Our report on real-time modelling of COVID-19: https://www.imperial.ac.uk/mrc-global-infectious-disease-analysis/covid-19/report-41-rtm/
(not yet peer-reviewed)
Our model is adapted to the situation in England. First: how R was changed by government policy announcements, seeing that only lockdown was capable of suppression (R < 1)
1/10 https://twitter.com/MRC_Outbreak/status/1341516576948940801

The epidemic started earlier in London, and a higher proportion of the population has been infected there, but still far below the herd immunity threshold
2/10

Starting lockdown a week earlier could have prevented around 20,000 deaths in the first wave (though the point of relaxation was less important). This emphasises the importance of acting early, as has been reported many times before
3/10

Showing again why COVID is such a threat: high rates of hospitalisation and death. Some good news is that fatality rate has halved due to improvements in care (such as proning, dexamethasone). Also noted is a lower rate in London, more so than can be explained by age alone
4/10

We specifically modelled care homes, which underwent a parallel epidemic during the first wave. For elderly people in care homes (CHR in the plot), the fatality rate was 3-4 times higher than those of a comparable age
5/10

This model synthesises many lines of evidence: death data, hospitalisations, serology, pillar two testing. On their own they paint a partial picture, but combined in a model which represents COVID-19 transmission allow us to produce directly usable estimates
6/10

Since March, the model has been one of the many sources SPI-M uses to summarise scientific knowledge on COVID each week. It was also used as a key part of the Academy of Medical Sciences winter scenario planning: https://acmedsci.ac.uk/policy/policy-projects/coronavirus-preparing-for-challenges-this-winter
7/10

The model will be expanded to include the B.1.1.7 lineage, and incorporate the effects of vaccination now rollout has begun. The model, data and fits will continue to adapt as the situation changes, so we can keep answering 'what if' and 'how many' questions each week
8/10

It's been a privilege to help out with some of the software used by this model. But this is all due to the constant work of Ed, @lilithwhittles, @pabloperguz, @rgfitzjohn, @dr_anne_cori and @MarcBaguelin, whose dedication and resilience has been inspiring
9/10

Model code is available (and has been from the start) at: https://github.com/mrc-ide/sircovid
data and scripts to run: https://github.com/mrc-ide/sarscov2-transmission-england
More info on software: https://wellcomeopenresearch.org/articles/5-288/v1
10/10