Using the Integrated Health Interview Series (IHIS) Harmonization of National Health Interview Survey (NHIS) Data to Assess Racial Health Disparities

The National Health Interview Survey (NHIS) is a cross-sectional, multistage area probability household interview survey that provides national estimates on health indicators, health care access, availability and utilization, and health-related behaviors for the U.S. resident civilian, non-institutionalized population. Using 1999-2005 NHIS made available in harmonized form by the Integrated Health Interview Series (IHIS), the presentation examines racial disparities among sample adults in health status, selected health conditions, and health insurance coverage using alternative classifications of racial groups. Health conditions examined in the presentation are whether a doctor or other health care professional ever told the person that he or she had (1) asthma, (2) cancer, or (3) a heart attack. The health insurance status examined in the presentation is whether the person was uninsured. The health status variable uses a five-point Likert scale ranging from "excellent" to "poor." The presentation compares outcomes using pre-1997 Office of Management and Budget (OMB) racial standards with those same outcomes using 1997+ OMB racial standards, and calculates measures of disparity. Several logistic regressions, two using health status and two using insurance status as the outcome variable, and the two different OMB standards as independent variables, demonstrate the effect of using different race standards on health services research. Measures of disparities using 1999–2005 NHIS data differ depending on whether one uses pre-1997 Statistical Policy Directive No. 15 standards, or standards based on the October 1997 revisions of OMB Statistical Policy Directive No. 15. Because of this difference, whenever possible statistical analyses incorporating NHIS race variables and data from multiple years should include both sets of standards, and present comparative analyses. Policies at local, state and federal levels to address health disparities depend on a clear understanding of how and whether health status, health conditions and insurance status (whether or not insured) vary when conditioned on race. The change in OMB guidelines regarding racial categories, as instituted in the post-1998 NHIS, provides a natural social experiment to determine to what degree such disparities are a function of social (political) construction. The presentation suggests that disparities in health status, health conditions and insurance status based on self-reported main racial background are not simply a statistical phenomenon; political decisions have a direct affect on the presence and magnitude of disparities. Thus, any debate on the nature of such disparities and the appropriate policy responses must understand and take account of the social (political) construction of racial categories.