Prevalence of unvaccinated and vaccinated: weaknesses in the calculation of the LGL policy

– Once a week, the Bavarian State Office for Health and Food Safety (LGL) publishes the seven-day prevalence of unvaccinated and vaccinated people. The calculation has weaknesses. A similar case at the Robert Koch Institute and a look at other federal states show that calculations can theoretically be done differently.

In mid-August, LGL published for the first time the seven-day incidence of unvaccinated and fully vaccinated persons in Bavaria. Even then, critics objected that these values ​​could not be compared one to one. One argument: people with full vaccination protection would probably be tested much less often than people without immunization.

How many and how often vaccinated people get tested compared to unvaccinated people is unclear and highly individual dependent. But the published occurrences of LGL have a completely different, concrete weakness in the calculation. There is apparently an under-registration of corona cases in vaccinated people.

Calculation of occurrences

How does this underreporting occur? The calculation of the incidences is based on reporting cases with a reporting date in the last seven days, which are divided by vaccination status. These are compared with the population figures for vaccinated and unvaccinated from the Robert Koch Institute’s (RKI) digital vaccination surveillance.

Vaccinated Covid-19 cases are counted as persons who had full vaccination protection at the time of infection. However, it is not only people who had not received a vaccination at the time of infection who are counted as unvaccinated Covid-19 cases. This also includes people with corona infection for whom there is no information on vaccination status from the health authorities. Persons with incomplete vaccination protection are not included in the calculation.

How does an unknown vaccination status occur?

LGL gives our editors an example. Admission of a person to a hospital in relation to Covid-19 must be reported. This notification is important in order to determine the occurrence of hospitalizations to which additional protective measures are attached. In addition, the form for this report also asks about the vaccination status of the admitted person. According to LGL, the information in reporting practices is often not completely filled out.

Underreporting of corona cases in vaccinated people

The fact that persons with unknown vaccination status are initially included in the group of unvaccinated – information, if available, is added with a certain delay – may result in a higher incidence among the unvaccinated and a lower incidence among the vaccinated.

Now the decisive question is how high the proportion of Covid-19 cases is with unknown vaccination status. This is the only way to actually determine how badly the two ratios are skewed. However, LGL cannot provide our editors with any more detailed insight into how high this underreporting of the vaccinated may be.

When asked about this weak point in the calculation, LGL replied that after a thorough analysis of the available registration data and considerations, it was decided to include the group with unknown vaccination status among the unvaccinated. This is based on “that otherwise a not inconsiderable number of reporting cases would initially be excluded due to missing information in the evaluation”. LGL is always working to improve the data situation.

We asked the authority’s press office to clarify the information and specifically wanted to know how high the proportion of corona cases with unknown vaccination status is – also, for example, for the incidences published last Thursday, November 11. The information from LGL: More precise information is not possible, due to numerous inquiries LGL cannot offer individual evaluations.

Similar case at RKI

The RKI’s calculation of vaccination effectiveness had a similar weakness, but they have now responded: firstly, when calculating vaccination effectiveness, the RKI also counted people with unknown vaccination status among the unvaccinated. Here, too, it had to be assumed that the number of corona cases in vaccinated individuals was underestimated and the vaccination effect therefore overestimated. However, the authority adjusted its calculation at the end of September and now only takes into account those Covid 19 cases for which there is complete indication of vaccination status.

Look at other states

There are no nationwide uniform calculation rules for the two indicators, and by no means all federal states record the two occurrences. One reason why data is not collected in North Rhine-Westphalia, for example, is corona cases with missing or incomplete information on vaccination status. A spokesman for Lower Saxony justified the rejection by saying that the calculation was “quite complex and time-consuming”.

In addition to Bavaria, Thuringia and Saxony, among others, also register separate occurrences. In Thuringia, corona cases are only taken into account if the information on vaccination status is complete. In Saxony, an incidence is calculated for “fully vaccinated persons” and one for “not (fully) vaccinated persons”. The “not (fully) vaccinated” also include all Covid 19 cases for which there is no information on vaccination status because “not collected” or “cannot be determined”.

In Bavaria, LGL justifies its calculation despite the difficulties: “Nevertheless, the data collected in this way allow for the analysis of general statements and trends about the relationship between the affected and unvaccinated population.”

Leave a Comment