Here is the SPI-M-O paper (7 Feb) which modelled easing NPI restrictions in stages. It combines the ICL and Warwick work and will have been foundation for deliberations and UK Government decisions shared on 22 Feb.
https://assets.publishing.service.g...lling_on_scenario_for_easing_restrictions.pdf
* All four scenarios modelled lead to a substantial resurgence in hospital admissions and deaths later this year.
*The scale and timing of these resurgences are critically dependent on very uncertain modelled assumptions, including real world vaccine effectiveness against severe disease and infection; vaccine coverage and rollout speed; behavioural factors; and the extent to which baseline measures (which could be voluntary) continue to reduce transmission once restrictions are lifted. Given this uncertainty, it would be inadvisable to tie changes in policy to dates instead of data.
* As restrictions are relaxed virus transmission will increase. The more slowly restrictions are relaxed, the greater the number of hospitalisations and deaths prevented by vaccination and hence it would be less likely that restrictions would need to be reimposed.
* A significant proportion of adults will remain who are not directly protected by vaccination.
e.g.: Proportion of adults directly protected against severe disease = 85% (assumes 15% not vaccinated (unable or unwilling)
Vaccine efficacy against severe disease x 79% (average of Pfizer and Oxford-AZ).
So final coverage = 67%
But only around 79% of the population are adults (53M over 18), which means population level protection is lower, at 53%. In addition, protection against infection is likely to be lower than that against disease. As a result, herd immunity is not reached without a large resurgence of transmission.
Comment: This relies on the unvaccinated catching it in numbers and developing antibodies.
Comment: If trials show vaccine can be used on under 16s then herd immunity becomes more achievable - once that cohort (12+M) is vaccinated.
Note: Modelling doesn't try to take account of seasonal effects (beneficial effect), any waning immunity (detrimental affect), or the adverse effect of any VoC that emerge in significant amounts in UK. There is emerging evidence (so less uncertainty) about vaccine effectiveness against serious disease (hospitalisation) and against transmission.
The paper perforce makes loads of assumptions and plenty of sensitivity analysis. But I'll paste one graph - the current England plan is slightly more prudent than the blue line in the graph below. And hospital occupancy is currently (8 Mar) about 80% of the black line (ie better than the model assumes). And summer's coming.
Assumes baseline (after all easing) measures reduce transmission to equivalent of (old) Tier 1. Assumes 4m doses per week from 22 Mar and uptake of 95% (over 80s), 85% (others). From the Warwick model.