Peak virus reproduction passed in Italy
“All conflict in the world is essentially about our differences in measurement.”
― Joseph Rain, The Unfinished Book About Who We Are
The new policies in Italy introduced to combat COVID-19, which divided the country in zones according to the urgency facing each region, imposed an effective lockdown in the red zones, similar to that of this Spring. The effect can be found in the virus reproduction number, which peaked on November 13 and now at a nationwide 1.08 according to the health authorities and corresponding to our estimated figure.
The daily routine of recording the national figures published every late afternoon has resulted in some important reflections on the numbers published and what they measure, as well as their nature as a sample. The first adjustment to new cases resulted from the swab test sample size variations. Having lived the entire procedure of having a positive case in the home and the quarantine procedures, it is evident that many of the swab tests reported each day are tests designated to monitor the already tested and quarantined population. The new cases tested are the ones we now use for adjusting the daily new cases. The number of swab tests effected to date are just under 22 million, whereas the total individuals tested stand at nearly 13 million. 1.6 million Italian residents have tested positive on tests. As infections rise, so do the ratio of swab tests to tested individuals.
The new corona-virus cases recorded vary with the actual incidence of cases in the population and the number of people sampled. Serological testing of entire municipalities have shown that the swab tests capture between 1/6 and 1/7 of actual cases. Hence, when we then run analysis on the samples, including the estimate of the virus reproduction number, the input number will determine your level of error. Since the virus reproduction is a non-linear function, we will get important differences in estimates of the virus reproduction number if our sample size is varying from one observation to the next. We easily get seduced into confusing the estimate and reality, preferring the concrete number estimated to the unknown actual number.
Following the revisions to the series, adjusted to variations in persons tested on a daily basis, it is notable how the timing of turning points change once you adjust the daily new cases for unique persons tested. Wave 2 started on July 5, almost as soon as the European summer vacations started in earnest. Adjusting for the incubation period, we can pinpoint the start to June 27. To me that sounds like the start of Northern European vacations and the flight to sunny beaches, the start of national school vacations of all levels, and so perhaps not entirely generated by domestic agents.
Wave 2 virus reproduction peaked on November 13 according to adjusted figures, while nominal published statistics indicate a November 1 peak. Looking ahead, the nominal figures suggest that the virus reproduction will break on December 13, while our forecast is substantially more cautious, suggesting that the virus reproduction number will fall below 1.0 on February 16, given where we are today.
It is a strong temptation to reopen the economy ahead of the holiday season if one is taking the nominal data vantage point, while holding back if the perspective is defined by test adjusted data.
When the emergency measures in Lombardy and other regions are relaxed from today, November 29 2020, the most significant change is the reopening of stores. In my previous post on the topic of COVID-19, I suggested that the measures that did not stop the virus reproduction should be reversed, including the closing of high street shops. The coming relaxation of measures are therefore likely acceptable risks, while we remain exposed to euphoria at Christmas as the idea that the virus is in decline is confused with the reality that the reproduction number is falling, but reducing at a slow pace. As long as it remains over 1.0, the new cases will rise and new deaths likewise. I would urge governments at all levels to resist opening for substantial social contact based on the peak in contagion, and analyses based on nominal figures, as these tend to overstate the positive picture.
In the current happy situation where we have observed a turning point in Italy in virus reproduction, it is evident that a Wave 3 is more than possible should governments at national and regional levels jump the gun and open for extended social contact in settings where the virus is most likely to find carriers. The ones who argue for opening up in order to save the economy would be better off avoiding a Wave 3, and the only way to constrain the virus before a vaccine is generally available, is through the simple hygiene measures and a substantial lockdown. Maintaining an Orange Zone regime is likely the most effective for now, until the virus reproduction is confirmed to be less than the 1.0 trigger level.