Article ID: | iaor201111703 |
Volume: | 14 |
Issue: | 4 |
Start Page Number: | 314 |
End Page Number: | 323 |
Publication Date: | Nov 2011 |
Journal: | Health Care Management Science |
Authors: | Koizumi Naoru, Rothbard B, Smith E, Mayer D |
Keywords: | statistics: regression, statistics: inference |
A discrete‐choice logit model was applied to study the determinants of mental health provider choice using data from a large urban county in the Northeast US. The study subjects were 9,544 adult Medicaid recipients who received outpatient treatment from the 20 Community Mental Health Center (CMHC) programs in 2001. In addition to a conventional set of variables representing client and provider characteristics, the regression model included several interaction terms to examine whether racial concordance level among patients influences the choice of an outpatient program. The results revealed that racial concordance among the clients seems to be a factor in choosing a program. In particular, Caucasian clients are much more likely to select a program with a higher percentage of Caucasian clients, even though they have to travel further. More generally, our results suggest that program choice may be driven more by the racial composition of the clients served than by spatial proximity to the program.