The imposition of a third national UK lockdown today presents a unique opportunity for gauging the effectiveness of lockdowns in managing the SARS-CoV-2 pandemic. It is unique because both the previous lockdowns began after deaths had already peaked, meaning that cases had reached their maximum perhaps 3 weeks before that.
Although the curves of both deaths and cases appeared to be unaffected by the additional effect of lockdown in either case, it is possible that they had some minor effect, as modelling by Imperial College has claimed for several countries, but not for Sweden, where in the absence of lockdown, interventions considered trivial elsewhere were concluded to be responsible for reversing the infections. The possibility that natural resolution occurred in all cases was not considered in that study, but is an important possibility.
At the onset of the present lockdown, though, cases, hospitalisations and deaths are all following a more or less linear increase according the latest reliable ONS statistics (24th December), as these ONS screenshots indicate (Figs. 1-3):
This overall pattern is what one would expect for early January in a bad influenza year, as the EuroMOMO data on excess deaths for the last few years shows (Fig. 4), though one should note that, there having been a run of fairly benign winter infection years recently, this chart shows spontaneous decrease in deaths early in early January in most years – a potential confounding factor for our study, unfortunately.
Be that as it may, the present data, combined with the knowledge gained about COVID-19 over the last year, makes it obvious that, if lockdowns are an effective policy, the sudden universal imposition of this one will necessarily produce a noticeable downward “kink” in the case-numbers (subject to the rate of testing remaining constant) in around 6 days from today (the average period from infection to positive PCR test), then a corresponding slowing of hospitalizations perhaps a week after that, and finally a reduction in deaths, or at least a noticeable slowing of increase (“flattening the curve”), around 3 weeks from today.
The last effect, it should be said, may be more difficult to detect, as the increase in deaths is so low already, and not clearly related in degree or natural timing to the other two measures.
Nevertheless, the observation of these effects, or their absence, will give a good indication, in a short time from now, of whether lockdown has had the effects desired and predicted by its advocates. If the effect of lockdowns is negligible (the null hypothesis), as many studies across the world have indicated this year, no such clear change will be seen. These are testable predictions.
Should this lack of effect be observed, experience shows that critics may blame non-compliance by the public rather than the policy itself. This, however, would be misguided: lockdown, with all its associated legal penalties and the extensive use of behavioural psychology (which the WHO has lauded as one of the great gains from this pandemic for future policy), is a practical policy, not a theoretical one. By analogy, one would not criticise laws against fraud on the basis that people still commit it: good laws deal with society as it is, and not as it ought to be: “Politics is the art of the possible” (Bismarck).
Furthermore, such indications of the failure of lockdown overall would also demonstrate that the individual componensts within it, additional to measures already in place in each of the tiers (such as closure of schools, limitations on movement or association and so on) are equally ineffective. Within the next month, then, or even within the next week based on case numbers, it should be possible to determine conclusively the effectiveness, or otherwise, of lockdowns. This time-course is short enough to exclude the possible benefits of mass-vaccination as a confounding effect, which is not yet sufficient to affect the outcomes substantially, but may well confuse the issue if both lockdowns and vaccination are continued into the spring, when in any case seasonal excess deaths would be expected to fall dramatically.
If this lockdown proves not to change the course of cases, hospitalisations and deaths measurably, according to the known natural history of the disease under study, then no doubt SAGE and NERVTAG will have solid grounds for advising the government to abandon such measures immediately, and for the future. Lockdowns will have proved to be an unprecedented and hugely costly experiment that has, regrettably, failed. We are pleased to remember that our government’s mantra throughout this pandemic has been “Follow the Science,” so we may expect this laudable policy to continue in the light of our study.