Cancer Incidence and Staging among American Indians in Oklahoma

Abstract

Background

This study describes overall and site specific cancer incidence among AI/ANs compared to whites in Oklahoma and differences in cancer staging.

Methods

Age-adjusted incidence rates obtained from the Oklahoma Central Cancer Registry are presented for all cancer sites combined and for the most common cancer sites among AI/ANs with comparisons to whites. Percentages of late stage cancers for breast, colorectal, and melanoma cancers are also presented.

Results

AI/ANs had a significantly higher overall cancer incidence rate compared to whites (629.8/100,000 vs. 503.3/100,000), with a rate ratio of 1.25 (95% CI: 1.22, 1.28). There was a significant disparity in the percentage of late stage melanoma cancers between 2005 and 2009, with 14.0% late stage melanoma for whites and 20.0% for AI/ANs (p-value: 0.03).

Conclusions

Overall, there were cancer disparities between AI/ANs and whites in Oklahoma. Incidence rates were higher among AI/ANs for all cancers and many site specific cancers.

BACKGROUND

Cancer is a serious cause of morbidity and mortality in the United States. Disparities exist in incidence, mortality, survival and stage of diagnosis between American Indians and Alaska Natives (AI/ANs) and whites.^1–23 When accounting for racial misclassification, AI/ANs have some of the highest mortality rates in the US.^24,25 In the US, regional differences exist with significantly disparate cancer incidence rates among AI/ANs compared to whites in the Alaska, the Northern Plains, and the Southern Plains regions.¹⁵

In 2011, approximately 3.8 million people resided in Oklahoma. Of these, 8.9% reported that they were AI/AN only. Within the AI/AN population, estimated at 337,000, there is a wide diversity of cultures represented by 39 tribal nations. Oklahoma has the second largest number and the second highest proportion of AI/AN individuals of all US states. The Oklahoma City Area Indian Health Service (IHS) had 332,658 health care and environmental services users in fiscal year 2012. Oklahoma is unique compared to other areas in the US because AI/ANs live with a higher level of integration into mainstream society, rather than on tribal reservations.

The purpose of this study is to describe cancer incidence among AI/ANs in Oklahoma. We describe statewide age-adjusted cancer incidence rates for AI/ANs and whites in Oklahoma, stage of diagnosis for specific cancers and cancer incidence rates for Oklahoma City Area IHS service units within Oklahoma.

METHODS AND MATERIALS

Central cancer registries receive case information from multiple sources, including hospitals, outpatient surgery centers, freestanding radiation centers, laboratories, and death certificates. Two federal programs fund the two central cancer registries in Oklahoma. The Centers for Disease Control and Prevention, National Program of Cancer Registries (CDC-NPCR) funds the Oklahoma Central Cancer Registry (OCCR) and the National Cancer Institute’s (NCI) Surveillance, Epidemiology, and End Results (SEER) Program funds the Cherokee Nation Cancer Registry (CNCR). OCCR is the central cancer registry responsible for all cancers diagnosed or treated in the state of Oklahoma and CNCR is responsible for those diagnosed or treated among AI/ANs in the Cherokee Nation Tribal jurisdictional area. OCCR and CNCR perform an annual data exchange and, thus for this study, data from OCCR will be used as it includes CNCR cases. OCCR has a greater than 95% estimated case completeness for each year of data collected between 1997 and 2009. The OCCR has received the highest level of certification available from the North American Association of Central Cancer Registries (NAACCR) based on quality, completeness and timeliness of data collected for 2001–2009.²⁶

To identify AI/AN cancer cases that were misclassified as other races, OCCR works with IHS to link cancer registry records with IHS patient registration files.⁵ AI/AN individuals must provide proof of membership in a federally recognized tribe to receive healthcare from the Indian health system which includes IHS, Tribal, and Urban facilities. An individual’s race was determined using the OCCR race 1 variable in combination with the IHS link variable or race 1 alone if it was coded as AI/AN. Thus, this study considers AI/ANs to be those coded as AI/ANs on the race 1 variable or the IHS Link variable and excludes that person from any other race codes. For this study only AI/ANs and whites are considered.

