Adolphe Quetelet's 1830s Convenience Still Beats Superior Science

The metric sitting in your electronic health record is nearly two hundred years old.

A Belgian astronomer named Adolphe Quetelet invented BMI in the 1830s by dividing weight by height squared — a mathematical convenience for population surveys when individual body composition could not be measured at scale. It persists today not because it predicts cardiovascular disease particularly well (it does not), but because it takes thirty seconds, requires no training. Can be entered into a form by anyone with a scale and a pencil.

In the 1990s, a rival challenger emerged with credible superiority. Waist-to-hip ratio, championed across multiple prospective studies, showed stronger correlation with heart disease risk than BMI in real patients followed for years — and a 2001 study published in Lancet followed nearly 30,000 people across eleven countries and found it outperformed BMI at predicting myocardial infarction.

By every scientific measure, the replacement should have happened. It did not. The barrier was not doubt. It was infrastructure. Waist-to-hip ratio requires a trained technician with a measuring tape present at the appointment. It cannot be self-reported reliably over the phone, insurance companies cannot underwrite it remotely. Electronic health records had no field to populate because no workflow existed to populate it consistently.

The metric does not compete alone.

Today's challengers. Body composition assessment, metabolic markers, visceral fat quantification. Face the identical obstacle wearing different clothes. Each one predicts disease risk better than BMI in controlled studies. Each one also requires either equipment, time, or expertise that clinical practice does not have built in. The actual question is not whether a better metric exists. It is whether institutions retrofit their infrastructure for superior evidence or keep using the tool that fits the system they already have. The metric does not compete alone. It competes as part of a locked system. Superior performance in isolation means almost nothing if the new tool cannot be inserted into the existing workflow without reorganizing the entire apparatus around it.