This project will leverage Centers for Medicare & Medicaid Services (CMS) data showing the number of Medicare beneficiaries assigned to Accountable Care Organizations (ACOs), in combination with enrollment in Medicare Advantage (MA) in rural counties, to describe implications for rural health care organizations (HCOs).

The following questions are addressed in this project:

1. How do policy changes in risk-sharing requirements and beneficiary assignment affect growth of ACO presence and percent of beneficiaries in ACOs in rural counties?
2. What is the relationship between changes in the number of beneficiaries assigned to ACOs and presence of local HCOs?
3. What are the trends in rural beneficiary enrollment in the combination of ACOs and MA plans, by county?

Four hypotheses are tested:

1. Time series analyses will show increases in rural beneficiary assignment to ACOs following initiation of advanced payment models.
2. The rate of growth in ACO assignment across rural counties will be higher in relation to proximity to counties with population centers (metro and micro counties), number of Traditional Medicare (TM)-enrolled beneficiaries in the county and presence of primary care.
3. Rural counties with increased ACO-assigned lives are more likely to include Critical Access Hospitals (CAHs) and Rural Health Clinics (RHCs) than those with continued low numbers (i.e., low volume) and percent (i.e., low penetration) of TM beneficiaries assigned to ACOs.
4. The number and percent of beneficiaries assigned to ACOs in each rural county will be positively correlated with the number and percent enrolled in MA plans.

Beneficiary enrollment in MA will be combined with assignment to ACOs to show the total number of beneficiaries by county in the two alternative payment models, as well as the percent of all Medicare enrollees in the county. Beyond testing statistical models that confirm trends seen in descriptive data, analysis will focus on implications for policy and practice. What can be learned about the impacts of policy changes? To evaluate changes due to rule changes, we will use an event study regression approach. We will evaluate three separate sample specifications for this analysis: urban counties, rural counties, and all counties combined. Using this sample of all counties, we will also interact an indicator for whether a county is rural with each year, thereby providing an estimate of year-by-year average differences in the change in assignments for rural counties compared to urban counties. In each of the event study models, we will use appropriate t-tests to show whether changes in outcomes are statistically significant and, in the model, using all counties, whether the changes in rural counties were statistically different compared to the change in urban counties.