Indian Health Service Care System and Cancer Stage in American Indians and Alaska Natives
Purpose. We aimed to determine whether the association between late-stage cancer and American Indian/Alaska Native (AI/AN) race differed by enrollment in the Indian Health Service Care System (IHSCS). Methods. We used Surveillance, Epidemiology, and End Results (SEER) data linked to Medicare files to compare the odds of late-stage breast, colorectal, lung, or prostate cancer between non-Hispanic Whites (NHWs) (n=285,993) and AI/ANs with (n=581) and without (n=543) IHSCS enrollment. Results. For AI/ANs without IHSCS enrollment, the odds of late-stage disease were higher in AI/ANs compared with NHWs for breast (OR=3.17, 95%CI: 1.82–5.53) and for prostate (OR=2.59, 95%CI:1.55–4.32) cancer, but not for colorectal or lung cancers. Among AI/ANs with IHSCS enrollment, there was not a significant association between late-stage disease and AI/AN race for any of the four cancers evaluated. Conclusion. Our results suggest that enrollment in the IHSCS reduced the disparity between AI/ANs and NHWs with respect to late-stage cancer diagnoses.
Cancer stage, American Indian, Alaska Native, SEER-Medicare, Indian Health Service
The most common cancers in American Indians/Alaska Natives (AI/ANs) are breast, colorectal, lung, and prostate cancers.1,2 Overall, AI/ANs have lower cancer [End Page 245] incidence, but worse five-year survival outcomes, than non-Hispanic Whites (NHWs) for these cancers.1–3 AI/ANs are also more likely to present with late-stage, metastatic cancers, and this is particularly notable for cancers for which screening is available, including breast, colorectal, and prostate cancers.4–6 Thus, differences in stage at diagnosis may be partly driving the disparity in health outcomes for AI/AN compared with NHW cancer patients.
Regular access to primary health care services, including screening in asymptomatic patients and early diagnosis in symptomatic patients, is associated with early stage at diagnosis for cancer patients.7,8 However, AI/ANs face several barriers to accessing primary care, including limited insurance coverage, lack of transportation, few nearby health care facilities, high costs of health care services, and mistrust of health care providers.9–11 The Indian Health Service Care System (IHSCS), including Indian Health Service, Tribal facilities, and urban facilities, provides a potential avenue to reduce barriers to primary care in AI/ANs, because IHSCS clinics offer free primary care services, including cancer screening, to eligible AI/ANs. Additionally, IHSCS clinics tend to be located close to, or within, tribal areas, making them accessible to AI/ANs who may have limited transportation options. However, IHSCS services are restricted to AI/ANs who are members of federally-recognized tribes or who meet other eligibility criteria. Federal recognition status is granted to tribes based on historical treaties, Acts of Congress, United States Court decisions, or by administrative procedures under the Federally Recognized Indian Tribe List Act. Most AI/ANs who are not members of federally-recognized tribes are not eligible for care through the IHSCS.
The objective of this study was to evaluate whether AI/AN race is associated with a higher likelihood of having a late-stage cancer diagnosis and to determine if this association differs between AI/ANs with and without IHSCS enrollment. The IHSCS may help reduce some barriers to primary health care and result in earlier diagnosis of symptomatic AI/ANs and increased cancer screening in asymptomatic AI/ANs. Thus, we hypothesize that AI/AN cancer patients who are eligible for IHSCS services will have a stage distribution that is more similar to NHWs than AI/AN cancer patients without IHSCS care.
We conducted a case-case comparison study using Medicare enrollment files and IHSCS enrollment records linked to the Surveillance, Epidemiology, and End Results cancer registry data (SEER-Medicare). We used the SEER-Medicare linked database to identify new cases of breast (women only), colorectal, lung, or prostate cancer who were: 1) diagnosed between 2001 and 2007, 2) residing in a SEER catchment area, 3) enrolled in Medicare parts A and B one year or longer prior to diagnosis, and 4) AI/AN or NHW. SEER race/ethnicity data were abstracted from medical records data which were primarily based on self-reported race/ethnicity. This may result in misclassifying a subset of cancer patients who are truly AI/AN as NHW.12 To address this, we ascertained AI/AN racial status through both SEER and IHSCS data and used probabilistic linkage data to correct misclassification of AI/AN race.13 Those who were enrolled in Medicare due to disability or end-stage renal disease were excluded. [End Page 246]
Approval for this study was granted by the institutional review boards of the Fred Hutchinson Cancer Research Center, State of Washington, Oregon Department of Public Health, Oregon Department of Health Services, California Rural Indian Health Board, Northwest Portland Area Indian Health Board, California Committee for the Protection of Human Subjects, and the California Department of Health Care Services.
