What is r? r is the Pearson correlation coefficient — a number between −1 and +1 measuring how closely two lines move together.
r = 1.0 is a perfect match. r = 0.9+ is very strong. r = 0.7–0.9 is strong. r = 0.2 is weak — essentially no meaningful relationship.
These are ecological (population-level) correlations. They show association, not proof of individual causation — but they are scientifically meaningful evidence of a shared biological trend across three large populations over nearly a century.
r = 0.96
USA
IR ↔ T2DM
Near-perfect
r = 0.94
UK
IR ↔ T2DM
Near-perfect
r = 0.95
Europe
IR ↔ T2DM
Near-perfect
Chart 1 — Estimated Insulin Resistance Prevalence
% of adults · 1925–2020
← % of adult population estimated to have insulin resistance
▲ Shaded zone pre-1985 = proxy estimates only (HOMA-IR not yet formalised)
USA
~5% (1925 est.) → 38.4% (NHANES 2017–18)
United Kingdom
~4% (1925 est.) → 30% (2020 est.)
Europe average
~4% (1925 est.) → 29% (2020 est.)
▲ Filled points = survey-anchored data (post-1985) · Open points = proxy estimates from obesity, T2DM incidence & glucose tolerance surveys (pre-1985)
Chart 2 — Diagnosed Type 2 Diabetes Prevalence
% of adults · 1925–2020
← % of adult population with diagnosed Type 2 diabetes
▲ Shaded zone pre-1980 = limited diagnostic data available
USA
~0.5% (1925) → 11.3% (2020, CDC)
United Kingdom
~0.3% (1925) → 7.4% (2020, Diabetes UK)
Europe average
~0.3% (1925) → 8.5% (2020, IDF Atlas)
▲ T2DM was largely undiagnosed before the 1950s — pre-1950 figures are conservative estimates. Modern figures reflect diagnosed prevalence only; true prevalence including undiagnosed is estimated 20–30% higher.
Why This Correlation Matters Clinically
The near-perfect correlation between insulin resistance and Type 2 diabetes (r = 0.94–0.96 across all three regions)
is not surprising to metabolic scientists —
Type 2 diabetes is insulin resistance that has progressed to the point
where the pancreas can no longer compensate.
They are not two separate diseases: they are early and late stages of the same condition.
This means that
insulin resistance is the clinically actionable window — the reversible phase
that precedes the irreversible pancreatic exhaustion of established Type 2 diabetes.
Identifying and reversing insulin resistance through dietary intervention is therefore not merely
preventive care — it is interception of a disease process already underway.
Reading the two charts together
The shape of the curves is the argument. Look at Chart 1 and Chart 2 side by side. The insulin resistance lines and the Type 2 diabetes lines rise in almost identical arcs — decade by decade, country by country. The correlation coefficients (r = 0.94–0.96) confirm what the eye immediately sees: these two epidemics are tracking together so closely that they are effectively measuring the same underlying metabolic process at different stages of severity.
Insulin resistance always precedes Type 2 diabetes — by 10 to 20 years. This is the critical clinical window. During that period, the pancreas is compensating by producing increasing quantities of insulin to overcome cellular resistance. Blood glucose may remain technically normal or only slightly elevated — easily missed by standard fasting glucose tests. Yet the metabolic damage is accumulating: endothelial dysfunction, chronic inflammation, visceral fat expansion, and progressive beta-cell stress are all advancing. HOMA-IR testing detects this phase. Standard NHS screening largely does not.
The gap between the two curves is the opportunity. At any given decade, the IR prevalence is substantially higher than the T2DM prevalence — because not everyone with insulin resistance has yet progressed to T2DM. That gap represents the population who are metabolically damaged but still in the reversible window. In the USA in 2020, approximately 38% of adults had insulin resistance while 11% had diagnosed T2DM — meaning roughly 27% of the adult population was in that critical intervention window. That is tens of millions of people.
Reversal is the correct clinical goal — not management. Because insulin resistance precedes T2DM by years and is driven by dietary and lifestyle factors, it is physiologically reversible through the same mechanisms. Eliminating refined sugar, reducing processed carbohydrate load, establishing structured fasting periods, and restoring insulin receptor sensitivity through constitutional nutrition can halt and reverse the trajectory shown in Chart 1 — and in doing so, prevent the progression shown in Chart 2 from ever completing.
⚠ Note on Data Availability
Insulin resistance prevalence before 1985 cannot be directly measured — the HOMA-IR model was not published until Matthews et al. (1985).
Pre-1985 IR estimates are back-projected from proxy markers: population obesity rates (WHO), T2DM incidence trends, glucose tolerance survey data, and metabolic syndrome criteria.
They should be treated as directional estimates with wide confidence intervals, not precise measurements.
Post-1985 data is anchored to NHANES III (1988–94), NHANES continuous surveys (1999–2018), NHS Health Survey for England (2003–2021), and the Frontiers systematic meta-analysis (2025).
T2DM figures before 1950 are conservative estimates; the condition was substantially under-diagnosed before systematic population screening.
```
Data sources:
Insulin Resistance: NHANES III 1988–94; NHANES 1999–2018 (PMC11601873 — IR prevalence 24.8%→38.4%); Frontiers meta-analysis 2025 (PMC12411212 — global IR 26.53%); NHS Health Survey for England 2003–2021; WHO obesity trends and Lancet Diabetes pooled analysis 1980–2014; pre-1985 back-projections from proxy markers.
Type 2 Diabetes: CDC National Diabetes Statistics Reports (USA 1980–2020); Diabetes UK Statistics (UK prevalence 1935–2020); IDF Diabetes Atlas 10th Edition 2021 (Europe); WHO Global Health Observatory; Hex N et al. Diabetic Medicine 2012 (historical UK); Gregg EW et al. NEJM 2014 (US trends).
HOMA-IR: Matthews DR et al. Diabetologia 1985;28:412–419.
Correlation r = Pearson coefficient computed from plotted data series (post-1985 anchored data). Ecological correlations — association does not prove individual causation.
```