Age-adjusted incidence rates are presented for all cancer sites combined and for the most common cancer sites among AI/AN bridged population estimates in Oklahoma; site categories are consistent with prevailing reporting standards.^27,28 Lymphomas (ICD-O histology codes 9590-9729) are presented as two separate categories (i.e., Hodgkin and non-Hodgkin lymphoma) and are not included with other tumors of specific anatomic sites. Mesothelioma (ICD-O histology codes 9050-9055) and Kaposi sarcoma (ICD-O histology code 9140) are not included with other tumors of specific anatomic sites. In situ and invasive bladder tumors are combined in a single category.²⁹ For gender-specific cancers including prostate, breast, corpus uteri, ovary, cervix uteri, testis, and penis cancers, the incidence rates were calculated using gender-specific populations. All other benign tumors (ICD-O behavior codes 0) and tumors of uncertain or unknown behavior (ICD-O behavior code 1) were excluded for the analysis.

Cancer stage at diagnosis for breast and colorectal cancer was determined using the derived SEER Summary Stage 2000. To review staging changes through time, we considered two periods, the earliest years available and the latest years available. For breast and colorectal cancer we used three years of data (1997–1999 and 2007–2009); for melanoma and cervical cancers we used five years of data (1997–2001 and 2005–2009) because of the small number of cases among AI/ANs for these sites. In situ was included for breast, colorectal cancer and melanoma but not included for cervical cancer. Early stage for cervical cancers was local stage and early stage for breast, colorectal and melanoma was in situ and local stages combined. Late stage included regional and distant stages for all sites. Cases that were identified as Death Certificate Only or had unknown stages were excluded from the analysis, but the percentages for unknown stage are shown.

Incidence rates were calculated per 100,000 population using bridged population estimates and were age-adjusted by the direct method³⁰ using the 2000 United States standard population. Incidence rates, rate ratios and 95% confidence limits were calculated. Denominators for rate calculations were derived from population estimates obtained from SEER population files. Incidence rates were calculated for the whole state and for each of the nine IHS service units in Oklahoma (Figure 1). These IHS service units were chosen because they are consistent with IHS reports and IHS service provision in Oklahoma. Incidence rates were calculated using the Oklahoma State Department of Health’s web-based query system, Oklahoma Statistics on Health Available for Everyone (OK2SHARE).³¹ For this analysis the two sample z-test for comparing a difference in proportions was calculated to allow us to test for a statistical significance for primary payer and for stage at diagnosis. To ensure confidentiality, numbers were not reported when total counts were less than ten.

Figure 1.

Indian Health Services Units in Oklahoma

RESULTS

Overall, we found an increase in the age-adjusted incidence rates in AI/ANs compared to whites in Oklahoma from 2005–2009. When accounting for misclassification, AI/ANs had a significantly higher overall cancer incidence rate than whites (629.8/100,000 vs 503.3/100,000) with a rate ratio of 1.25 (95% CI: 1.22,1.28). Females had an increased rate ratio between AI/ANs and whites at 1.31 (95% CI: 1.27,1.35) compared to males at 1.19 (95% CI: 1.15,1.23; Table 1).

Table 1.

Age-Adjusted Incidence Rates and Rate Ratios of Cancers for American Indian/Alaska Natives Compared With Whites by Gender: Oklahoma 2005–2009

	White		American Indians/ Alaska Native
	Count	Age-Adjusted Incidence Rate (100,000)	Count	Age-Adjusted Incidence Rate (100,000)	Rate Ratio	95% Confidence Interval
Male	43,136	570.7	3,703	677.5	1.19	(1.15–1.23)
Female	41,367	459.6	4,155	602.7	1.31	(1.27–1.35)

Age-group specific cancer rates were higher among AI/ANs compared to whites among each age group, with the exception of those ages 5 to 9. Rates were significantly higher among AI/ANs compared to whites in all but four of these age groups (Table 2). All age-specific groups 25 years and older showed significantly increased rates among AI/ANs when compared to whites.

