Coronavirus modelers in wonderland
The Italian press and academia celebrates the latest paper in the realm of disease modelling by Andrea Rinaldo and colleagues. The professor proudly states in the press release:
The increase in confirmed infections and their much lower average age are of great concerns for us. As the cold season approaches, indoor activities will make airborne infections much more aggressive, unless there is a strong discipline in personal protection. In my opinion, the mask should remain mandatory until a vaccine is widely available.
We already see here the two big theses of the Coronazis:
- Human behavior must be restricted to whatever degree is necessary to reverse the natural tendency of respiratory viruses to spread especially during the winter. Remeber the winter cause it will come back.
- The harsher and harsher restrictions must be in place until a vaccine is widely available.
There is no indication that either 1. or 2. are actually possible: the vaccine will not be available at least until spring 2021, and may in fact never be available. The hypothesis that even the harshest lockdowns can stop the spread of a respiratory disease is extremely weak, but I don’t even want to go into that: what do you do once you’re done locking everybody in for months and you had success isolating your country à la New Zealand? People paid with their money and lives an “achievement” that is in fact a terrible disadvantage, for the country must now remain isolated and locked down forever.
But enough for the general picture, let’s talk about the SCIENCE. Lockdowners love SCIENCE.
The geography of COVID-19 spread in Italy and implications for the relaxation of confinement measures
Published no less than by Nature Communications, the paper predicts the number of infected people in Italy according to different “relaxation” scenarios, starting from May 1st, 2020. Let’s go through the paper together.
A 40% increase in effective transmission would yield a rebound of infections. A control effort capable of isolating daily ~5.5% of the exposed and highly infectious individuals proves necessary to maintain the epidemic curve onto the decreasing baseline trajectory.
Interesting conclusion: a relatively small increase in transmission (40%) is enough for a nightmare scenario of collapsing hospitals, yet to prevent it it’s enough to isolate 5% (5.5% because the model is that precise) of the highly infectious individuals. This already smells: so 95% of the infected people could go around freely and the model predicts it’s gonna be Gucci, with the disease still driven to extinction by the fact that the other 5% is isolated. To top the absurdity, the authors use this calculation to recommend that not only all exposed individuals are isolated, but that everybody else is restricted.
We estimate a large reduction in the effective disease transmission rate in each province. This reduction, expressed as a ratio of effective transmission estimated on May 1 to the initial uncontrolled one, ranges between 0.3 and 0.4 depending on location (Fig. 1d). Technically, this reduction is computed via the product of the reduction in transmission rates times the fraction of the population still susceptible to the infection on May 1. The latter, however, is very sensitive to the fraction of infections that develop heavy symptoms (parameter σ in the model, see “Methods” section). The reference value assumed is σ = 25%, which is consistent with empirical evidence. […] The highest value (σ = 50%) matches the empirical results found by testing for two weeks an entire community (Vo’ Euganeo, IT, ~3000 inhabitants)12, whereas the lowest (σ = 10%) is likely to be a lower bound, unachievable in the actual geographic context, because it may reflect also the age structure of a much younger population38. Another way to assess the plausibility of the assumed fraction of infections that develop heavy symptoms, is to compare the IFR estimated with different values of σ. The values of IFR corresponding to σ = 25% and σ = 10% (2% and 0.8%, respectively) bracket the available estimates of IFR for western countries39. Thus the median value assumed (σ = 25%) seems like a sensible choice to probe the actual role of the unobserved epidemics in Italy.
The authors don’t define what counts as “heavy” symptoms for them. They take 25% and consider up to 50% which is absolutely inconceivable. The whole reasoning behind compulsory masks for everybody everywhere is that the large majority of people who are contagious do not show any symptom. Official statistics for Italy show that heavy symptoms where detected in 25% of people with a confirmed test only above 60 years old. Undiagnosed cases vastly dwarf confirmed cases, and most of the population is below 60 anyway. But the authors continue:
Different assumptions for σ result in different values of the infection fatality rate (IFR), defined as the ratio between the official death count (at a certain date) and the corresponding total number of infections estimated by the model. As of May 1, we estimate an IFR of 4%, 2%, and 0.8%, respectively for σ = 50%, 25%, and 10%.
It is absolutely certain that the IFR is below all of these three numbers. The CDC gives in their meta analysis a most likely scenario of 0.65%. 4% and 2% are plain ridiculous at this point in time. In fact, in the summer of crazy testing 2020, the confirmed fatality ratio was as low as 0.15% in Spain. So while 0.8% could be borderline (im)plausible, also matching a more plausible 10% for heavy symptoms, the other two are just bonkers.
But enough chat, let’s check the money plot.
Shown here are reconstructed data (empty circles), and model results (solid lines and confidence intervals). The blue solid line represents the baseline scenario, i.e., the median of the computed results with transmission estimated during lockdown maintained indefinitely beyond May 3, 2020. The green and purple solid lines represent the scenarios corresponding to a release of the containment measures determining an effective increase in the overall transmission rates of, respectively, 20% and 40%.
