Tuberculosis in Indigenous Persons — United States, 2009–2019

Abstract

Background

Populations of indigenous persons are frequently associated with pronounced disparities in rates of tuberculosis (TB) disease compared to co-occurring nonindigenous populations.

Methods

Using data from the National Tuberculosis Surveillance System on TB cases in U.S.-born patients reported in the United States during 2009–2019, we calculated incidence rate ratios and risk ratios for TB risk factors to compare cases in American Indian or Alaska Native (AIAN) and Native Hawaiian or other Pacific Islander (NHPI) TB patients to cases in White TB patients.

Results

Annual TB incidence rates among AIAN and NHPI TB patients were on average ≥10 times higher than among White TB patients. Compared to White TB patients, AIAN and NHPI TB patients were 1.91 (95% confidence interval (CI): 1.35–2.71) and 3.39 (CI: 1.44–5.74) times more likely to have renal disease or failure, 1.33 (CI: 1.16–1.53) and 1.63 (CI: 1.20–2.20) times more likely to have diabetes mellitus, and 0.66 (CI: 0.44–0.99) and 0.19 (CI: 0–0.59) times less likely to be HIV positive, respectively. AIAN TB patients were 1.84 (CI: 1.69–2.00) and 1.48 (CI: 1.27–1.71) times more likely to report using excess alcohol and experiencing homelessness, respectively.

Conclusion

TB among U.S. indigenous persons is associated with persistent and concerning health disparities.

Appendices

Appendix 1

Population data for rate calculations were obtained from the U.S. Census Bureau’s American Community Survey (ACS) through queries of the Census Bureau website (data.census.gov/mdat) on January 14, 2021. These data are provided in Online Resources 6, 7 and 8 for reference.

Crude annual TB case rate calculations:

To calculate crude annual TB case rates (per 100,000 persons) by patient race/ethnicity group, we obtained annual population estimates from the public use microdata sample (PUMS), 5-year population estimates using queries involving the RAC1P, RAC2P, RAC2P12, RAC3P, HISP, and NATIVITY variables and person-weight weighting as follows:

• A. For the AIAN, Asian, Black, and White groups, we obtained annual population estimates using queries involving the RAC1P, HISP, and NATIVITY variables (e.g., https://data.census.gov/mdat/#/search?ds=ACSPUMS5Y2019&cv=NATIVITY(1),HISP(01)&rv=RAC1P&wt=PWGTP). For the HISP and NATIVITY variables, we only included counts associated with the “Not Spanish/Hispanic/Latino” and “Native” categories, respectively. For the AIAN group, we summed counts associated with three RAC1P categories (“American Indian alone,” “Alaska Native alone,” and “American Indian and Alaska Native tribes specified; or American Indian or Alaska Native, not specified and no other race”) to generate the population estimate for each year.

• B. For the Hispanic group, we obtained annual population estimates using queries involving the HISP and NATIVITY variables (e.g., https://data.census.gov/mdat/#/search?ds=ACSPUMS5Y2019&cv=NATIVITY(1)&rv=HISP(02,03,04,05,06,07, 08,09,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24)&wt=PWGTP). For the HISP variable, we included counts associated with all categories except “Not Spanish/Hispanic/Latino.” For the NATIVITY variable, we only included counts in the “Native” category. We summed counts associated with all 23 HISP categories to generate the population estimate for each year.

• C. For the NHPI group, we obtained annual population estimates using queries involving the RAC1P, RAC2P, RAC2P12, RAC3P, HISP, and NATIVITY variables. We did this as follows to obtain estimates as consistent as possible with our NHPI TB case definition, which excluded some NHPI subraces:

  • i. For years 2016–2019, we first obtained annual population estimates using queries involving the RAC2P, HISP, and NATIVITY variables (e.g., https://data.census.gov/mdat/#/search?ds=ACSPUMS5Y2019&cv=NATIVITY(1),HISP(01)&rv=RAC2P(60,61,62,63,64,65,66)&wt=PWGTP). For the RAC2P variable, we included counts associated with all seven NHPI-related categories (i.e., NHPI subraces): “Native Hawaiian alone,” “Samoan alone,” “Tongan alone,” “Guamanian or Chamorro alone,” “Marshallese alone,” “Fijian alone,” and “Other Native Hawaiian and Other Pacific Islander.” For the HISP and NATIVITY variables, we only included counts associated with the “Not Spanish/Hispanic/Latino” and “Native” categories, respectively. Next, we compared these annual population estimates to those obtained for the “Native Hawaiian and Other Pacific Islander alone” category using the queries involving the RAC1P, HISP, and NATIVITY variables (described above in section 1A). The population estimates generated using the RAC2P queries and the RAC1P queries were the same in each of the four years during 2016–2019. Given this concordance, we summed the counts associated with the three RAC2P categories that we sought to exclude from our NHPI population estimates (“Tongan alone,” “Marshallese alone,” “Fijian alone”) in each year during 2016–2019 and subtracted that sum from the population estimate obtained using the RAC1P query for that year.

