Then, focusing on hospitalisations & deaths, we get this.

We see we get only 12,029 post-lockdown deaths - a bit less than for a modest flu season. There are quite a lot of hospitalisations, though, peaking at 4251/day. That could test NHS capacity. Maybe with Nightingales reactivated they'd cope but I doubt they want to try that.

There are a couple of reasons that hospitalisations peak might get smoothed or cut. First, although R0=4 doesn't seem extravagant, especially given full release might trigger a large amount of pent-up spreading activities, it would be summer. Seasonality might materially cut R.

A potentially bigger factor, however, is behavioural response. A lot of people wouldn't want to end up in hospital even if they weren't going to die. And experience over the pandemic has taught them what to do to limit spread.

So suppose that whenever hospitalisations went >500/day people started to moderate their behaviour & when hospitalisations went >2500/day they panicked enough to take R all the way to 1. Then we'd get this. Note the curious wiggles such endogenous behaviour models produce.

Overall the msg of these extended models is that the concern about a summer 3rd Wave isn't silly, but if the vaxxes work ok it shldn't produce many deaths & the hospital capacity issues, if they wldn't be automatically moderated by behaviour change, shld be feasible to plan for.

One last observation is that in these models nothing bad happens until that final step. Maybe the govt's hope is that TTI can just shave enough off R in the final step, if case numbers prior to final relaxation are low enough, to smooth out that last hospitalisations phase a bit.