How many people become infected with the corona virus? And how many become seriously ill? How many people have already contracted Covid-19 and are immune for some time? These questions are important to decide when and under what conditions we resume our social lives. Because as long as neither medicines nor vaccines are available, the strain on health care systems must be prevented.
However, exact figures on how many people have been infected with the SARS-CoV-2 virus are still missing. People are currently tested only under certain conditions. Furthermore, 80 percent of virus carriers develop only mild symptoms or no symptoms at all. These infections often do not appear in the statistics. Accordingly, the number of unreported cases of actually infected persons is higher than the cases reported because they were tested (*).
Lethality and immunity
This is where the thinking of Allen Caldwell comes in. If we know the basic death rate for Covid-19, we can draw conclusions about the total number of people infected. This would also provide insight into what proportion of the population is immune to Covid-19 – at least for a certain period of time.
Although many studies have dealt with the question of mortality, the results are sometimes mixed and typically only report on one study at a time. “I am following developments in the Corona crisis with great interest”, explains Caldwell. “As a particle physicist, I have extensive experience with statistical methods and evaluations.
It therefore made sense to take a closer look at the conflicting death toll figures”. Caldwell and his team have accumulated the available data on antibody tests for the virus to see if they could better quantity the true mortality rate.
Evaluation of 7 international studies
For his analyses, Caldwell and his co-workers Vasyl Hafych, Oliver Schulz und Lolian Shtembari used reports from 7 randomized tests for antibodies to the Covid-19 virus (**). These studies examined the proportion of people in certain groups already infected – regardless of whether they had symptoms.
They then related the infected fraction to the number of deaths in connection with COVID-19 and found that the calculated mortality ranged from below 0.2% to close to 0.6%. Given the uncertainties in the reported numbers, they quote an even wider interval, ranging from 0.1% to close to 1%. The actual value for a given region depends on specifics related to the tests and to the population.
This range is nevertheless considerably lower than when mortality is calculated on the basis of case numbers (i.e. the number of Covid-19 cases actually reported) (**). In Germany, this figure is currently around 5 percent (as of 15 May 2020).
Further analyses intended
“However, the data we used comes from countries with good medical care”, admits Caldwell. “The death rates may be higher in less developed regions”. In addition, the antibody tests are not 100 percent specific to the SARS CoV-2 virus. Depending on the study design, there may also be statistical biases.
Caldwell would therefore like to continue the evaluations. “Every day, there are new studies and figures. It also appears that extremely precise and broadly applicable antibody tests will soon be available. These will provide us with a much larger database – and statistically more reliable results”.
(*) In the case of an infection, the genetic material of the virus can be directly detected for a short time. These PCR tests provide figures for the case statistics of the Robert Koch Institute, for example. During the infection, those affected produce antibodies that are still detectable after several months. These show that a person has overcome an infection.
(**) A distinction is made between two methods for calculating deaths.
The case fatality rate (CFR) is calculated from the number of cases reported after positive PCR tests and the death rate. A variation of the case fatality rate is the infection fatality rate (IFR). Here, the number of known infections is compared with the number of deaths.