, and other stakeholders. Model participants universally screened all Medicaid and Medicare beneficiaries who received care from clinical partners for five upstream drivers of health, also referred to in the report as “core needs.” Those who were identified as having at least one of these needs were universally referred to community-based organizations. 

When the model concluded in 2023, participants had screened more than 1 million Medicaid and Medicare beneficiaries for core needs. Of those screened, 37% screened positive for at least one core need, and 18% also reported at least two ED visits in the past 12 months and were community-dwelling, making them eligible to receive navigation services. Navigation services helped connect beneficiaries with community-based organizations who could address their needs.

The report, submitted by RTI International, describes two AHC Model tracks. In one track, the model provided navigation services to support those who needed help finding community resources. In the other track, the model provided these same navigation services while strengthening the relationships between clinical partners and community-based organizations. The two tracks were known as the Assistance and Alignment Tracks:

• Assistance Track: Navigation-eligible beneficiaries in the Assistance Track were randomly assigned to an intervention or control group. Beneficiaries assigned to the intervention group received their usual clinical care, a community referral summary with a list of community resources available for their specific needs, and an offer of navigation services. Beneficiaries in the control group received everything beneficiaries in the intervention group received except for an offer of navigation services.

The Assistance Track tested universal screening and referral to identify Medicaid and Medicare beneficiaries with core needs and refer them to services, adding navigation assistance to connect eligible beneficiaries to the community services they needed.

• Alignment Track: Navigation-eligible beneficiaries in the Alignment Track were not randomized, so all beneficiaries received the same intervention as the Assistance Track intervention group. In addition to the beneficiary-level intervention, Alignment Track bridge organizations performed a variety of community-level activities, such as community-level continuous quality improvement. The Alignment Track tested universal screening, referral, and navigation combined with engaging key stakeholders in community-level continuous quality improvement to align community service capacity with the community’s service needs.

Key findings

Among the key findings of the report were that navigation-eligible Medicaid and FFS Medicare beneficiaries in the Assistance Track intervention group had lower total healthcare expenditures than beneficiaries randomized to the control group.

In the Assistance Track, Medicaid beneficiaries had a 3% reduction in total expenditures and FFS Medicare beneficiaries had a 4% reduction in total expenditures. In the Alignment Track, Medicaid beneficiaries had a 7% reduction in total expenditures. 

Medicaid beneficiaries in the intervention group also had lower inpatient admissions and unplanned readmissions relative to the control group, indicating that reduced inpatient use—including unplanned readmissions—was a key driver of the lower observed expenditures among Medicaid beneficiaries. 

FFS Medicare beneficiaries in the intervention group had lower ambulatory care sensitive condition (ACSC) admissions and ED visits, suggesting that use of other emergent services drove lower expenditures among FFS Medicare beneficiaries, the report noted. 

Navigation-eligible Medicaid beneficiaries in the Alignment Track also had lower total healthcare expenditures relative to the comparison group. Lower inpatient admissions and ED visits may have driven the observed reduction in total healthcare expenditures for Medicaid beneficiaries. The analysts did not find significant impacts for most outcomes among FFS Medicare beneficiaries in the Alignment Track. A set of complementary analyses that aimed to overcome this limitation suggests that the lack of significance was primarily because the sample size was too small to detect impacts, not because the model was ineffective in this population. 

Moreover, many of the impact estimates among FFS Medicare beneficiaries in the Alignment Track were in the same direction and had a similar magnitude as those observed among FFS Medicare beneficiaries in the Assistance Track.

Across both tracks and payers (i.e., Medicaid and FFS Medicare), the AHC Model generated net savings of more than $200 million. AHC was associated with lower expenditures and hospital-based utilization in both the Medicaid and Medicare intervention groups

The report highlighted some mechanisms that may have led to AHC’s successes:

● Navigators built trust with beneficiaries. Having a trusted relationship with a navigator may have increased trust in the healthcare system overall. This could have led to better connection to the healthcare system and associated providers making care more effective for beneficiaries.

●  Navigators may have had direct impacts on healthcare utilization. Interviews with navigators revealed that navigators often went above and beyond in helping their clients. In some cases, navigators would help their clients remember important appointments for ongoing treatment, such as for mental health conditions. Thus, one mechanism that could explain the observed impacts is that navigators leveraged their relationships with clients to help them better navigate the healthcare system, and better access to care led to improvements in acute healthcare outcomes.

● Need resolution could have been more likely among certain subpopulations of those navigated. Results from the Third Evaluation Report showed that certain subpopulations of beneficiaries, such as those with chronic conditions, were more likely to have their needs resolved than others. Thus, navigation could have had more impact on need resolution for these subpopulations.

The report also noted that navigation services were more impactful when provided alongside other Medicare alternative payment models. One possible explanation for this finding is that some beneficiaries also in APMs received enhanced care management services alongside the navigation services provided through the AHC Model.

In addition, qualitative data collected for this evaluation suggest that navigators performed complementary functions as a care manager or care coordinator for their clients (e.g., providing appointment reminders for medical appointments) in addition to their work connecting beneficiaries to community-based organizations. These care management and care coordination services combined with navigation services may have had synergistic effects on the cost and utilization outcomes.

In its conclusion, the report notes that the AHC Model demonstrated that it is possible to screen for upstream drivers of health on a large scale and that screening can be integrated into the clinical workflow in flexible, patient-centered ways. The model also showed that core needs are prevalent among Medicaid and Medicare beneficiaries, and that most beneficiaries with core needs are receptive to navigation services. The lessons learned from the report can help as screening and navigation efforts that started under the AHC Model continue and expand throughout the healthcare system and communities across the country.

One example of a bridge organization was Camden Coalition in New Jersey. Its model focused on the needs of Medicare and Medicaid beneficiaries living in Camden, Burlington, and Gloucester counties who sought healthcare services at participating clinical delivery sites. Our clinical delivery site partners were Cooper University Health Care, Jefferson Health, Virtua Health, CAMcare, and Oaks Integrated Care.

The Camden Coalition, through partnerships with clinical and community service providers serving these counties, implemented four key elements of the Accountable Health Communities model:
● Screening;
● Referral;
● Community navigation services; and
● Regional partner alignment.

By systematically identifying and addressing the health-related social needs of Medicare and Medicaid beneficiaries with participating providers, Camden Coalition sought to reduce healthcare costs and utilization, and improve the health of all patients across these three counties.