Table 2.

Age-Specific Incidence Rates and Rate Ratios of Cancers for American Indian/Alaska Natives Compared With Whites: Oklahoma 2005–2009

	White		American Indian/Alaska Native
Age Group	Count	Age-Adjusted Incidence Rate (100,000)	Count	Age-Adjusted Incidence Rate (100,000)	Rate Ratio	95% Confidence Interval
0–4 years	204	26.2	36	29.2	1.11	(0.72–1.51)
5–9 years	135	14.6	17	12.2	0.84	(0.41–1.26)
10–14 years	125	13.8	27	18.7	1.36	(0.79–1.92)
15–19 years	200	20.7	58	37.6	1.82	(1.29–2.35)
20–24 years	393	36.5	66	43.7	1.20	(0.89–1.51)
25–29 years	566	56.4	102	82.3	1.46	(1.15–1.77)
30–34 years	840	95.2	133	128.2	1.35	(1.10–1.59)
35–39 years	1,295	143.2	215	217	1.52	(1.30–1.73)
40–44 years	2,344	243.3	341	339.7	1.40	(1.24–1.55)
45–49 years	4,360	405.9	536	516.5	1.27	(1.16–1.39)
50–54 years	6,562	638.6	729	787.8	1.23	(1.14–1.33)
55–59 years	8,540	928.3	911	1152.8	1.24	(1.16–1.33)
60–64 years	10,372	1343.4	1,045	1738.4	1.29	(1.21–1.38)
65–69 years	11,894	1948.1	1,090	2485.9	1.28	(1.20–1.36)
70–74 years	11,495	2292.6	961	2901.8	1.27	(1.18–1.35)
75–79 years	10,652	2539.2	730	3019.1	1.19	(1.10–1.28)
80–84 years	8,067	2536.0	500	2951.2	1.16	(1.06–1.27)
85+ years	6,462	2180.2	364	2494.2	1.14	(1.02–1.26)

Primary payer at diagnosis is an important variable for understanding access to care and to screening services. Primary payer at diagnosis was significantly different for AI/ANs and whites for all payer sources (Table 3). AI/ANs were more likely than whites to have unknown insurance (11.2% vs 8.3%), to have Medicaid (7.8% vs 4.1%), and to have IHS/Public Health Service coverage (12.7% vs 0.0%) as their primary payer. AI/ANs were less likely to be uninsured (2.3% vs 3.7%), to have military insurance (0.5% vs 1.1%), to have Veterans Affairs insurance (0.8% vs 1.1%) and have private insurance (24.0% vs 31.7%) as their primary payer.

Table 3.

Percentage of Cancer for American Indian/Alaska Natives Compared With Whites by Primary Payer at Diagnosis: Oklahoma 2005–2009

	White		American Indian/ Alaska Native
	Count	Percent	Count	Percent	p-value
Unknown	7,042	8.3%	877	11.2%	<0.001
Not insured	3,135	3.7%	183	2.3%	<0.001
Private insurance	26,791	31.7%	1,884	24.0%	<0.001
Medicaid	3,494	4.1%	613	7.8%	<0.001
Medicare	42,199	49.9%	3,199	40.7%	<0.001
Military	894	1.1%	41	0.5%	<0.001
Veterans affairs	921	1.1%	66	0.8%	0.039
Indian/ Public health service	30	0.0%	998	12.7%	<0.001