  • ii. For years 2012–2015, we first obtained annual population estimates using queries involving the RAC2P12, HISP, and NATIVITY variables (e.g., https://data.census.gov/mdat/#/search?ds=ACSPUMS5Y2015&cv=NATIVITY(1),HISP(01)&rv=RAC2P12(60,61,62,63,64,65,66)&wt=PWGTP). For the RAC2P12 variable, we included counts associated with all seven NHPI-related categories (i.e., NHPI subraces): “Native Hawaiian alone,” “Samoan alone,” “Tongan alone,” “Guamanian or Chamorro alone,” “Marshallese alone,” “Fijian alone”, and “Other Native Hawaiian and Other Pacific Islander.” For the HISP and NATIVITY variables, we only included counts associated with the “Not Spanish/Hispanic/Latino” and “Native” categories, respectively. Next, we compared these annual population estimates to those obtained for the “Native Hawaiian and Other Pacific Islander alone” category using the queries involving the RAC1P, HISP, and NATIVITY variables (described above in section 1A). The population estimates generated using the RAC2P12 queries and the RAC1P queries were different in each year during 2012–2015. Given this discordance, we summed the counts associated with the three RAC2P12 (or RAC2P) categories that we sought to exclude from our NHPI population estimates (“Tongan alone,” “Marshallese alone,” “Fijian alone”) in each year during 2012–2019 and quantified the percent of the total population estimate obtained using the RAC2P12 (or RAC2P) query that this sum represented. For each of the three categories that we sought to exclude, these percentages were highly consistent across the seven years (mean+stdev: “Tongan alone” = 5.77%±0.37%, “Marshallese alone” = 2.48%±0.42%, “Fijian alone” = 2.76%±0.19%). Given this consistency we summed the counts associated with the three RAC2P12 categories that we sought to exclude in each year during 2012–2015, calculated the percent of the total population estimated using the RAC2P12 query that this represented, and then reduced the population estimate obtained using the RAC1P query for that year proportionally.

  • iii. For years 2009–2011, we first obtained annual population estimates using queries involving the RAC3P, HISP, and NATIVITY variables (e.g., https://data.census.gov/mdat/#/search?ds=ACSPUMS5Y2011&cv=NATIVITY(1),HISP(01)&rv=RAC3P(02,03,04,05,06)&wt=PWGTP). For the RAC3P variable, we included counts associated with all five NHPI-related categories (i.e., NHPI subraces): “Native Hawaiian alone,” “Samoan alone,” “Guamanian or Chamorro alone,” “Other Pacific Islander alone,” and “Native Hawaiian and Other Pacific Islander groups only.” For the HISP and NATIVITY variables, we only included counts associated with the “Not Spanish/Hispanic/Latino” and “Native” categories, respectively. Next, we compared these annual population estimates to those obtained for the “Native Hawaiian and Other Pacific Islander alone” category using the queries involving the RAC1P, HISP, and NATIVITY variables (described above in section 1A). The population estimates generated using the RAC3P queries and the RAC1P queries were the same in each year during 2009–2011, but counts for the three NHPI subrace categories that we sought to exclude from our NHPI population estimates (“Tongan alone,” “Marshallese alone,” “Fijian alone”) were not reported separately. Given the aforementioned interannual consistency in the percent of the annual total population estimate associated with counts for these three NHPI subrace categories during 2012–2019, we summed the three percentages in each year during 2012–2019 (range=10.4–11.4%), averaged these sums across years (mean+stdev: 11.02%±0.003%), rounded this average to the nearest whole percent (11%), and reduced the population estimate obtained using the RAC1P query for each year during 2009–2011 by that proportion.

Age-adjusted annual TB case rate calculations:

To calculate age-adjusted annual TB case rates (per 100,000 persons) by patient race/ethnicity group, we used the direct method, the 2010 U.S. Standard population, and age-stratified annual population estimates from the U.S. Census Bureau’s American Community Survey (ACS). We applied the six-level age categorization scheme used in NTSS: <5 years, 5–14 years, 15–24 years, 25–44 years, 45–64 years, and ≥65 years.

The 2010 U.S. standard population (308,745,538 persons), as estimated using 2010 decennial census data from the U.S. Census Bureau, was partitioned into the six NTSS age categories based on published age-stratified counts [20] as follows: <5 years=20,201,362 persons; 5–14 years=41,025,851 persons; 15–24 years=43,626,342 persons; 25–44 years=82,134,554 persons; 45–64 years=81,489,445 persons; ≥65 years=40,267,984 persons.