Stage at diagnosis was dichotomized into early-stage (SEER local/regional stages) or late-stage (SEER distant stage) disease. We used logistic regression models to compare the odds of a late-stage diagnosis between AI/AN and NHW cancer patients. Models included the following adjustment variables: age, sex, diagnosis year, marital status, geographic region, ZIP code-level median income, type of residence according to ZIP code rural-urban commuting area codes, and Klabunde co-morbidity score at diagnosis.14 We also conducted separate analyses to determine the association between late-stage disease and AI/AN race for those who were, and were not, eligible for IHSCS. Ineligibility for the IHSCS was determined using the SEER-Medicare data that had been linked to IHSCS files.
After excluding patients missing stage at diagnosis (approximately 5% of cancer patients), analyses included 285,993 NHWs and 1,124 AI/ANs diagnosed with breast, colorectal, lung, or prostate cancer during the study period. Compared with NHWs, AI/AN cancer patients tended to be younger, were less likely to be married, and were more likely to live in rural areas and in areas with lower ZIP code-level median incomes (Table 1). In analyses not accounting for IHSCS eligibility, the adjusted odds of presenting with late-stage disease was higher in AI/ANs compared with NHWs diagnosed with breast (OR=2.20, CI: 1.40–3.44) or prostate (OR=2.01, CI: 1.39–2.90) cancer (Table 2). The association between late-stage cancer and AI/AN race was strongest for AI/ANs without IHSCS eligibility for both breast and prostate cancers. AI/AN breast cancer cases who were not eligible for the IHSCS were greater than three times more likely (CI: 1.82–5.53) to be diagnosed with late-stage cancers than NHWs. For prostate cancer, AI/AN patients who were not eligible for the IHSCS were 2.5 times more likely (CI: 1.55–4.32) to be diagnosed with late-stage disease than NHWs, but there was not a significant difference in the distribution of late-stage breast or prostate between NHW and IHSCS eligible AI/AN patients. For colorectal and lung cancers, there was a 17–21% increased odds of late-stage disease in AI/ANs, but this was not statistically significant, and these associations did not change according to IHSCS eligibility status (Table 2).
Our results suggest that older AI/AN cancer patients enrolled in Medicare were more likely than NHW cancer patients to be diagnosed with late-stage cancer, particularly for breast and prostate cancers. These results highlight the reality that barriers to care in AI/ANs extend beyond differences in health insurance coverage.9–11 However, our analyses also suggested that enrollment in the IHSCS reduced the disparity in the [End Page 247]
[End Page 248]
[End Page 249] proportion of patients with late-stage cancer diagnoses between AI/ANs and NHWs. This may be due, in part, to increased access to primary care and cancer screening services in AI/ANs who are eligible for IHSCS services. Although prior research suggests gaps in preventive care through IHSCS,15 results from a recent qualitative study of mammography in AI/ANs are consistent with our analyses and support the hypothesis that the IHSCS may help bridge the gap in access to cancer early detection resources for AI/ANs.16
Our study included robust data from a large population of American cancer patients enrolled in Medicare, but we acknowledge several study limitations, including: 1) limited sample size for AI/AN cancer patients which resulted in reduced power to detect potential modest differences between AI/ANs and NHWs for late-stage disease in lung and colorectal cancers; 2) lack of qualitative data to assess cultural barriers to primary care and cancer screening among AI/ANs in our study population; and 3) likely mis-classification of a subset of AI/AN cancer patients as NHW.12 As noted in the Methods section, we used both SEER and IHSCS data to reduce misclassification of AI/ANs,13 but this will not eliminate all misclassification. Because AI/ANs make up a small percent of the U.S. population over the age of 65 years old, the percent of cancer patients that were classified as NHW, but who are actually AI/AN, will be small and will not have a large impact on study results. This type of misclassification may slightly bias our results towards the null, so the overall significant, positive association between AI/AN race and late-stage cancer that we reported is likely conservative. This conservative estimate in our overall analysis does not affect the interpretation of our stratified analyses, which suggest a reduction in the disparity between NHWs and AI/ANs for late-stage cancer diagnosis in breast and prostate cancer patients who are eligible for the IHSCS. Thus, the practical implications of this research point to developing and testing interventions aimed at increasing access to primary care and cancer screening services in AI/ANs. Additionally, future research to address potential cultural and social barriers to cancer screening in AI/ANs is needed to improve the early detection of cancer in this population and improve survival outcomes.
ANDREA N. BURNETT-HARTMAN and STACEY A. COHEN are associated with the Fred Hutchinson Cancer Research Center, Seattle, WA. Andrea N. Burnett-Hartman is also associated with Kaiser Permanente Colorado Institute for Health Research, Denver, CO. SCOTT V. ADAMS and AASTHAA BANSAL are associated with the University of Washington, Seattle, WA, along with STACEY A. COHEN. SCOTT D. RAMSEY is associated with both the Fred Hutchinson Cancer Research Center and the University of Washington. JEAN A. MCDOUGALL is associated with the University of New Mexico, Albuquerque, NM. ANDREW KARNOPP is associated with Kaiser Permanente Center for Health Research, Portland, OR. VICTORIA WARREN-MEARS is associated with Northwest Portland Area Indian Health Board, Portland, OR.