UA minimal relaxation of these measures would lead to disaster, increasing the number of daily cases to more than 4,000, and quadrupling the number of new hospitalizations, without significantly affecting, however, the number of deaths.
As these projections are used by the government to justify maintaining strict measures that undermine the freedoms of citizens, it is essential to assess the quality of these and to put in context the significance of these figures.
Theoretical physicist, researcher in biophysics, I have been working for several decades on issues of science and society. I published last November an essay, Pandemic, When Reason Falls Ill, in which I offer a critical reflection on the profound issues that the health crisis has revealed on the functioning and communication of science, the heaviness of the state apparatus, the role of the media, the behavior of elected officials and the place of the most vulnerable in political orientations.
Epidemiologists get carried away
Indeed, since the start of the pandemic, the comparison between the INSPQ projections and the reality have shown that these have systematically, with the exception of the December projections, significantly overestimated the number of cases. hospitalizations and deaths, even for the so-called “optimistic” scenarios. As for the projections of December 17, 2020, they completely missed an upward trend that began at the end of November.
A month later, the INSPQ models published on January 21 were equally incapable of predicting the drop in the number of cases, which began in early January. We observe a similar bias in the federal government’s projections, where, almost systematically, optimistic projections are higher than reality: despite the upward projections, in Quebec, the number of cases in the field continues to decline, despite the predictions. alarmists linked to the spring break and the increase in the proportion of variants. They are now reaching very low levels, comparable to September 2020, which do not justify maintaining current health constraints.
The poor assessment of the evolution of COVID is not limited to Quebec and Canada. While everywhere, epidemiologists had predicted a rise in cases after the major holidays, be it American Thanksgiving or Christmas and the end of the year, no sign of runaway is visible when looking at the evolution of cases retrospectively. On the contrary, after peaking in late December or early January, the United States, Great Britain and Sweden saw the number of cases drop rapidly, following a curve very similar to that of Quebec and Ontario. .
However, if the Canadian experts immediately attributed this decrease to the reinforced sanitary measures, the experts abroad showed much more humility, recognizing that they were unable to explain this sharp decrease, whereas they had rather expected a rise in the number of cases after the holidays.
By digging a little deeper into these figures, several observations are in order:
-
While it is impossible to explain the sharp decrease in cases in the United States and England, even taking vaccination into account, it is difficult to see how we can, in Quebec, establish significant causal links between the measures. imposed since January 8 (including the curfew) and the evolution of cases since the beginning of the year.
-
The absence of a spike in cases following Thanksgiving in the United States and the holiday season suggests that meeting privately with family or friends is much less risky than it is claimed, as people have adopted preventive behaviors. These data call into question the relevance of the ban on such gatherings imposed on the majority of Quebecers since October 1.
-
Although the variants are worrying, the marked drop in the number of cases in the United Kingdom and the United States, where they circulate much more than in Quebec and Canada, suggests that their real impact is relatively weak in a context where distancing measures are already in place. Indeed, this fall in cases far exceeds what one would have expected with the current vaccination, as shown by the case of Israel, where the spread remains marked, despite a high level of protection, which makes it possible to present a very low death rate.
Despite these data, for methodological and training reasons, the Quebec and Canadian public health authorities, supported by the modellers, seem to favor a fragmented approach to risk analysis: each measurement, each problem is analyzed independently, without with regard to the transformations in our daily lives over the past year or to the interactions between them.
–
