I wrote fifteen hundred words about high heels in 2008. The central mechanism doesn't hold up.

The story goes like this, and you've heard it, because everyone has. Heel goes up. Pelvis tips forward. Lower back arches harder to compensate. Hence back pain.

It is a beautiful mechanism. You can draw it on a napkin. I have drawn it on a napkin.

In 2017 a group in Zurich put forty women through it with proper measurement — barefoot, four centimetres, ten centimetres — and found lumbar curvature went down, not up.

The following year a German group tested it again in women wearing heels of seven to eleven centimetres and reported, in about as flat a sentence as you'll find in a biomechanics journal, that their hypothesis of increased lumbar lordosis "had to be rejected for standing as well as for walking." What moved was the pelvis, by roughly one degree — and it moved backwards.

And it turns out somebody noticed this a while ago. In 2010 a researcher went through the literature and found three studies showing lordosis decreased, three showing no change, and three showing an increase — with all three of the increase studies having something wrong with them. He then wrote a paper whose subtitle is, and I want you to appreciate this, "a discussion of the disconnect between Internet content and peer-reviewed literature."

I was the internet content.

One caution while I'm correcting myself, so I don't simply install a new dogma: the honest position is not "heels flatten your back." It's that the popular increase claim isn't supported, and studies find either no change or a decrease. The author of that review called the literature equivocal, and he was right.

Bunions. Obviously the heels, right?

The largest study of its kind — 2,627 women aged 50 to 89 — found hallux valgus associated with narrow and very narrow toe boxes worn in a woman's twenties and thirties. The odds roughly doubled. And it found no association with heel height at either age period.

It is the shape of the front of the shoe, not the height of the back of it, that the best evidence points at. Heels and narrow toe boxes are almost perfectly confounded in real footwear, which is exactly why everyone blamed the visible half.

Nothing here is an argument for throwing out your shoes, and I notice that 2008 me spent several hundred words being fairly smug at women who weren't going to. That was neither useful nor particularly attractive, and I'd like it struck from the record.

What I'd say instead:

Duration matters more than height. In fifty habitual wearers over a six-hour day, back discomfort appeared first at around two and a quarter hours and general discomfort by about three and a half. Below about seven and a half centimetres, reported pain stayed under one out of ten. One study, young women, so hold it loosely — but "how long" is clearly doing more work than the industry's obsession with "how high."

Look at the toe box. That's where the best foot evidence actually points, and almost nobody shops for it.

If flats feel wrong, that's information, not damage. Reduce height gradually and give the calf and ankle some real work. The structural studies that found shortened calf fascicles studied women wearing heels essentially all working hours, for years — not occasional wear.

And if the ankle has genuinely lost range, that is assessable and often changeable, which is the one part of this whole subject where somebody like me is any use at all.

So the arched-back story doesn't survive testing, the arthritis story was never demonstrated, the bunions are about the front of the shoe, and the thing that actually changes is an ankle that stops absorbing and a knee and hip that pick up the difference.

Which is a smaller, quieter claim than the one I made in 2008. It also has the advantage of being what the measurements say.

Anyway. You know it now. And you're going to think about your toe box in a shoe shop for the rest of your life, which I'm afraid is on me.