Thread by @ewanbirney

With Scottish, English and South African data all in hospitalisation risk given infection all coming in below 50% (range I think 80% lower SA to 60% lower, English some endpoints) this key parameter is firming up. Frustratingly in the balance from my reading of the SPI-M models

(plus "what does xx% lower mean - xx% per infection or per equivalent infection knowing that Omicron reinfects etc," and how does one factor this vs vaccination and age - so much detail here to nail down)

Basically, good news, and provides narrower spaces for models (both forward models and backward models on infection levels as hospitalisations are more completely ascertained than cases etc).

I doubt however we'll have resolution on best path forward though for a while.