*22 III. Results

Consistent with national estimates, there are considerable racial/ethnic differences in the socio-demographic characteristics of African Americans, Latinos, and whites (table 1). Most notably, African Americans and Latinos are poorer than whites and a larger proportion are uninsured. Latinos have the highest uninsured rate among all racial and ethnic groups. Whites are more likely to be privately insured than African Americans or Latinos. African Americans are more likely to have Medicaid than Latinos or whites. Racial/ethnic differences in health coverage as well as differences in other factors are important to consider when comparing usual sources of care. For African Americans and Latinos, the sources of medical care of the publicly insured and uninsured are as important to examine as the sources of care of the privately insured. Moreover *23 these differences in health coverage highlight the importance of making racial comparisons among similarly insured population groups.

 A. No Usual Source of Care

On average, the vast majority of white, African-American, and Latino children and adults have a usual source of medical care. However, a sizable share of the population, 18.5%, lack a usual source of medical care. Applying this percentage to the non-elderly population in 1996 yields approximately 40.4 million people without a usual source of medical care. The uninsured, regardless of race/ethnicity, are more likely to lack a usual source of care than persons with private coverage or Medicaid. However, striking racial/ethnic differences even exist among persons with similar insurance coverage. This analysis focuses on the site of care for individuals who identify a usual source of medical care.

 B. Use of a Particular Site of Care

Among persons with a usual source of care, office-based providers (generally physicians) are clearly an important source of care, regardless of race/ethnicity and health coverage. Analysis of the MEPS data shows that office-based providers are the usual source of care for most (88.2%) persons under age sixty-five (figure 3). Only 11.2% of respondents under age sixty-five identify a hospital-based clinic or OPD as a usual source of care, and less than 1% (0.6%) identify an ER as a usual source of care.

*24 Despite the perception that ERs are widely misused by minority Americans, only a small fraction of respondents identify the ER as their usual source of care. Racial differences are observed when examining the usual sources of medical care for children and adults.

Health insurance does not appear to be a major factor affecting the use of an office-based provider as a usual source of care by white children, but it does appear to play a role for African-American and Latino children (table 2). About 90% of white children who are either covered by Medicaid or private insurance, or who are uninsured, have an office-based provider as a usual source of care. African-American and Latino children with private coverage report the use of an office-based provider at rates similar to those of white children. However, minority children who are covered by Medicaid or who are uninsured are far more reliant on a hospital clinic or OPD as a usual source of care than are white children. Among Medicaid beneficiaries, more than twice as many African-American children (22.8%) as white children (9.9%) use a hospital-based provider as their usual source of care. Similarly, almost twice as many Latino children (18.8%) as white children rely on a hospital-based provider. As for the uninsured, African-American (24.1%) and Latino (17.2%) children are at least twice as likely as their white counterparts (8.3%) to use a hospital-based clinic or OPD as their usual source of care.

 Findings on the usual sources of medical care of adults parallel those of children (table 3). Racial/ethnic differences are largest among adults enrolled in Medicaid and among the uninsured. A hospital-based clinic or OPD is the usual source of care for approximately twice as many African-American (29%) and Latino (25%) adults with Medicaid as their white *25 counterparts (12.5%). Similar findings are observed for the uninsured, with uninsured African Americans and Latinos being nearly twice as likely to obtain care from a hospital clinic or OPD as whites. Also, the percentages who identify an ER as a usual source of care are small and are not statistically different by race/ethnicity.

 C. Multivariate Analysis

When holding measures of socio-demographic and health status constant, race/ethnicity persists as a factor significantly and strongly associated with the use of a hospital-based provider as a usual source of medical care. Findings are strikingly similar for children and adults (tables 4 and 5), and excluding those who identify the ER as a usual source of care does not appreciably change these results.

African Americans and Latinos, regardless of insurance coverage, are more likely than whites to have a hospital-based provider as a usual source of medical care. For example, African-American children are 2.5 times as likely as their white counterparts, and Latino children are twice as likely as their white counterparts, to have a usual source of medical care that is not an office-based provider. The racial differential persists for adults as well, although the effect is modestly diminished. The lack of a statistically significant interaction term for race and insurance coverage provides evidence that these findings consistently apply across racial/ethnic groups and insurance categories. Thus, privately insured African Americans and Latinos are more likely than their privately insured white counterparts to use hospital-based providers as a usual source of medical care. Similarly, African Americans and Latinos with Medicaid are more likely than their *26 white counterparts with Medicaid to use hospital-based providers as a usual source of medical care.

 Insurance status is also associated with where medical care is obtained. Uninsured children are roughly 1.5 times as likely as privately insured children to use a hospital-based provider as a usual source of care. Similarly, uninsured adults are at least 1.4 times as likely as privately insured adults to use a hospital-based provider as a usual source of care. There is some indication that Medicaid beneficiaries and the privately insured, other factors being equal, do not differ substantially in their major sources of health care. In other words, when comparing individuals, for example, of similar race/ethnicity or health status, those with Medicaid *27 and private coverage do not statistically differ in the likelihood of having a hospital-based provider as their usual source of care.

Finally, the regression results show that two factors, in addition to race and insurance, are associated with where adults obtain health care (table 5). Family income and geographic region also are related to the usual source of care for adults. Adults with family incomes at or below the federal poverty level are more likely to use a hospital-based clinic as a usual source of care than non-poor adults. Also, adults living in the West and Midwest are more likely than adults in the Northeast to use a hospital-based provider as a usual source of care.