We obtained age-stratified annual population estimates from the public use microdata sample (PUMS), 5-year population estimates using queries involving the RAC1P, HISP, NATIVITY, and AGEP variables and person-weight weighting as follows:

• A. For the AIAN, NHPI, Asian, Black, and White groups, we obtained age-stratified annual population estimates using queries involving the RACE1P, HISP, NATIVITY, and AGEP variables (e.g., https://data.census.gov/mdat/#/search?ds=ACSPUMS5Y2019&cv=NATIVITY(1),HISP(01),RAC1P&rv=AGEP&wt=PWGTP). For the HISP and NATIVITY variables, we only included counts associated with the “Not Spanish/Hispanic/Latino” and “Native” categories, respectively. For the AIAN group, we summed counts associated with three RAC1P categories (“American Indian alone,” “Alaska Native alone,” and “American Indian and Alaska Native tribes specified; or American Indian or Alaska Native, not specified and no other race”) to generate the population estimate for each year.

• B. For the Hispanic group, we obtained age-stratified annual population estimates using queries involving the HISP, NATIVITY, and AGEP variables (e.g., https://data.census.gov/mdat/#/search?ds=ACSPUMS5Y2019&cv=NATIVITY(1),HISP(02,03,04,05,06,07, 08,09,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24)&rv=AGEP&wt=PWGTP). For the HISP variable, we included counts associated with all categories except “Not Spanish/Hispanic/Latino.” For the NATIVITY variable we only included counts in the “Native” category. We summed counts associated with all 23 HISP categories to generate the population estimate for each year.

• C. For the NHPI group, age-stratified counts were not available for the annual population estimates obtained using the queries involving the RAC1P, RAC2P, RAC2P12, RAC3P, HISP, and NATIVITY variables (described above in section 1C). To approximate these, we obtained age-stratified annual population estimates for the “Native Hawaiian and Other Pacific Islander alone” category using queries involving the RACE1P, HISP, NATIVITY, and AGEP variables (described above in section 2A) and used these to determine the percent of the total population associated with each of the 6 NTSS age categories. We then calculated the products of these percentages and the annual population estimates obtained using the queries involving the RAC1P, RAC2P, RAC2P12, RAC3P, HISP, and NATIVITY variables to complete the age-stratified approximations.

Appendix 2

Among incident TB cases with a count date during January 1, 2009–December 31, 2019, counted within any of the 50 states or DC, and in a patient classified as being U.S.-born and who reported their race/ethnicity as AIAN, NHPI (or both), 332 were not classified into either the AIAN or the NHPI patient race/ethnicity group because they met one of the follow criteria:

• A. A total of 106 cases were classified into the Multiple/unknown patient race/ethnicity group because they were in patients who reported their race as AIAN or NHPI (or both) and one or more other races (i.e., multiple race/ethnicity). Of these, 76 were in patients who reported their race as AIAN and another race other than NHPI (the most common other races reported by these patients were White (57) and Black (21)), 28 were in patients who reported their race as NHPI and another race other than AIAN (the most common other races reported by these patients were White (16) and Asian (11)), and 2 were in patients who reported their race as AIAN and NHPI.

• B. A total of 9 cases were classified into the Multiple/unknown patient race/ethnicity group because they were in patients who did not self-report their ethnicity (i.e., unknown race/ethnicity). Of these, 7 reported their race as AIAN, 1 reported their race as NHPI, and 1 reported their race as both AIAN and NHPI.

• C. A total of 63 cases were classified into the Hispanic patient race/ethnicity group because they were in patients who reported their ethnicity as Hispanic. Of these, 58 were in patients who reported their race as AIAN, 4 were in patients who reported their race as NHPI, and 1 was in a patient who reported their race as both AIAN and NHPI.

A total of 154 cases were excluded from the analysis entirely because they were in patients who reported a country of birth or an NHPI subrace (or both) not included in our AIAN and NHPI case definitions. Of these, 2 were in patients who reported their race as AIAN and their country of birth as France or India (these patients were likely misreported as AIAN); the remaining 152 were in patients who reported their race as NHPI. These 152 cases were in patients who reported their NHPI subrace and country of birth as follows: 52 Micronesian (of which 23 and 20 reported their country of birth as the United States or the Federated States of Micronesia, respectively), 47 Marshallese (of which 26 and 18 reported their country of birth as the United States or the Marshall Islands, respectively), 19 Chuukese (of which 10 reported their country of birth as the United States), 16 Tongan (of which 15 reported their country of birth as the United States), 4 Samoan, 1 Carolinian, 1 Fijian, 1 Solomon Islander, and 11 cases in patients who did not self-report an NHPI subrace (of which 8, 2, and 1 reported their country of birth as the Marshall Islands, Federated States of Micronesia, or Palau, respectively).