When you use a corrected BMI calculator, you're relying on a set of numbers that most people have never heard of: body-segment weight percentages. These figures — that a hand is 0.7% of body mass, a thigh is 16.3%, and so on — are the foundation of the correction. But where do they come from, and why should you trust them?
The story begins in the 1950s at the University of Michigan, where aerospace engineer-turned-biomechanist WT Dempster led a project to measure the physical properties of the human body. Working with human cadavers, Dempster's team carefully dissected individual body segments — hands, forearms, upper arms, feet, lower legs, thighs, and the head/trunk — and measured each one's mass as a percentage of total body weight.
The results were published in a Wright Air Development Center technical report (55-159) that became one of the most cited documents in biomechanics. Despite the limited sample size (the cadavers were predominantly older males), the relative proportions proved remarkably consistent with subsequent studies and remain the standard reference more than 70 years later.
Dempster's data wasn't originally intended for BMI correction. It was designed for engineering applications — designing cockpits, calculating forces on joints, and modeling human movement. But the data turned out to be invaluable for prosthetics, rehabilitation, and nutrition science.
DA Winter's Biomechanics and Motor Control of Human Movement — first published in 1979 and updated through multiple editions — compiled and standardized the Dempster data, making it accessible to practitioners, researchers, and students. Winter's tables, published in 1990, are now the go-to reference for body-segment parameters in practice.
Winter's contribution was organizational and interpretive rather than original data collection. He presented the segment mass data alongside the center of mass and radius of gyration for each segment, creating a complete set of parameters for biomechanical modeling. His textbook has been through four editions and remains in print — a rare longevity for a technical reference.
While Dempster and Winter provided the raw data, it was LA Osterkamp who connected it to the problem of BMI calculation in people with limb loss. In a 1995 paper in the Journal of the American Dietetic Association, Osterkamp demonstrated that standard BMI systematically underestimates weight status in this population and proposed a correction formula based on the Dempster/Winter segment data.
Osterkamp's paper was significant because it translated biomechanical data into a practical tool. Before this work, many providers were either unaware of the BMI discrepancy in people with limb loss or had no practical way to correct for it. The Osterkamp correction gave them a simple, evidence-based method that required only height, weight, and amputation level — data routinely available in practice.
In an era of DEXA scans, bioelectrical impedance devices, and AI-powered body composition analysis, you might wonder why 70-year-old cadaver data is still relevant. The answer is accessibility and transparency.
Modern body composition tools are more accurate but also more expensive, less available, and harder to verify. A DEXA scan requires a trained technician, specialized equipment, and a calibrated device. Bioelectrical impedance results vary by device, hydration status, and individual factors. The Dempster/Winter tables, by contrast, are freely available, mathematically transparent, and can be applied with nothing more than a calculator.
For routine screening — the setting where BMI is most commonly used — the population-average correction provides sufficient accuracy without the cost and complexity of specialized equipment. It's not a replacement for DEXA when precision is critical, but it's far better than ignoring the problem entirely.
Here are the segment weight percentages used in BMI correction, with their Dempster/Winter origins:
The Dempster data comes from a limited cadaver sample — primarily older, male, and likely of European descent. Body composition varies by sex, age, ethnicity, and fitness level. A young, muscular person will have different segment proportions than the average in Dempster's sample. The correction is an estimate, not a measurement, and should be used with that understanding.
For most purposes, the population-average correction is sufficiently accurate. For high-stakes decisions — bariatric surgery candidacy, research protocols — direct body composition measurement remains preferable. But for screening, goal-setting, and general awareness, the corrected BMI based on these tables is a vast improvement over the alternative.
This article is for informational purposes only and is not a substitute for professional medical advice.