Primary Site

For 36 specific cancer sites, 21 had rate ratios that were significantly higher among AI/ANs compared to whites(Table 4). The highest site specific incidence among both AI/ANs and whites was found to be prostate cancer. The prostate cancer rate for AI/ANs was significantly higher among AI/ANs compared to whites, with a rate ratio of 1.10 (95% CI: 1.01, 1.16). Lung cancer was the most commonly diagnosed cancer for both AI/ANs and whites. AI/ANs had a significantly higher rate of lung cancer, with a rate ratio of 1.50 (95% CI: 1.37, 1.58). Primary sites demonstrating large disparities between the two populations included: kidney and renal pelvis (RR: 1.90, 95% CI: 1.70, 2.10), liver and bile duct (RR: 2.21, 95% CI: 1.84, 2.58), stomach (RR: 2.02, 95% CI: 1.63, 2.41), gallbladder (RR: 3.40, 95% CI: 2.20, 4.60), and penis (RR: 4.00, 95% CI: 1.88, 6.12). Four of the primary sites had rate ratios that reflected lower incidence among AI/ANs, but only two of those sites had rates that were significantly higher among whites: melanomas of the skin (RR: 0.60, 95% CI: 0.52, 0.68) and bladder (RR: 0.87, 95% CI: 0.75, 0.99; Table 4).

Table 4.

Age-Adjusted Incidence Rates and Rate Ratios of Cancers for American Indian/Alaska Natives Compared With Whites by Primary Sites: Oklahoma 2005–2009

	White		American Indian/Alaska Native
Site	Count	Age-Adjusted Incidence Rate (100,000)	Count	Age-Adjusted Incidence Rate (100,000)	Rate Ratio	95% Confidence Interval
Total	84,506	503.3	7,861	629.8	1.25	1.22–1.28
Prostate*	10,985	142.4	817	154.7	1.09	1.01–1.16
Breast (Female, Excluding In Situ)*	10,818	121.5	973	140.5	1.18	1.08–1.23
Lung and bronchus	13,511	78.9	1,282	107.2	1.36	1.28–1.44
Colon and Rectum	8,082	47.7	861	70.5	1.48	1.37–1.58
Corpus uteri*	1,840	20.5	220	31.1	1.52	1.30–1.73
Kidney and renal pelvis	2,718	16.2	392	30.8	1.90	1.70–2.10
Non-Hodgkin Lymphoma	3,287	19.6	282	22.8	1.16	1.02–1.30
Breast (Female, In Situ only)*	2,146	24.4	160	22.6	0.94	0.78–1.08
Bladder	3,508	20.5	211	17.9	0.87	0.75–0.99
Leukemia	2,275	13.9	227	17.8	1.28	1.11–1.46
Ovary*	1,113	12.3	121	17.7	1.44	1.17–1.71
Cervix uteri*	720	9.5	117	16.5	1.74	1.40–2.08
Melanomas of the skin	4,467	27.4	214	16.4	0.60	0.52–0.68
Pancreas	1,803	10.5	156	13.1	1.25	1.04–1.45
Liver and bile duct	981	5.8	160	12.8	2.21	1.84–2.58
Oral cavity and pharynx	2,162	12.8	162	12.6	0.98	0.83–1.14
Thyroid	1,403	9.2	142	10.0	1.09	0.90–1.27
Stomach	807	4.8	119	9.7	2.02	1.63–2.41
Brain	1,152	7.3	130	9.3	1.27	1.04–1.50
Other nervous system	1,172	7.2	114	9.0	1.25	1.01–1.49
Other female genital system*	449	5.1	57	7.9	1.55	1.12–1.98
Multiple myeloma	860	5.0	96	7.8	1.56	1.23–1.89
Testis*	371	5.3	49	6.0	1.11	0.79–1.47
Larynx	757	4.5	75	6.0	1.33	1.02–1.65
Esophagus	824	4.8	71	5.5	1.15	0.87–1.42
Soft tissue (including heart)	472	3.0	53	4.0	1.33	0.95–1.71
Other digestive system	412	2.4	46	3.9	1.63	1.13–2.12
Other endocrine glands	420	2.8	55	3.8	1.36	0.98–1.74
Penis*	69	0.9	17	3.6	4.00	1.88–6.12
Gallbladder	167	1.0	38	3.4	3.40	2.20–4.60
Hodgkin Lymphoma	373	2.5	42	2.7	1.08	0.74–1.42
Anus	339	2.1	28	2.1	1.00	0.61–1.39
Small intestine	334	2.0	25	2.1	1.05	0.62–1.48
Bones and joints	137	0.9	23	1.4	1.56	0.87–2.24
Eye	175	1.1	14	1.1	1.00	0.46–1.54
Mesothelioma	152	0.9	13	1.0	1.11	0.48–1.74

