mental health across racial and ethnic groups. Finally, we have mainly compared disease prevalence across groups, but groups may also differ in such other factors as age at disease onset, duration, and severity.
The uncertainties and inconsistencies in this sketch of mortality and health differences suggest various directions for future research: to clarify the health status at older ages of particular groups, to define more clearly what appear to be contradictions in particular areas of health and disease, and to resolve apparent conflicts between indicators.
Research Need 1: Attempt a systematic decomposition of racial and ethnic health differences in mortality and morbidity among older people to determine the relative contribution of particular diseases and conditions. Try to assign differences in the prevalence of specific diseases and conditions to differences in the prevalence of major risk factors.
Racial and ethnic differences in mortality have in fact been decomposed in relation to various diseases and conditions with reference to people of all ages (Wong et al., 2002). A focus on older people would be useful. Moving beyond such decomposition to assess the contribution of risk factors of all types to differences would be a complex exercise, though perhaps a more manageable challenge than the World Health Organization’s (2002) attempt to assess the contribution of all major risk factors to the global disease burden. If such partitioning can be accomplished, it should be possible to refine the research agenda to focus on the most critical areas.
The health disadvantages of blacks and American Indians and Alaska Natives appear to be the most troubling. Black disadvantages are the best documented, so the research emphasis for blacks might be more appropriately on understanding differences, which we consider in subsequent chapters.
Research Need 2: Clarify the contrasts between mortality rankings and morbidity rankings, particularly between older whites and Hispanics, and assess the relative contributions of diseases and conditions to differences in mortality and overall health.
One problematic contrast is the consistently poorer self-reported health among Hispanics than would appear warranted by mortality levels. Another is the higher incidence of heart disease mortality among blacks than whites, despite the higher reported diagnoses of the disease among whites. To understand such apparent contradictions, it would be useful to be able to assess, on a continuing basis, how much each disease or condition adds to mortality or detracts from overall health, as has been done for the population as a whole (Wong et al., 2002) but not for older adults or different ethnic