Thread by @puddleg

Here's a thread on the concept of "causality" used in the LDL cholesterol argument, explaining why people end up at cross-purposes.
I will use the example of the recent Waiho bridge collapse. 1/-
radionz.co.nz/news/national/…

This bridge, a multi-span Bailey bridge for the engineers out there, collapsed when its supports were wrecked by rocks (and ice) driven downstream by a flood, caused by a truly massive rainfall dump. Collision with the rock was the direct "cause" of failure. 2/-

The rock is a discreet, identifiable unit, comparable with ApoB in the atherosclerosis model. The flood is comparable with excess insulin and glucose; rocks are always there, rain can just turn up. 3/-

GWAS is our map equivalent. On the map, we can see that the river bed is rocky, and that rocky river beds greatly increase the risk of bridge failure. Rain is not on the map, although factors that may focus a flood on the bridge are. 4/-

So we can predict that bridges built on sandy river beds will fail less often, and that removing rocks before floods might help mitigate the damage. But that doesn't alter the fact of the strict temporal association between floods and bridge collapses at Waiho in its history. 5/-

But there's something we're missing - the flood caused other damage. It damaged houses, and it broke up an old landfill and polluted 100 km of pure unspoiled Westland coastline. The rocks did no other harm whatsoever besides the failure of the Waiho bridge. 6/-

And, whereas hyperinsulinaemia and hyperglycaemia will increase the risk of multiple causes of death and disability, including cancer, CVD, diabetes, dementia, I can't find much evidence linking LDL to disease processes other than athero - it's even protective for T2D 7/-

So, assuming you could either a) remove all the large rocks from the stream above the next Waiho bridge, or b) build a diversion that prevented future floods, which would be the better use of Westland rate-payers' money? 8/8