^*Incidence rates calculated using gender-specific populations

Stage at diagnosis

From 1997–1999 there was a significant disparity in the percent of female breast cancers diagnosed at the late stage between whites and AI/ANs (29.5% vs 41.5% respectively, p-value <0.001); for 2007–2009 there was no significant disparity between whites and AI/ANs (31.0% vs 34.2% respectively, p-value 0.097) in late stage diagnosis of female breast cancer. There was a 17.5% decrease in late stage diagnosis of female breast cancer among AI/ANs (41.5% vs 34.2%). A 5% increase was found among whites (29.5% vs 31.0%; Table 5). There was a decrease in late stage colorectal cancers diagnosed from 1997–1999 to 2007–2009 among both whites and AI/ANs (59.2% vs 54.9% and 63.9% vs 58.2% respectively) and no significant difference was found between the two populations (Table 5). There was a statistically significant disparity in stage at diagnosis of melanoma of the skin between whites and AI/ANs in the later time period of 2005–2009 (p-value 0.030) but no significant difference between the percent diagnosed at late stage in the earlier time period, 1997–1999 (p-value 0.113; Table 5). There was a large decrease in the percentage of late stage melanoma cancer when the two time frames were compared. There was a 32% decrease among whites and a 27% decrease among AI/ANs (Table 5). There was a 40% increase in percentage of late stage diagnoses for cervical cancer among both whites (44%) and AI/ANs (43%) between 1997–2001 and 2005–2009.

Table 5.

Percentage Late Stage of Female Breast, Colorectal, Melanoma, and Cervical Cancers for American Indian/Alaska Natives Compared With Whites by Year Groups: Oklahoma

	White			American Indian/Alaska Native
	Total	Percent Unknown	Percent Late Stage	Total	Percent Unknown	Percent Late Stage	p-value % late stage
Female Breast
1997–1999	7490	7.2%	29.5%	523	7.8%	41.5%	<0.001
2007–2009	7911	6.2%	31.0%	714	9.0%	34.2%	0.097
Colorectal
1997–1999	5244	10.9%	59.2%	377	14.1%	63.9%	0.093
2007–2009	4767	12.0%	54.9%	535	15.5%	58.2%	0.176
Melanoma
1997–2001	2094	18.1%	20.6%	101	20.4%	27.5%	0.113
2005–2009	4467	9.9%	14.0%	214	22.9%	20.0%	0.030
Cervical
1997–2001	770	12.5%	38.3%	108	12.0%	39.0%	0.900
2005–2009	720	10.0%	55.1%	117	12.8%	55.9%	0.882

Geographic Distribution

Figure 1 displays the IHS Service Units in Oklahoma. There were disparities in age-adjusted incidence rates for each of the IHS service unit geographic designations (Table 6). Among the nine service unit areas, only two (Clinton and Lawton) did not have a statistically significant rate ratio of AI/ANs rates compared to white rates (Table 6; Figure 1). All other areas had significant rate ratio differences between AI/ANs and whites within the region (Table 6). The Tahlequah service unit, at 1.47 (95% CI: 1.38, 1.56), had the highest age-adjusted incidence rate ratio with the Claremore service unit being next highest at 1.30 (95% CI: 1.25, 1.35).

Table 6.

Incidence Rates and Rate Ratios of Cancers for American Indian/Alaska Natives Compared With Whites by Oklahoma Indian Health Service Units: Oklahoma 2005–2009

	White		American Indian/Alaskan Native
	Count	Age-Adjusted Incidence Rate (100,000)	Count	Age-Adjusted Incidence Rate (100,000)	Rate Ratio	95% Confidence Interval
Ada	6,440	519.3	669	620.7	1.20	(1.10–1.29)
Claremore	24,768	509.3	2,653	663.2	1.30	(1.25–1.35)
Clinton	7,016	493.2	222	539.5	1.09	(0.95–1.24)
Lawton	6,728	495.0	431	506.6	1.02	(0.92–1.12)
Pawnee	6,409	494.5	485	610.8	1.24	(1.12–1.35)
Shawnee	23,251	510.4	1,120	604.4	1.18	(1.11–1.26)
Tahlequah	4,068	463.1	1,340	681.1	1.47	(1.38–1.56)
Talihina	4,729	479.0	717	584.1	1.22	(1.12–1.32)
Wewoka	1,008	527.8	193	645.9	1.22	(1.04–1.41)