With indefinite lockdown the blue line is predicted. With 20% more transmission between people after the end of the lockdown the estimate is at least 1000 new hospitalizations per day, with an increasing trend. Well, how did that go? The numbers are in the low tenths since at least June. So how do the authors comment the fact that their model is off by a factor of 10 or even 100 and is therefore complete garbage? No biggie: the fact that it failed spectacularly (exactly like others: Colorado, Imperial College, etc.) won’t discourage us from destroying your life.
They show this plot:
Gray empty circles show the newly available data up to June 17th.
And they are not ashamed to comment
Among the three scenarios explored, the actual progression of the outbreak in the Italian territory after the lifting of lockdown measures is consistent with the baseline scenario
- It is not consistent, it is much lower, and astronomically far away from your “95%” confidence band.
- The blue line is the prediction for the continuation of house arrest for the whole population, who could only get out to work in essential businesses and buy groceries.
To pretend that this prediction isn’t absolutely ridiculous, the authors say that people probably meet more than when they couldn’t get out of their houses (at least they don’t dare deny that…), but hum yeah it must be that they wear masks all the time and follow distancing rules. Which, if you have been to any corner of Italy or even elsewhere in the world, you must know for a fact is not true.
Several intertwined factors could explain these patterns of disease progression after lifting the lockdown. In the following we discuss what we deem most relevant. Restrictive measures have been only partially released: education, from pre-school to higher education, will resume in-presence activities only in September; large gatherings are still forbidden; every commercial activity and workplace is still subject to stringent protocols to ensure social distancing and avoid infections; PPE are mandatory in every indoor setting but one’s own household and recommended outdoors whenever preventing social contact is not possible. Altogether, these measures may have contributed significantly to avoid a recrudescence of virus transmission. In particular, evidence accumulates for the crucial role of face masks in reducing transmission
So our prediction is complete garbage, but wait it’s not because of that, it must be… MASKS. Hence mandatory masks forever.
The definition of circular reasoning:
The effect of the containment measures was parameterized by assuming that the transmission parameters (βP, βI and βA) had a sharp decrease after the containment measures announced on February 24 and March 8.
So nothing can ever be wrong in this “model”: we assume that imposing a lockdown will sharply decrease the number of infections. Therefore, we can prove that lockdown is necessary to sharply decrease the number of infections. Even better, after the lockdown ceases to exist infections still went down. What was introduced then? Masks. So then it must be masks, with a massively stronger effect of even lockdown itself. Now cases are going up, what is it? Of course people not using masks enough! Lock them up! or down! Whatever as long as everything is closed and my bank account receives the full professor salary all the time!
Maybe if you want to see if the parameters have had a sharp decrease due to lockdown you can’t just assume it in the first place? Maybe none of this makes any sense?
Isn’t our model missing something? Something that could be important for coronavirus? There are other coronaviruses around, and they cause common cold. Which is named cold because you have it most often when it’s COLD.
There are even studies on that. Wanna read them? They could be SCIENCE too.
Seasonality effects5, not explicitly accounted for here, might also have had a role in the reduction of transmission. […] Some authors have suggested that COVID-19 transmission may be subject to seasonal variations in analogy with other human coronaviruses (see, e.g., ref. 5). However, this hypothesis is still debated among experts, and a clear evidence has not emerged yet. Should warmer, drier weather be proven to actually hinder transmission, such factor could have played a role in Italy during the timeframe analyzed herein.
So they might (thank goodness) also have had a role and a clear evidence has not emerged (the fact that the epidemic arrives and leaves exactly at the same time in places with the same climate but different lockdown strategies, including NO lockdown is not evidence, as well as the fact that all coronaviruses are highly seasonal), but we didn’t consider them. In fact, in their equations the only things that can change infection rates are immunity and people meeting less. Therefore whenever the observed number of infection decreases, their model attributes it to lockdowns, then when the lockdown is not there (and it’s summer) to indoor masks and tracing which are supposedly more powerful than full hard lockdowns, an absurdity of academic level (a PhD is required to believe it). Let’s go back to the interview at the beginning:
As the cold season approaches, indoor activities will make airborne infections much more aggressive
Aha. So like, much less aggressive in the summer…
What is the consequence of this disastrous piece of toilet paper? Winter will come because at least they haven’t published any peer-reviewed paper stating that wearing masks can prevent the Earth from going around the Sun, and coronavirus spread will undoubtedly resume everywhere, since the virus is present all over the world. We cannot prevent this from happening, and all the policies, informed by such disgusting “science”, attributing the change to human behavior, not enough masks, and similar dogshit arguments are bound to destroy our civilization — and to a larger degree than what we have already seen in 2020.
The cold-flu season hasn’t even started yet, coronavirus hospitalizations and deaths are at record lows, cases are probably driven by the absurd increases in the number of tests and masks are mandatory everywhere indoors in most of Europe, outdoors in Italy, Spain, France. Borders are closing rapidly (I don’t even know what that is supposed to achieve by the way). If you believe that people aren’t doing “enough” and that this is the reason of the increase what are the next steps? Lockdowns from October to May? That’s going to be so good for the health of everyone, right?
Shame these incompetent so-called scientists and stop this madness.