Diabetes (T2) 2X to 6X higher prevalence in some dark-skinned or poor people

Claude AI May 2026

Question: What is the higher incidence/prevalence of diabetes in the following groups? Native Americans, African Americans, Africans, Aborigines, Mongolians, Americans living below the poverty line

Bottom line

Type 2 diabetes prevalence in the six target groups ranges from about 0.5× (rural sub-Saharan Africa) to ~7× (Pima Indians of Arizona) the non-Hispanic-white U.S. baseline of ~6.9–7.5%. The most striking gradients are not racial per se — they tracked

  • Westernization,
  • urbanization and
  • poverty almost everywhere studied

Pima vs. Mexican Pima, urban vs. rural Africa, remote vs. non-remote Aboriginal Australia, U.S. <100% FPL vs. ≥500% FPL FPL = Federal Poverty Level).

TL;DR

  • Reference: Age-adjusted diagnosed diabetes in non-Hispanic White U.S. adults is 6.9% (CDC, NHIS 2019–2021); total (diagnosed + undiagnosed) NHANES estimate is ~9–10%.
  • Highest-burden groups (relative to NH whites): Pima Indians of Arizona ~5–7× (≈38–50% of adults), Aboriginal Australians in remote Northern Territory communities ~4–6× (29–40% of adults ≥20 y), Navajo ~3–4× historically, U.S. AI/AN overall ~2.0× (13.6%), African Americans ~1.75× (12.1%), Mongolians ~1.2–1.4× (8–10%), Americans <100% FPL ~2.6× (13.1% vs. 5.1% at ≥500% FPL).
  • Confirms the user's framing: Rural sub-Saharan Africa is genuinely low (often 0.5–1.5× white U.S. baseline, i.e. ~1–3%), 2–4× lower than urban African rates, supporting the "epidemiologic transition" hypothesis.

Key Findings — Comparative Table

Population (year/source) Absolute T2D prevalence Ratio vs. NH-White U.S. (6.9%) Notable variation
Non-Hispanic White U.S. adults (CDC NDSR, NHIS 2019–2021, age-adjusted) 6.9% diagnosed; ~9–10% total (NHANES Aug 2021–Aug 2023) 1.0× (reference) Men 7.7%, women 6.2%
All American Indian / Alaska Native (AI/AN) adults (CDC NDSR 2019–2021; IHS NDW) 13.6% ~2.0× IHS 2017: 14.6%; declined from peak 15.4% (2013); regional range 5–25%
Navajo adults ≥20 y (Navajo Health & Nutrition Survey, 1997; Brown UJPH 2023 update) 22.9% age-standardized in 1997; >40% in those ≥45 y; combined diabetes+pre-diabetes ~50% in 2020 ~3–4× Women > men; rises sharply with age
Pima Indians, Gila River, Arizona (NIDDK / Schulz 2006; Bennett 1971; Pearson 2015) ~38% age/sex-adjusted adults; ~50% in adults ≥35 y ~5–7× (highest documented in world) Mexican Pima, similar genes but traditional lifestyle: only 6.9% — proves environmental dominance
Alaska Natives (overall) ~6–9% (lower than Lower-48 AI/AN; some IHS regions report ~16–17%, others (Alaska) lower) ~1.0–1.3× High regional heterogeneity; Yup'ik historically very low (<2%)
African Americans (NH Black) U.S. (CDC NDSR 2019–2021; NHANES) 12.1% diagnosed; total (NHANES 2009–2018) ~14.6% ~1.75× diagnosed (~1.6× total) Black women > Black men in some surveys; ESRD from diabetes 2× whites
Sub-Saharan Africa — RURAL (Lancet/SSA reviews; Aspray Tanzania; Maher Uganda; Price Malawi 2018; AWI-Gen 2024) Typical range 1–4%; classic studies: 0.87% rural Tanzania (1989), 0.6–1.4% rural Uganda, ~2% rural Malawi (2013–16), 3.6% rural Ghanaian men/5.5% women (RODAM) ~0.15–0.6× (well below white U.S.) Doubles to triples in urban centers; AWI-Gen middle-aged (40–60 y) pooled rural+peri-urban: 5.5% baseline, doubling within 6 y
Sub-Saharan Africa — URBAN (for contrast) (same sources) ~6–18%: urban Senegal 17.9%; urban Ghana ~10%; urban Kenya 12%; urban Malawi 3% ~1–2.5× Often 2–4× higher than rural counterparts in same country
Aboriginal & Torres Strait Islander Australians (ABS NATSIHMS 2022–24; AIHW; Burrow & Ride; Hare 2022) Adults ≥18 y: 15.5% (measured); Indigenous Eye Health Survey ≥40 y: ~37% self-report ~3× (vs. ~6.1% non-Indigenous; AIHW: "2.9× as likely after age-adjustment") Remote 21.7% vs. non-remote 14.3%; Torres Strait Islander adults 8.6% vs. Aboriginal 7.2% (NATSIHS 2022–23 self-report)
Aboriginal Australians, REMOTE Northern Territory (Hare et al., BMJ Open Diabetes 2022) Adults ≥20 y: 28.6%; Central Australia 39.5%; Top End 24.2% ~4–6× Among highest documented worldwide; women > men; very high youth-onset T2D
Mongolia (national) (WHO STEPS 2019; Dovepress 2023 OGTT study; Tandfonline 2018) STEPS 2019 (15–69 y): 8.3%; reanalyzed: 9.5% (urban 8.9%, rural 7.7%); OGTT-confirmed adults ≥30 y: ~10% ~1.2–1.5× Tripled since 1999 (3% → ~10%); pre-diabetes 17.4%; men > women; Inner Mongolian Chinese rural: only 3.7%
U.S. adults below 100% FPL (CDC NDSR 2019–2021; PCD CDC 2026; NHIS) 13.1% age-adjusted diagnosed (vs. 5.1% at ≥500% FPL) ~2.6× vs. high-income; ~1.9× vs. NH-white average In 2011–2014 the relative increase in DM prevalence vs. high-income was +100% for poor, +74% for near-poor, +40% for middle-income

