Comparison of Methods to Estimate Asthma Burden Among the Tribal Population in Wisconsin

BACKGROUND: In Wisconsin and elsewhere, many American Indian and Alaska Native (AI/AN) populations live in sparsely-populated areas such that rates of health outcomes of interest are difficult to quantify for small geographical areas. To overcome this problem, Contract Health Services Delivery Area (CHSDA) counties (geographic areas that include reservations and areas surrounding those reservations) are often used as a geographical proxy so that county-level data can be used to generate rates that may be applied to Tribal populations. However, there is some question about whether this approach generates reasonably accurate rates for Tribal populations. To inform this question, we undertook an analysis of Wisconsin hospitalization and emergency department (ED) data to compare estimates of AI/AN asthma burden based on CHSDA to estimates based on race. This analysis compares 1) CHSDA and non-CHSDA county rates vs. overall state rates by race; 2) within Tribe-specific CHSDAs, CHSDA rates vs. rates by race. 

METHODS: Wisconsin asthma inpatient and ED data (primary diagnosis ICD-9-CM code 493) were utilized to calculate age-adjusted rates for overall, child and adult populations. Wisconsin’s 31 CHSDA counties and their contribution to eleven Wisconsin Tribes were provided by the Great Lakes Inter-Tribal Epidemiology Center. Aggregate rates for all CHSDA counties vs. non-CHSDA counties excluding Milwaukee-metro counties were compared to state-wide rates by race during 2011-2013. CHSDA rates were compared to those by race within Tribal CHSDAs during 2007-2013.  

RESULTS: Wisconsin age-adjusted asthma ED rates overall and among adults for non-CHSDA counties (30.9/10,000 and 28.3/10,000) were higher than CHSDA rates (28.0/10,000 and 24.3/10,000). By race, Blacks had the highest rates overall (135.5/10,000) followed by significantly higher rates for AI/ANs (45.9/10,000) compared to whites (19.1/10,000). For the majority of Tribes, the CHSDA rates across all ages were 50-70% lower than the AI/AN race-based rates. Hospitalization data showed similar trends. Examination of race-based rates showed that adult AI/ANs had higher rates than children in some Tribes, which was opposite that seen among whites; this finding was not captured through CHSDA rates.  

CONCLUSIONS: CHSDA rates do not approximate AI/AN asthma rates in Wisconsin. It appears that CHSDA rates reflect the majority (white) population, rather than AI/AN population.  Furthermore, using CHSDAs as a geographical basis for calculating rates introduces misclassification, since some counties fall into two different Tribal CHSDAs. When race is not available, a methodology based on zip codes associated with tribes may produce more specific rates to highlight Tribal disease burden and disparities.