What Changed Since WHI 2025

WHI 2025 was a single photograph of the urban workforce's metabolic and behavioural health. WHI 2026 is the second frame, on a much larger and broader sample — 11,427 biomarker reports against roughly 2,800 last year, 9,097 survey respondents, 74,339 claims. With two frames, the question stops being what is the rate and becomes what moved. The honest answer separates into two columns that look similar on a chart and mean opposite things: the workforce changed, or only the measurement did.

This page draws every metric that has a defensible 2025 baseline — 58 comparable readings across the report, plus eight that are flagged not-comparable — and sorts them by whether the change is real. The discipline matters because a year-on-year report is most useful when it is least flattering: the temptation is to read every downward number as progress and every upward number as alarm. Most of the large swings in this dataset are neither.

The direction summary: more held than moved#

The headline is the balance. Twenty-three metrics held flat and twenty-three fell, but the two columns are not symmetric in what they mean. The flat column is almost entirely the large-sample biomarker headlines — the readings we trust most, on both sides of the comparison. The "lower" column is thinner than it looks: more than half its rows are city-by-city lipid and liver numbers whose 2025 baselines came from samples of a few hundred reports. Those are corrections, not improvements.

What genuinely changed: the behaviour got worse#

The cleanest real movements are on the survey side, where both years draw on samples in the thousands and the same questions.

Stress is the largest single movement in the comparable set: high stress rose 6.5 points across the workforce and 6.2 points among women specifically, who now cross the 47% mark. Short sleep rose 3.4 points. The share doing no structured exercise in a typical week rose to nearly a third. None of these is a sampling story — the survey base is large on both sides and the questions are stable enough to compare directionally.

The supplementation reversal is the one to sit with. Vitamin D supplementation fell almost six points, to one in five professionals — and it fell in a population where 81.4% tested below the sufficiency threshold this year, barely down from 84.0%. The deficiency stayed put; the corrective behaviour retreated.

Two behaviours moved the right way: nicotine use fell 3.3 points and reported alcohol consumption edged down. Both are modest, and the report treats them as directional rather than precise — but in a year where most behavioural arrows pointed the wrong way, they are the two that didn't.

What held: the metabolic floor did not move#

The biomarker headlines are where the larger 2026 sample earns its keep. A rate that survives a four-fold increase in sample size is structure, not noise.

Glucose dysfunction held at 38.7%. Pre-diabetes widened by 1.7 points and the diabetic share was flat. B12 deficiency stayed near 70%, Vitamin D deficiency near 80%, male anemia under 9%. Each of these matched its 2025 value inside the margin a four-fold sample increase allows. The workforce did not bend any of these curves in a year — it walked further along them.

The one biomarker headline that moved enough to flag is female glucose dysfunction, up 3.5 points to 39.9%. The honest read is in the sample: 2025 measured 770 women; 2026 measured 3,262. The cleaner interpretation is that the 2026 number is the more accurate estimate of where women already were — not that women's metabolic health deteriorated in twelve months. Male glucose dysfunction, measured on large samples both years, did not move at all.

What only changed in the measurement: the city rankings reshuffled#

The largest deltas in the entire comparable set are city-level. They are also the least trustworthy. WHI 2025 published a city-by-city ordering built on biomarker samples of a few hundred reports per metro. WHI 2026 maps every report to a city via its pincode and lands samples several times larger. Where the two disagree sharply, the 2026 number is the estimate; the swing is a correction.

Pune was 2025's "healthiest" metro on glucose and Delhi its "sickest". Both reversed. Pune's apparent 13-point rise and Delhi's 10-point fall are the same artefact in opposite directions: 2025 sampled 384 reports in Pune and 130 in Delhi, too few to place either city on the metro ranking with confidence. On the 2026 samples — 2,562 in Pune, 689 in Delhi — Pune sits near the top of the distribution and Delhi near the middle. Neither workforce changed. The measurement did.

What survives the city reshuffle is narrow but real. Mumbai's worsening glucose holds up: its sample is stable across years, and it leads every metro on glucose dysfunction in 2026. Delhi's liver-dysfunction rate replicates almost exactly — 34.8% to 34.4% — and remains the worst among the metros, the one place where the 2025 city finding was solid enough to stand. Everything else in the city column should be read as the 2026 baseline, not as a trend.

The not-comparable column, and why it's honest to show it#

Eight metrics have a 2025 number but no defensible 2026 comparison. The five city average-HbA1c values move by hundredths of a point — too fine to call against this much sampling noise. Two more are 2025 career-stage cuts ("Gen Z normal glucose", "Gen X total glucose risk") built on a generational framing WHI 2026 dropped in favour of actual decade bands. Listing these as not-comparable rather than forcing a delta is the same discipline as the rest of the page: a number you can't trust is worse than a gap you're honest about.

What the second frame shows#

Two photographs of the same workforce, a year apart, on a much bigger camera. The metabolic floor — glucose, lipids by gender, B12, Vitamin D deficiency — held exactly where WHI 2025 found it, which on a four-fold sample is the strongest confirmation the inaugural report could have asked for. The behaviour underneath it got worse: more stress, less sleep, less movement, less supplementation. And the city-by-city map redrew itself, mostly because the 2026 sample is finally large enough to draw it.