1. American Indian : Pima 6X, Navajo 3.5X, other 2X

The CDC National Diabetes Statistics Report (2019–2021 NHIS + IHS data) puts age-adjusted diagnosed diabetes in AI/AN adults at 13.6%, roughly double the non-Hispanic-white rate of 6.9%. The Indian Health Service active-clinical-population analysis (Bullock et al., BMJ Open Diabetes 2020) showed AI/AN prevalence rose from 14.4% (2006) to 15.4% (2013), then fell to 14.6% in 2017 — the first documented decline in this population, attributed in part to the IHS Special Diabetes Program for Indians. The IHS's own communications cite a 2.3× ratio vs. non-Hispanic whites and note that "in some communities prevalence is as high as 60%." Lifetime risk for AI/AN children born around 2000 is estimated at ~50%.

Tribal extremes: - Pima Indians (Gila River, Arizona): The most-studied group in the world. Bennett's 1971 Lancet paper documented ~50% prevalence in those ≥35 years using OGTT — the highest ever recorded. Schulz et al. (Diabetes Care 2006) reported age/sex-adjusted prevalence of 38% in Arizona Pima vs. only 6.9% in genetically similar Mexican Pima living traditionally — one of the strongest natural experiments in diabetes epidemiology. Modern figures: ~50% of Arizona Pima adults aged ≥35 y. - Navajo: The 1997 Navajo Health and Nutrition Survey (Will et al., Journal of Nutrition) found age-standardized prevalence of 22.9% in adults ≥20 y, with >40% of those ≥45 y affected — about 4× the contemporary U.S. rate. More recent reports describe ~50% of Navajo adults having either diabetes or pre-diabetes (2020). - Alaska Natives historically had much lower rates than Lower-48 tribes (e.g., 6% in some Alaska studies vs. 24% in southern Arizona tribes), though they have been catching up.

2. African Americans 1.75X

CDC NDSR (2019–2021): 12.1% diagnosed, vs. 6.9% in non-Hispanic whites → 1.75× higher. NHANES total (diagnosed + undiagnosed) data 2009–2018 gave ~14.6% in Black participants vs. ~10.6% in whites (~1.4× total). The HHS Office of Minority Health (2024) reports Black/African American adults are 24% more likely than U.S. adults overall to have diabetes, 78% more likely to die of it, and >2× more likely to develop diabetic kidney failure. The disparity has widened over time: in 1999–2002 the Black/white ratio for total diabetes was ~1.6×, by 2009–2012 it had grown.

3. Rural Sub-Saharan Africa (continental Africans, rural setting)

The user's framing is well-supported. Multiple high-quality studies show rural prevalence dramatically lower than urban prevalence in the same countries:

  • Rural Tanzania (McLarty et al., Lancet 1989; Aspray 2000): 0.87% adults ≥15 y (1989); rural 1.7% vs. urban 5.9% (2000).
  • Rural Uganda (Maher 2010; Mayega 2012): 0.6–1.4% — among the lowest rural rates documented.
  • Rural Malawi (Price et al., Lancet Diabetes & Endocrinology 2018): 2% men and 2% women in rural Karonga vs. 3% in urban Lilongwe.
  • Rural Ghana (RODAM study, 2016): 3.6% men, 5.5% women — vs. 10.3% men, 9.2% women in urban Ghana, and 21.4%/54.2% obesity in Ghanaian migrants in London.
  • West Africa pooled (5 countries, WHO STEPS, BMJ Open 2023): age/sex-adjusted T2DM 2.5% rural vs. 6.2% urban (RR urban vs. rural ~2.5).
  • Systematic review (Hilawe et al., 2019): Pooled undiagnosed-DM 3.93% rural vs. 8.68% urban — urban roughly double rural across SSA.
  • AWI-Gen (Lancet Global Health 2024/2025): Among adults aged 40–60 y in South Africa, Kenya, Ghana, and Burkina Faso, baseline T2D prevalence was 5.5%, but doubled within ~6 years — the rural advantage is closing rapidly.

