The UK Covid-19 lockdown weakened in April and May 2020: implications for the size of the epidemic and for outcomes had lockdown been earlier

The number of active cases in the UK Covid-19 epidemic, the case fatality rate, the susceptible proportion of the population, and how well the lockdown was maintained during April–May 2020 are unknown. These four have a relationship with the shape of the daily mortality curve once one considers the intervals from infection to death or recovery. Without an understanding of this relationship we cannot say that an earlier lockdown would have saved lives. Using a small stochastic model, the lockdown had to be weakened, in April and May, for simulated deaths to match ongoing actual daily deaths. Google mobility data was found to be consistent with the weakening required in the model with similar changes from baseline in time and magnitude. If in an earlier lockdown, mobility and interactions would have followed a similar course, then with a large epidemic curve an earlier lockdown might be associated with many more deaths than some currently believe. This was confirmed in the stochastic model and in two modified SIR models of epidemics of various sizes. The first SIR model had a fixed period to recovery and the second used random periods, both models had random periods to death. Weakening of the mitigations was required to tune the output in large but not in small epidemics. This gives weight to the epidemic having affected many more individuals than some reports currently suggest. In both one and two-week earlier lockdowns, total deaths were found to depend on the size of the epidemic and to vary from 2,000–49,000 deaths. There was a linear relationship between the peak proportion of the population infected and the reciprocal of the case fatality rate. This work questions the low prevalence of < 0.1%, reported by the Office for National Statistics in May and June 2020, since to accommodate a weakening lockdown, the shape of the daily mortality curve, and an acceptable case fatality rate a much larger epidemic curve is required.