BMI was invented to study populations, not individuals — and its creator, Adolphe Quetelet, explicitly said it was unsuitable for judging an individual's health. It took about 150 years for that caveat to be systematically ignored
The previous articles on this site covered why BMI is flawed, childhood BMI percentiles, BMI and ethnicity, and BMI during pregnancy. This article addresses the history of BMI and how a population statistics tool became a clinical screening tool — and what the insurance industry's adoption of BMI in the mid-20th century had to do with that transformation.
Quetelet's original purpose: the "average man" concept
Adolphe Quetelet (1796-1874), a Belgian statistician and astronomer, developed the measure now called BMI in the 1830s. He called it the Quetelet Index and described it as:
"If man increased equally in all dimensions, his weight at different ages would be as the square of his stature."
His purpose was entirely statistical — Quetelet was interested in the statistical properties of human populations, specifically in understanding the distribution of physical characteristics across populations. He was developing what we now call descriptive statistics — means, distributions, normal curves.
He was explicit that the Quetelet Index described populations, not individuals. The "average man" (l'homme moyen) was a statistical construct, not a health ideal. Finding that you deviate from the statistical average of a population is a mathematical fact about the distribution, not a clinical finding about an individual.
The insurance industry's adoption: turning statistics into risk categories
For most of the period from the 1830s to the 1940s, the Quetelet Index was used primarily in academic anthropological and statistical research. It was not used as a clinical screening tool.
The transformation came through life insurance. Insurance companies, seeking actuarial data to price policies, collected extensive height and weight measurements from applicants and correlated them with mortality outcomes in their policyholder databases. By the 1940s and 1950s, major US and UK life insurance companies (particularly Metropolitan Life) had developed "ideal weight" tables — which, implicitly, used body-mass-related ratios to categorize policyholders.
The insurance industry's motivation was risk pricing, not individual health. A policy applicant with a high body mass ratio was, statistically, more likely to die early (on average, across a population) — making them a higher actuarial risk. The individual-level implication (you as a person have elevated health risks) was an extrapolation from population statistics that the data supported on average but couldn't confirm for any specific person.
Ancel Keys and the naming of "Body Mass Index"
The term "Body Mass Index" was coined by physiologist Ancel Keys in 1972, in a paper evaluating different measures of relative body weight. Keys compared several indices and found Quetelet's original (weight/height²) performed best at correlating with body fat percentage in the populations he studied — which is why he recommended it as the standard measure.
Ancel Keys also explicitly described BMI as a population measure — useful for epidemiological studies comparing groups, not for assessing individual health status. His paper's conclusions have been selectively cited over the decades in ways that departed from his stated caveats.
The clinical adoption accelerated through the 1980s and 1990s as health organizations — seeking simple, widely-applicable screening tools — embraced BMI for its practical advantages: it requires only a scale and a measuring tape, no blood tests, no specialist equipment, no training.
The 1998 threshold change: millions became "overweight" overnight
In 1998, the US National Institutes of Health lowered the "overweight" BMI threshold from 27.8 to 25 (aligning with World Health Organization guidelines). This single policy change reclassified approximately 29 million Americans from "normal weight" to "overweight" overnight — without any change in their actual body composition.
The threshold itself is largely arbitrary: it was derived from statistical analysis of population mortality data — populations with BMIs in certain ranges had higher mortality rates on average. But the specific choice of 25 as the cutoff, versus 24 or 26, was a policy decision, not a precise physiological boundary. The mortality risk increase near BMI 25 is modest and gradual, not a sharp step function at exactly 25.
The 1998 change was not based on new scientific findings about individual health at specific BMIs — it was a standardization decision for consistency in international comparisons and clinical guidelines.
What BMI actually predicts — and with what accuracy
At the population level, BMI is a reasonable predictor of several health outcome distributions. Groups with higher average BMI tend to have higher rates of type 2 diabetes, cardiovascular disease, and certain cancers. This is a real statistical relationship.
At the individual level, BMI has substantial limitations as a predictive tool:
- It doesn't measure body fat directly — it measures a ratio of weight to height squared, which correlates imperfectly with actual body fat percentage
- It cannot distinguish muscle from fat — a muscular athlete may have a high BMI but low body fat
- It doesn't indicate fat distribution — visceral fat (around organs) is more metabolically harmful than subcutaneous fat, and BMI doesn't indicate which type predominates
- Its predictive accuracy varies significantly by ethnicity — as covered in the previous ethnicity-specific article, the thresholds appropriate for European-origin populations underestimate risk for some Asian populations at the same BMI
How to use the BMI Calculator on sadiqbd.com
- Understand it as a screening signal, not a diagnosis — a high BMI warrants discussion with a healthcare provider but doesn't itself confirm any health condition
- Interpret in context of other measurements — waist circumference, waist-to-hip ratio, and metabolic markers (blood pressure, blood glucose, lipid panel) provide more diagnostic information than BMI alone
- Apply ethnicity-appropriate thresholds — for individuals of South or East Asian descent, the standard 25/30 thresholds may underestimate risk; the previous article covered recommended adjusted thresholds
Frequently Asked Questions
If BMI is such a crude measure, why do healthcare systems still use it? Practical utility at scale. A measure that requires only height and weight is applicable in primary care settings worldwide, with no equipment or specialist training, at essentially zero cost. Even if BMI is imprecise for individual assessment, it's the only practical option for population-level screening in many healthcare systems — and "imprecise but universally applicable" beats "more accurate but requiring equipment most clinics don't have" for large-scale public health purposes. The goal isn't eliminating BMI from clinical use; it's ensuring both clinicians and patients understand what it does and doesn't tell you, and that it's interpreted alongside other information.
Is the BMI Calculator free? Yes — completely free, no sign-up required.
Try the BMI Calculator free at sadiqbd.com — calculate your Body Mass Index and understand what it does and doesn't indicate.