Implied ratio vs. non-Hispanic whites: Truly rural, traditional-lifestyle SSA populations typically run 0.15–0.6× the U.S. white rate (i.e., 1–4% vs. ~7%). This is the only group in the analysis where the ratio is materially below 1.

4. Aboriginal 5X

  • National adult prevalence (ABS NATSIHMS 2022–24, biomarker-confirmed): 15.5% of Indigenous adults ≥18 y — up from 11.1% a decade earlier. Males 16.7%, females 14.3%; remote 21.7%, non-remote 14.3%; ages 55+ 34.6%.
  • Self-report (NATSIHS 2022–23): 7.4% overall, with Torres Strait Islanders 8.6% slightly higher than Aboriginal 7.2% — but biomarker surveys show self-report substantially under-reports.
  • AIHW summary: Indigenous adults are ~2.9× as likely as non-Indigenous adults to have diabetes after age adjustment (17% vs. 6.1%). The National Indigenous Eye Health Survey (≥40 y) found age-adjusted prevalence of 43.8% in Indigenous vs. 11.5% non-Indigenous (~3.8× ratio).
  • Remote Northern Territory (Hare et al., BMJ Open Diabetes 2022, linked-data study of 21 267 Aboriginal residents): Adult (≥20 y) prevalence 28.6% in 2018/19, with Central Australia at 39.5%, comparable to historical Pima rates; type 2 accounted for 99.6% of incident cases.

Ratio vs. non-Indigenous Australians (≈ comparable to NH-white U.S.): 3–4× nationally, ~5–6× in remote NT communities.

5. Mongolia

  • WHO STEPS 2019 (ages 15–69): 8.3% by fasting glucose; a reanalysis of the same data covering ages ≥18 reported 9.5% overall (urban 8.9%, rural 7.7%) with pre-diabetes 17.4%.
  • OGTT-confirmed national survey (Dovepress 2023, ages ≥30): age-standardized diabetes prevalence ~10% — a 3.2× rise since 1999 (when it was ~3%).
  • Ulaanbaatar (Plos One 2025, STEPS-style 2018): 6.27% in adults ≥20, men 7.5% > women 5.3%.
  • Inner Mongolian rural Chinese (Han + Mongol mixed, 2009 cross-section): much lower at 3.7%, with high IFG (18.5%), suggesting traditional pastoralist lifestyle still protective.

Ratio vs. NH-white U.S. (6.9% diagnosed; ~9% total): roughly 1.0–1.4× depending on metric. Mongolia is a clear example of a country mid-transition: rural nomadic populations historically had very low rates, urbanization is driving prevalence up rapidly. The IDF Western Pacific 2024 Region (which includes Mongolia) average is 12.4%.

6. Americans Below the Federal Poverty Line (1.9x) 2.6X

This is where data are clearest. From the CDC National Diabetes Statistics Report (NHIS 2019–2021, age-adjusted):

Family income vs. FPL Diagnosed diabetes prevalence
<100% FPL (poor) 13.1%
100–299% FPL ~10%
300–499% FPL ~7%
≥500% FPL 5.1% (men 6.3%, women 3.9%)
  • Ratio (poor vs. high-income): 13.1% ÷ 5.1% = 2.57×
  • Ratio (poor vs. NH-white average 6.9%): ~1.9×
  • A separate CDC PCD analysis (2026) of NHIS 2021–2024 found that among adults below the FPL, rural residents 16.1% vs. urban 13.4% had diabetes — i.e., poverty + rurality is additive.
  • The 2016 MMWR report on disparities (1999–2002 vs. 2011–2014) showed that the absolute income-related prevalence gap widened from 4.7 to 7.1 percentage points. CDC/ADA 2021 analysis: between 2011 and 2014, relative diabetes prevalence vs. high-income was +100% for poor, +74% near-poor, +40% middle-income.

The poverty effect is partly mediated by race/ethnicity (minorities are over-represented in low-income groups) and partly independent (operates within every racial/ethnic group through obesity, food insecurity, physical inactivity, and reduced healthcare access).