DISCUSSION

Overall, we see disparities between AI/ANs and whites in Oklahoma similar to or greater than those seen in the US overall.^15,32 Incidence rates were higher and stage at diagnosis was often worse for AI/ANs compared to whites. These differences persisted in each IHS service unit. Moreover, this study shows that while we have made significant progress in understanding the AI/AN cancer burden in the US, we have yet to realize Judith Swan’s plea of almost ten years ago to provide local AI/AN data.³² This study is one step in that direction.

While the age-adjusted incidence rate was lower in females compared to males in both whites and AI/ANs, the rate was higher in AI/ANs compared to whites for both genders, with females having a higher rate ratio suggesting there is a larger racial disparity among females. This gender gap bears further investigation.

In those under 25 years, the only significant difference by age between whites and AI/ANs was in the 15–19 year age group. This age group is an interesting group as they are transitioning from childhood to adulthood and preventive/primary care may be lacking.³³ More investigation on access to care and risk factors in AI/ANs compared to whites is needed to better understand potential disparities in this group. When examining those 25 years and older, the higher rate ratios are in those 25 to 44 years, with the ratios appearing to decline as age increases.

There were differences by primary payer at diagnosis for all payer types. It is important to note that the primary payer at diagnosis for most AI/ANs in Oklahoma was Medicare (40.7%) followed by private insurance (24.0%) and 2.3% of AI/ANs reported no primary payer or were uninsured. The slightly lower percentage of cancer cases with a primary payer of Medicare is consistent with the lower percentage of the population age 65 and older among Oklahoma AI/ANs; 7.4% AI/ANs compared to 16.0% among whites were 65 and older in 2011.³⁴ The higher percentage of AI/ANs with a primary payer of Medicaid is also consistent with the higher poverty level among AI/ANs; 21.9% of AI/ANs lived below the poverty level compared to 14.2% of whites in 2011.³⁴ Additionally, AI/AN women participating in the Breast and Cervical Cancer Treatment program are eligible at a higher poverty level (250% of Federal Poverty Level compared to 185% for non-AI/ANs). Only 12.7% of all AI/ANs had IHS/Public Health Service as their primary payer.

When examining the most commonly occurring cancers listed by NCI, AI/ANs had a significantly higher rate ratio for breast, colon and rectal, cervix uteri, kidney and renal pelvis, leukemia, lung and bronchus, non-Hodgkin lymphoma, prostate, and pancreatic.³⁵ AI/ANs had a significantly lower rate ratio than whites for bladder and melanomas of the skin. There was no significant difference in thyroid cancers. AI/ANs had a more than 3-fold higher rate for both penis and gallbladder cancers compared to whites. However, these are rare cancers and small numbers were used to determine these rates, resulting in imprecise estimates. Future studies should examine the reasons for these differences, including preventive care and risk factors in the two populations.

When comparing the percentage of late stage cancers between whites and AI/ANs over two periods, the percentage of late stage breast cancer was significantly higher among AI/ANs from 1997–1999, while there was no significant difference between the two racial groups seen in the later time period of 2007–2009. The percentage of late-stage breast cancer decreased in AI/ANs but increased in whites. This may reflect enhanced screening through Tribal programs, including the Cherokee Nation Breast and Cervical Cancer Early Detection Program, Kaw Nation Breast and Cervical Cancer Early Detection Program, Cherokee Nation Comprehensive Cancer Control Program, as well as other strong Indian health screening programs. We also saw a difference in melanoma between the two populations from 2005–2009, with AI/ANs having a significantly higher percentage of late stage cancers than whites. However, there was a decrease in the percentage of late stage cancers from 1997–2001 to 2005–2009 in both AI/ANs and whites. While there was no significant disparity between AI/ANs and whites in late stage cervical cancer diagnosis in either of the two time periods, we were concerned about the large increase in late stage cervical cancer diagnosis among both AI/ANs and whites. Finally, it was very concerning that although AI/ANs have a lower rate ratio of melanoma incidence, they have a higher percentage of late stage cancers, particularly in the more recent time periods (2005–2009). Future studies should focus on both the disparity of melanoma staging as well as the increase in late stage cervical cancer.

