The Great Recession and Racial and Ethnic Disparities in Health Services Use

The “Great Recession” of 2007 to 2009 affected Americans of all backgrounds, across education, age, race/ethnicity, and household type, but took a far greater toll on African Americans and Hispanics than on whites.^1,2 In 2009, unemployment rates of African Americans (14.8%) and Hispanics (12.1%) were significantly higher than the rate for whites (8.7%).³ Median wealth fell 66% among Hispanic households, 53% among African American households, and 16% among white households.² Rates of employment-based health insurance declined more steeply for minorities than for whites, as 25% of African Americans and Hispanics lost their job during the recession compared with 15% of whites, and minorities were more likely to become uninsured.⁴

Loss of insurance coverage deteriorates access to care and is associated with reduced use of health services,⁵ particularly during recessions.⁶ Preventive service use is also sensitive to recessions.^6,7 Taken together, these forces suggest that health services use patterns of minorities may have been significantly altered during the recession. The objective of the present study was to examine differences in health services use regarding office-based physician visits, inpatient stays, emergency department visits, and prescription drug fills for racial and ethnic minorities before and during the Great Recession.

To investigate the association between the economic recession and health care use, we used data nationally representative of the civilian noninstitutionalized US population from the Medical Expenditure Panel Survey⁸ (MEPS) for 2005 to 2006 and 2008 to 2009 for adults aged 18 to 64 years. The total sample (N = 54 007) included non-Hispanic whites (whites) (n = 30 760), non-Hispanic African Americans (n = 9822) (African Americans), and Hispanics (n = 13 425). Our outcome variables included use counts of office-based physician visits, inpatient stays, emergency department visits, and prescription drug fills over the calendar year. We estimated negative binomial models for the count data. We report the incident rate ratios (IRRs), which indicate the estimated rate ratio of an explanatory variable relative to its reference category, if all the other variables are held constant in the model.⁹

The key explanatory variables included a dichotomous indicator equal to 1 for 2008 to 2009 and equal to 0 for 2005 to 2006. We also controlled for the interaction terms of this recession indicator with race and ethnicity. We controlled for respondents' demographics, socioeconomic status, health insurance coverage, self-reported health, and chronic conditions. Characteristics of the sample are available in the eTable.

There were significant differences in health services use by race/ethnicity before and during the recession. Unadjusted statistics show that the mean number of office-based physician visits over the 2-year prerecession period were lower for African Americans (5.75) and Hispanics (4.51) than for whites (7.34) (Table). During the recession, whites and Hispanics reported fewer physician visits compared with the period before the recession. Prescription drug use was highest for whites in the prerecession period (14.08), with a statistically significant drop during the recession to a mean of 13.44 prescription drug fills. Hispanics had the lowest mean prescription drug fill counts before the recession (8.40) and no significant drop during the recession (8.09). The drop in prescription drug fills for African Americans (from 12.93 to 12.74) was not significant. For African Americans, the mean number of inpatient stays was higher before the recession (0.16) and lower during the recession (0.14), but inpatient stays were not significantly reduced for whites or Hispanics. Emergency department use did not differ significantly across the 2 periods. African Americans reported the most visits, whereas use was similar for whites and Hispanics.

The IRRs from the negative binomial regressions show that there were fewer prescription drug fills (IRR, 0.91; P < .01) and inpatient stays (IRR, 0.90; P < .05) across all races and ethnicities during the recession relative to the prerecession period. There were no statistically significant overall reductions in physician visits or emergency department use related to the recession. The coefficients for the interaction terms of the year indicator (2008-2009) and “Hispanic” or “African American” show the differential associations of the recession compared with the prerecession period for minorities relative to whites. The Hispanic × year indicator shows that Hispanics had a greater (IRR, 0.91; P < .05) reduction of physician visits than did whites during the recession.

Our results show the recession of 2007 to 2009 is associated with downward trends in health care utilization relative to the period before the recession across race/ethnicity and services. Physician visits, prescription drug fills, and inpatient visits were statistically significantly lower in the recession period for whites, African Americans, and Hispanics. Negative binomial results show that the recession is associated with lower counts of prescription drug fills and inpatient stays across race/ethnicity.

Although minorities bore the brunt of the recession in terms of losses in employment, income, and insurance, our findings suggest that trends in use patterns were similar across race and ethnicity. The only evidence of ethnic disparities is the statistically significant finding that Hispanics reduced office-based physician visits more than whites during the recession.

Correspondence: Dr Mortensen, Department of Health Services Administration, University of Maryland, 3310 School of Public Health Bldg, College Park, MD 20742 (Karoline@umd.edu).

Published Online: January 7, 2013. doi:10.1001/jamainternmed.2013.1414

Author Contributions:Study concept and design: Mortensen and Chen. Acquisition of data: Chen. Analysis and interpretation of data: Mortensen and Chen. Drafting of the manuscript: Mortensen. Critical revision of the manuscript for important intellectual content: Mortensen and Chen. Statistical analysis: Chen. Administrative, technical, and material support: Mortensen. Study supervision: Mortensen.