Evaluation of the Next Generation Accountable Care Organization Model

Close up of health care worker in scrubs looking at a tablet with a patient

Assessing CMS’s innovative ACO model to improve outcomes and reduce costs for Medicare beneficiaries

Client

Center for Medicare & Medicaid Innovation within the Centers for Medicare & Medicaid Services
Dates

2016 – 2023

Problem

CMS wanted to determine whether an innovative cost-cutting model was working.

In 2016, the Centers for Medicare & Medicaid Services (CMS) launched its Next Generation Accountable Care Organization (NGACO) model. The NGACO model tested whether strong financial incentives, flexible payment options, and tools to support care management improved value and lowered expenditures for aligned populations of Medicare fee-for-service (FFS) beneficiaries from 2016 to 2021. It allowed ACOs (groups of doctors, hospitals, and other health care providers and suppliers) to keep part of the savings when they were more efficient and improved patient care. In funding the program, Congress required CMS to evaluate whether the NGACO model was working and why. The model was originally scheduled to run from 2016-2020, but due to the pandemic, it was extended for a sixth year.

Solution

NORC’s mixed-methods evaluation assessed the NGACO model’s implementation and impact.

In partnership with the CMS Innovation Center, NORC used a mixed-methods approach to evaluate 62 participant ACOs, 35 of whom remained in the model’s sixth year. Those ACOs comprised over 91,000 practitioners in hospital-affiliated and physician practice ACOs serving 4.2 million traditional Medicare FFS beneficiaries. We evaluated them according to the following key performance indicators:

What approaches work best, and under what circumstances?
What yields the biggest savings while maintaining or improving the quality of care?
How can successful approaches be scaled in other areas of the country?
What other cost-cutting strategies might NGACO participants suggest?

Our evaluation involved analyzing several types of ACO data. We conducted a quantitative analysis of Medicare claims data and surveys of executives and clinicians, as well as qualitative analyses of information from interviews conducted with ACO leaders and staff. NORC created a database of organizational, survey, and interview data to support mixed-methods analyses. Our evaluation used qualitative comparative analysis (QCA) and coincidence analysis (CNA) to test the hypothesis that the different contexts and structures in which NGACOs operate and their implementation approaches create distinct pathways to outcomes in the model.

NORC’s final summative evaluation report evaluated the 35 remaining NGACOs. We analyzed the structural and contextual factors that affected NGACOs’ cost-cutting efforts and ACO responses to the COVID-19 public health emergency (PHE). The infrastructure, partnerships, and resources developed from participation in the NGACO Model better positioned them to respond to needs during the PHE. NGACO leaders noted at the conclusion of the model that the expansion of data analytics to assess risk and target resources was a significant organizational change during the model.

Result

The evaluation produces insights into how ACOs operate in CMS models.

Our evaluation showed that the NGACO model reduced gross total Medicare spending by increasing amounts over time totaling 1.7 billion dollars over six years. However, the model did not save Medicare overall after considering payouts to ACOs. Larger gross spending reductions were associated with ACO physician practice affiliation, election of 100 percent risk level and risk caps greater than 5 percent, and use of population-based payment mechanisms.

Throughout the model, NGACOs reduced utilization in the most intensive care settings and increased the use of preventive care through population health strategies.

NGACOs coordinated care to prevent hospitalizations and ED visits, reducing spending for patients with eight or more chronic conditions.
NGACOs engaged beneficiaries through Annual Wellness Visits (AWVs), increasing AWVs by 21 percent.
Partnerships with skilled nursing facilities (SNFs) to manage post-acute care resulted in decreased SNF spending and days with a modest increase in SNF stays.

NORC’s QCA found that combinations of factors, or pathways, were more influential on spending throughout the model than was any single factor, such as a physician practice affiliated ACO that embedded care management in inpatient hospital settings. The CNA analysis showed that no single condition's presence or absence precluded spending reductions, meaning failure to reduce spending was associated with varying NGACO characteristics over the course of the model, leading to the decision to exit the model.