Federally-Qualified Health Centers and the Use of and Costs of Care for Medicaid

Federally-Qualified Health Centers (FQHCs) serve many uninsured patients as well as those on Medicaid.  Although some previous work has compared the use of health care and some measures of quality of health care among FQHC users and non-users, very little has addressed possible unobserved differences between users and non-users and none considers any differences in costs of care.   Even if patterns of use of health care are similar for Medicaid recipients who are FQHCs users and non-users, costs of care may be different since Medicaid pays FQHCs differently than it pays other providers.

This paper estimates the impact of FQHC availability on the use of and costs of health care among non-elderly individuals on Medicaid.  Data from Medicaid Medicaid Analytic eXtract (MAX) claims for 1999-2007 for those under age 65 and in the National Health Interview Survey (NHIS) 1994-2005 are used.   Differences in personal characteristics from the NHIS, self-reported health conditions from the NHIS, and Medicaid spending and  measures of the use of health services and some simple measures of quality of care (such as ambulatory care sensitive hospitalizations) from Medicaid claims are first compared for FQHC users and non-users.   Then differences in the use of and quality of care and in total costs of care are analyzed controlling for other factors including measures of education and family income available from the NHIS, plus state and year fixed effects.   Since FQHC placement may also vary with the level of need in a community in unobservable ways, some models  instrument for FQHC use with distance between an individual Medicaid recipient’s place of residence (exact latitude and longitude from newly-geocoded data from the NHIS) and the nearest FQHC (latitude and longitude of FQHCs in existence for 1999-2007).  The growth in the number of FQHCs between 1999 and 2007 yields in substantial over time variation in distance to FQHCs within states. 

Compared to non-FQHC users, FQHC users are younger and more likely to be female, non-Hispanic black compared to white, and less likely to be married or to have completed high school. Adjusted for age, they are more likely to have certain health conditions (asthma, hyperlipidemia, hypertension, congestive heart failure) and less likely to have others (arthritis, cancer, ischemic heart disease). They have statistically significantly higher Medicaid expenditures and more physician visits, visits to hospital emergency rooms, and inpatient stays (including those due to ambulatory care sensitive conditions) compared to Medicaid recipients who are not FQHC users.   The instrument consisting of distance-to-the-nearest FQHC predicts well in the first-stage, with an F-statistic of about 50 in models controlling for  observed personal characteristics and area and year fixed effects.  Preliminary second-state estimates no longer find statistically significant differences in measures of the use of health care services including hospitalizations due to ambulatory care sensitive conditions but still find statistically significantly higher total Medicaid expenditures for those using FQHCs.   Future models may replace area fixed effects with person fixed effects since those who use FQHCs may differ from others in unobservable ways within areas.