JONES RESPONDS

Of the many daunting challenges faced by the Indian Health Service (IHS) in 1955, one of the most immediate was the lack of systematic health data. Over the last six decades, the IHS has worked consistently to overcome this deficit, from its first field health surveys in 1956 to its extensive data collection today.1^,2 Population-wide data remain invaluable to the IHS, revealing the persistence of wide disparities between American Indians and Alaska Natives (AIAN) and the general population even as overall health conditions improved among all groups in the United States. However, as Lanier reminds us, conglomerated data obscure significant disparities within AIAN populations. Such disparities have long been known. When Smithsonian anthropologist Ales Hrdlicka surveyed health conditions around 1900, he found enormous disparities in tuberculosis, with mortality rates ranging from 60 per 100 000 persons among the Navajo and Pueblo, to 3080 per 100 000 persons among the Sioux.3 Disparities between specific AIAN groups persist today for diseases and conditions as varied as HIV, diabetes, infant mortality, and many others.4^–6

Cancer data are especially revealing. When the IHS began collecting health data, it found lower rates of cancer (and heart disease) in American Indians and Alaska Natives than in the general population; infectious diseases remained prevalent among the AIAN population.1 This pattern fit well with popular notions that American Indians and Alaska Natives remained an underdeveloped society and did not share the burden of chronic disease that afflicted Whites in the United States. This low reported rate of cancer among AIAN groups likely reflected differences in life expectancy, risk factors, expectations, and disease surveillance. The cancer gap, however, has now closed. For many AIAN populations, not just those in Alaska, there already exists excess cancer mortality compared with the general population.7 The nature of the disparity varies widely for different AIAN groups (defined geographically) and for different cancers. Although population-wide data can reveal broad trends, health policies must be based on nuanced analyses of precisely defined populations.

Variations in rates of cancer and other diseases within AIAN groups should come as no surprise. These groups exhibit enormous diversity in socioeconomic status, educational attainment, geography (particularly urban vs rural), and many other factors that affect disease risk. The existence of disparities in the prevalence of disease within AIAN populations raises a more fundamental question: is race the best variable for parsing patterns of disease? Health data in the United States are steeped in race, with the disparity literature dominated by racial and ethnic comparisons. The use of crude racial and ethnic categories obscures important disparities within each group. It also obscures more important determinants of health outcomes, especially socioeconomic status and the growing income inequalities in the United States.