The number of Federally Qualified Health Centers (FQHCs) in the United States has grown substantially over the past two decades. In 1996, fewer than 10% of Medicaid enrollees had had contact with an FQHC. This percentage had doubled by 2016, to about 20%. This change occurred since the number of CHC delivery sites in the U.S. increased 51% since 1996, reaching 2,500 sites by 2016. This increase strongly differed across states with the number of delivery sites slightly falling in 15 states, while the number more than doubled in 10 states. Analysis of these data from HRSA shows that Medicaid physician payment rates are negatively correlated with the growth in the number of CHC delivery sites in a state. Prior work (Decker,
Inquiry, 2009) also shows that Medicaid physician payment rates are negatively correlated with the probability that an ambulatory care visit takes place in hospital outpatient department (OPD) and, to a lesser extent, a hospital emergency department (ED). Clearly, state Medicaid policy has a strong effect on place of ambulatory care for Medicaid enrollees, with some states relying much more heavily on care provided in FQHCs and OPDs (and possibly EDs) than others. Since FQHCs and OPDs are paid differently than care provided in physician offices, these decisions are likely to affect costs of care and may also affect quality of care. Sixteen states continued to use cost-based reimbursement methods for OPDs in 2016. The introduction of prospective payment for FQHCs under Medicaid in 2000 likely changed the payment rates for FQHCs compared to other sites over time, though how payment rates compare for similar care across ambulatory care sites is not known.
We analyze data for those continuously enrolled in Medicaid and not Medicare for each calendar year in the Medical Expenditure Panel Survey (MEPS), 1996-2016. We first produce descriptive information following methods in Biener and Selden (Health Affairs, 2017) to calculate average national standardized Medicaid payment rates for ambulatory care visits provided in FQHCs, OPDs and physician offices over time. We will then estimate the effect of FQHC use on the total number of ambulatory care visits, total expenditures on ambulatory care, and measures of quality of care, including those coming from the Consumer Assessment of Healthcare Providers and Systems (CAHPS) measures included in MEPS starting in 2002. Our main independent variable will be whether the beneficiary is an FQHC user. This will be defined as either more than half of ambulatory care contacts take place at an FQHC in a given year, or the beneficiary exceeds a minimum threshold of annual visits to an FQHC. Other independent variables such as age, sex, race, and state and year fixed effects will be included. Since FQHC users are likely different from non-FQHC-users in unobservable ways, we instrument for whether a beneficiary is an FQHC user with their distance to the nearest FQHC using geocoded data for all 1996-2016 HRSA delivery sites and geocoded household addresses in the MEPS.
