I read in yesterday's Times about a Canadian study which claims to show that face masks can almost halve the number of new Covid-19 cases, but with the caveat that it has not yet been peer-reviewed. So, being more than mildly cynical about the efficiency of wearing non-medical face coverings to mitigate against the transmission of microscopic organisms, I've been and found it....
For the study, the authors looked at the difference in case growth rate between Ontario’s 34 Public Health Units (PHUs) and between Canada’s ten provinces during the period when some had mask mandates in force and some did not. They allowed two weeks for mandates to have an impact on reported cases, and did additional primary research to show that the mandates did in fact coincide with a change in reported behaviour (i.e. more people reported wearing masks).
 |
| Add caption |
The first point to be made is that the mask mandates came in during the general decline of the epidemic and the period of low prevalence in July and August when there were only around 400 cases per day nationally in Canada. This makes the data highly sensitive to the testing regime and susceptible to false positives. It is noteworthy that there was a substantial increase in testing in July, showing up as a bump in the otherwise declining cases curve. At the time Ontario was recording just 100 cases or so per day. Divided between the 34 PHUs, that’s just three each per day. At a time of very low prevalence and high levels of testing such data cannot be considered reliable.
From a big picture point of view, it is of significance that once masks were mandated everywhere in Ontario (from July 8th) and Quebec (from July 18th) by the end of August cases started climbing again and have continued to do so as autumn has drawn in. Additionally, if you look at the cases curve for each province the impact of increased testing is clear but there is no sign of any impact of masks (or other intervention), just a smooth curve. Neither of these points supports the hypothesis that masks make a difference.
The main graphs the authors present do show an average difference in case growth rate between mask mandate and non-mandate PHUs and provinces for the four weeks or so in July and early August when mask mandates were not universal. In line with the authors’ hypothesis, mask mandate PHUs and provinces had a lower overall growth rate, though this was not consistently the case. It is noteworthy that the time lag from the mandate dates to the divergence in case growth of mandate and non-mandate regions is very different on PHU level and province level – two weeks versus four weeks – again counting against a claim to causative effect.
The most obvious explanation for the divergence between the two categories (mandate and non-mandate) is that the different PHUs or provinces are on different trajectories and so when they move from one category to the other (non-mandate to mandate) this has a corresponding effect on the growth rates. There is strong evidence this is the case.
On the PHUs graph we see initially in the second week of July the few PHUs with mask mandates in place have a higher average growth rate than those without one – the opposite of the hypothesis. Then, as some of those without mandates impose mandates and switch categories in the following week, the mandate PHU average growth rate quickly declines while those remaining in the non-mandate category quickly begin to show a higher rate of growth.
Finally, in the second week of August (when the study period comes to an end), most of the remaining PHUs impose mandates and switch categories, and the growth rate in the mandate PHUs spikes rapidly. The fact that this shift results in a spike rather than continued decline is indicative that any difference was an artefact of which PHUs happened to be in which category rather than the impact of masks. In other words, they brought their higher growth rate with them into the mandate category rather than seeing it curbed by their masks. In this regard it is significant that the remaining few PHUs with no mask mandate in the second week of August actually trend below those with mandates – again, the opposite of the hypothesis.
In the case of provinces, in addition to the problem mentioned above that the average difference takes four weeks to show up rather than two, the data is skewed by the fact that only two of the ten provinces adopt a mandate during the study period so the sample size is very small and, again, the effect could be merely a function of the different trajectories of the different provinces. As noted, there is no visible sign in the case curves of the mask mandates altering the trajectories, and most provinces have seen a large growth in cases during the autumn.
The paper includes an elaborate attempt to account for the impact of other government interventions, but as this is all modelling, based on a lot of estimates and guesses, it does not address the criticisms raised here.
The evidence that the general use of cloth masks by the general population is ineffective for preventing the spread of Covid-19 (and other similar viruses) is well established. This is why the World Health Organisation did not endorse them until June, and even then stated: ‘At the present time, the widespread use of masks by healthy people in the community setting is not yet supported by high quality or direct scientific evidence.’ . It listed 11 ‘likely disadvantages of the use of mask by healthy people in the general public’, including ‘potential increased risk of self-contamination’. It all but admitted to BBC medical correspondent Deborah Cohen on July 14th that the change of policy was due to political lobbying rather than evidence. Cohen said:
‘We had been told by various sources WHO committee reviewing the evidence had not backed masks but they recommended them due to political lobbying. This point was put to WHO who did not deny.’
The study from Canada is the latest effort to produce evidence for something that has long been established to be false. Viruses are too small to be caught by pieces of cloth and the general public will never use masks properly or keep them clean. Airborne infection is largely a function of the viral load or concentration that has built up in a non-ventilated space, not the projection of droplets from cough and sneezes, and cloth masks are useless in preventing such a build-up of virus particles or protecting from the inhalation of them. They are also bad for health in multiple ways because they obstruct breathing clean air, are bad for social interaction, and particularly bad for social psychology.
I've been told - repeatedly - that resistance to wearing masks (especially as they don't need to be medical grade and could in theory be fashioned out of old string vests) indicates a susceptibility to conspiracy theories, possible David Icke related madness, selfishness or even a sign of low intelligence. However, when the evidence in favour of their use is either flimsy (no pun intended), or non-existent, they are not much more than the 2020 version of wearing tin foil hats to prevent abduction by aliens.