We saw a significantly higher rate of cancer incidence among AI/ANs compared to whites in all but two of the Oklahoma IHS service units from 2005–2009. Future studies should further investigate more specific locations within these service units where there are increases and understand the demographics structure and risk behaviors of these populations, including tribal differences. Studies should also aim to understand why there are higher rate ratios in some service units compared to others, particularly the Tahlequah Service Unit. One possible explanation to these differences may be from environmental factors. Various sources of environmental contamination may play a role in the disparity of cancer incidence among service units. Future studies should be done to address the possible connection between these environmental factors and the differences we see in cancer incidence.

There are several limitations in this study. One possible limitation is the changes in reporting quality over time. During the reporting period, there have been changes in the quality of some variables collected by the OCCR since reporting requirements to the NPCR have changed. There have also been changes to categorization within certain variables, one of which is staging. We did, however, use SEER summary stage which is fairly consistent. This could result in misclassification primarily through stage migration.^36–38 However, it is expected to be non-differential between whites and AI/ANs as these changes were not specific to a particular racial group. Another limitation is the large percentage of unknown or missing stage. This could be differential and affect the results since the percent of missing stage varied by cancer site, time-period, and racial group. More investigation should be done to understand the potential impacts of these differences in the interpretation of the results. Regarding the inclusion of all AI/AN cases in Oklahoma, there could be missing cases since Indian health facilities are not required to report to the OCCR. However, we believe that this number is very small as many Indian health facilities choose to report cases and OCCR works with laboratories and with Indian health facilitates for follow-back on these laboratory reported cases. Additionally, Indian health facilities typically have limited capacity to provide complete cancer treatment, thus most of the cases are contracted to non-Indian health facilities for at least some of their care and are, thus, reported from that facility. There is also still some misclassification in the AI/AN population. Not all AI/ANs seek care through Indian health system during their life and may still have their race reported as white or another race as opposed to AI/AN. This study also did not consider mortality rates, thus an important piece of the AI/AN cancer burden was not considered. A final limitation of this study was the lack of tribal specific data. While reviewing data by IHS service units shows some differences by regions, these service units are not consistent with tribal jurisdictional areas. Additionally, because of the large number of tribal headquarters and tribal heterogeneity in the state, tribal members live throughout the state rather than on “reservations” or in tribal jurisdictional areas. The geographic dispersion of the AI/AN population leads to an additional challenge in reviewing any health data for AI/AN populations in Oklahoma. This challenge is the issue of “rurality”. Since much of the Oklahoma AI/AN population resides outside of the three major cities (those being Oklahoma City, Tulsa and the Greater Lawton Areas), it is uncertain whether these disparities occur based on the AI/AN ethnicity or if the disparities are a result of the “rural culture”, which historically has less and/or sometimes difficult access to health services compared to those persons living in metropolitan settings. To fully answer that question, further detailed comparisons of health data between the AI/AN and non-AI/AN population in rural Oklahoma settings would be necessary.

In conclusion, there are clearly disparities in the incidence of cancer in AI/ANs in Oklahoma and within regions of Oklahoma. This study highlights some of the areas where disparities exist, including gender, age, primary payer, site, staging, and geographic location. This study presents more questions than answers as it is the first study to describe the differences in cancer incidence between whites and AI/ANs in Oklahoma. Future studies should expand on these areas and investigate how these factors work together to better understand how the risk of cancer differs between whites and AI/ANs in Oklahoma.