Recommendations (for VitaminDWiki content use)

  1. Use the CDC 2019–2021 figures (NDSR) as the primary U.S. comparator (NH-white = 6.9%), and cite NHANES Aug 2021–Aug 2023 for total (diagnosed + undiagnosed) figures (~10–11% adults).
  2. For the Pima Indian "headline" figure, cite both Bennett 1971 (Lancet — original 50% in adults ≥35 y) and Schulz 2006 (Diabetes Care — 38% age/sex-adjusted; the Mexican-vs-Arizona Pima comparison is the strongest argument that lifestyle, not just genetics, drives the disparity). Avoid using "Pima Indians = 50%" without specifying the age band; for the entire population the figure is closer to 38%.
  3. For Aboriginal Australians, use two figures — the national ~15.5% (NATSIHMS 2022–24, all adults) and the remote NT ~29% (≥20 y) with Central Australia 39.5% (Hare 2022). Cite AIHW's 2.9× age-adjusted ratio; note the remote rate is higher than for women than men, opposite of the global pattern.
  4. For rural sub-Saharan Africa, use the systematic review numbers — rural prevalence typically 1–4%, urban often 2–3× higher. Best primary sources: Hilali (BMC Public Health 2011), Price 2018 (Lancet Diabetes Endocrinol), AWI-Gen 2024 (Lancet Global Health), McLarty 1989 (Lancet). State explicitly that rural rates are below the white-U.S. baseline.
  5. For Mongolia, cite WHO STEPS 2019 (8.3%) as the official figure; the Dovepress 2023 OGTT-confirmed study (~10%, ages ≥30) is more rigorous. Note pre-diabetes ~17%.
  6. For poverty, the cleanest single statistic is 13.1% (<100% FPL) vs. 5.1% (≥500% FPL) → ~2.6× gradient (CDC NDSR 2019–2021). This is a robust, government-published figure suitable for wiki content.
  7. Update threshold: If CDC publishes NDSR with 2022–2024 data (likely 2026–2027), figures may shift modestly downward for AI/AN and upward for African Americans/Hispanics — recheck annually.

Caveats

  • Definitional variability: Studies use different criteria (self-report, fasting plasma glucose ≥126 mg/dL, OGTT 2-h ≥200 mg/dL, HbA1c ≥6.5%). HbA1c thresholds underestimate diabetes in some African populations (Khosla 2021), so older sub-Saharan Africa figures using HbA1c alone may underestimate. AWI-Gen and Malawi studies used FPG.
  • Self-report vs. measured: In remote Aboriginal Australia and rural SSA, self-reported diabetes substantially underestimates true prevalence — the NATSIHMS 2022–24 found 1 in 5 Indigenous Australians with biomarker diabetes had not been previously diagnosed; the SSA review by Hilawe et al. found undiagnosed rates 41–75%.
  • Pima Indians "highest in the world" claim: Now contested — Aboriginal Australians in Central Australia (39.5% in those ≥20 y), the Marshall Islands (~31% ≥20 y), and Nauru have approached or exceeded historical Pima rates. The Pima rate has been roughly stable since the 1980s–2000s while others rose.
  • Type 2 vs. all-types: Most national surveys (CDC, AIHW, IDF, WHO STEPS) report "diabetes" without subtype. In adults, type 2 is 90–95%+ of cases in all six target groups, so the figures cited are an excellent proxy for type 2; type 1 in AI/AN is rare (~1% of cases in NT Aboriginal data).
  • AI/AN data are regional: IHS data come predominantly from on-reservation and clinic users; urban AI/AN populations may be undercounted. Tribe-level prevalence varies from ~6% (some Alaska Native communities) to ~50% (Pima).
  • Rural SSA: heterogeneity: Reported rural prevalence ranges from 0.6% (Uganda) to 8.8% (some rural South African studies); the user's "2–3% rural vs. 8–12% urban" framing captures the modal pattern but not the full range. Some rural populations are already mid-transition.
  • Mongolia STEPS 2019 included urban-dominant Ulaanbaatar (~half of national population); pure rural-nomadic-herder prevalence is likely lower than the 7.7% reported, but no recent OGTT-confirmed nomadic-only study exists.
  • Poverty figures: Cross-sectional, so causality runs in both directions (diabetes worsens income, low income increases diabetes risk). The CDC PCD 2023 analysis shows the diabetes/income gradient narrowed 2001–2011 then widened 2011–2018.
  • Some forward-looking figures noted in sources (e.g., IDF projecting Western Pacific +19% prevalence by 2050; AFR Region +142%; ADA 2026 NDSR estimating 40.1 million Americans with diabetes) are projections, not measured prevalence, and should not be presented as established fact.

Related in VitaminDWiki

Americans

Diabetes

image

Increase Health Problems for Blacks

image