A number of tales about numbers

Ecclesiastes 3:7

In Germany, the Robert Koch-Institut (RKI) has been monitoring all things Covid and has been weekly (or should I say weakly?) publishing reports. Among other things, ICU cases are being counted, and vaccine efficacy is being computed. In the report as of January 20th, 2022 the RKI lists (cf p. 27), for the four-week period from calendar week 51/2021 to calendar week 2/2022, 618 unvaccinated cases, 158 double vaccinated, and 39 boostered. Partially vaccinated cases, ie, cases in the time period between the first jab and 14 days after the second, or less than seven days after the third, are disregarded.

Since December 14, 2021 there is a second data source regarding Covid ICU cases. The DIVI Intensivregister is now counting as well, and these numbers have been included in the RKI report (cf p. 18). Here it is 3,156 unvaccinated cases, 562 partially vaccinated cases, 1,115 double vaccinated, and 442 boostered (it remains unclear how patients who have received the booster jab less than seven days ago are being counted).

I would love to study these data using statistical methods, but it doesn’t make any sense. Instead, I will tell some tales. Who doesn’t love a thrilling tale? Note that all patients on ICUs everywhere have my sympathy, and I wish them the very best and full recovery, but there are things going on that need to be addressed.

A tale of vaccine efficacy: both sets of figures exhibit proportions of unvaccinated patients that are much larger than the proportion of unvaccinated people in the population. Therefore, the vaccines protect well against severe outcome in case of a Covid infection. And it’s even better with booster!

A tale of impotence: the RKI is unable to make hospitals transmit data in a way that is useful for calculation of vaccine efficacy. They have been lucky that proportions in the sub-sample available to them are comparable to those recorded by the DIVI.

A tale of symptoms: one main difference between the DIVI and RKI figures is the RKI’s restriction to symptomatic cases. Are five out of six ICU “cases” asymptomatic? And what would this imply for the whole containment circus in which we are all whipped across the arena?

A tale of noise: in a typical 4-week period during winter, more than 120,000 patients will be squeezed through German ICUs (20,000 capacity, 90% utilization, 4-day stay on average). If 1 in 100 people turn out positive when tested for Covid (eg, 7-day incidence of 1,000 per 100,000 considering under-reporting, 7-day window for positive test), a baseline of around 1,200 infections will distort the statistics.

A tale of incentives: it really seems to be profitable for hospitals to report large numbers of Covid patients, and even more profitable to report large numbers of Covid ICU patients. It might even make sense to set up separate Covid (ICU) stations and transfer all patients tested positive to these. A tale of noise, reloaded: in a 4-week period, more than one million people will be registered in German hospitals. The baseline of positive tests rises to 10,000, more than the total DIVI figure. How many of these will end up in ICU?

A tale of fraud: no, I won’t tell that one.

A tale of bad vaccination strategy: we are vaccinating all those healthy, younger people that would never end up in ICU if they got Covid. At the same time, we are too stupid to find and convince those that would really profit from vaccination. That’s why we end up with both large numbers of unvaccinated people in ICUs and splendid vaccine efficacy figures.