How can you assess the risk exactly

med-wiss blog

The current AHA formula includes that you should keep at least 1.50 m away from other people in order to reduce the risk of infection with corona. Other countries even speak of a distance of 2 m. But how well founded are these recommendations given the new evidence that Aerosols play an important role in the transmission?

A current publication from Oxford, UK, comes to a much more nuanced view [1]. The authors state that the distance rules are based on outdated scientific knowledge. Coronaviruses can appear in droplets of liquid significantly further than 2 m fly through space when people cough, sing or shout loudly. The distance is then less important for the risk of infection.

When people meet, what matters is how many people are together, how long they have been in contact, how good the ventilation is and whether everyone present is wearing a mask or not. Based on these points of view, Nicholas Jones and colleagues developed an easily understandable traffic light scheme:

Jones et al. [1], Figure 3 translated into German

Contagion risk: green = low, yellow = moderate, red = high.

The scheme applies in the event that none of the people present has any symptoms. A high-risk situation would be staying in a club or bar, for example, where many people stand close together for hours without a mask, speak loudly and the room is very poorly ventilated. Situations where few people meet outside bring only a low risk.

These considerations neglect one important factor, namely how high the probability is that one of the people present is infected with the coronavirus. That's what the current one is for Infection events on site crucial.

A working group from San Francisco took up exactly this aspect [2]. The rationale is that that Risk from the person you meet, is just as important as the risk in the situation one is in. If a region is practically virus-free, you cannot get infected even in a "high-risk" environment. The authors have thought the whole thing through to the point that they can find an English language online Tool for calculating the risk provide. In an additional document, they describe in detail their considerations and the research results on which each individual calculation step is based. Incidentally, I owe the reference to this website to the statistician Katharina Schüller, who runs the blog "Unstatistics of the month" together with Gerd Gigerenzer, Walter Krämer and Thomas K. Bauer.

In order to use the calculation tool from San Francisco that Personal risk To calculate, you have to know how many people in the respective region have been newly diagnosed as corona infected in the last 7 days. This number of weeks of new infections is based on 1 million inhabitants. In Germany you can call up the 7-day incidence per 100,000 inhabitants for each of the 412 districts at the Robert Koch Institute. Multiplying this number by 10 gives the 7-day incidence per 1 million population.

Because infected people can be contagious or even before the onset of their disease no symptoms at all can develop, the authors also take into account the Dark figure. For our conditions in Germany, they are given with a factor of 6. The personal risk per 1 million population results from the 7-day incidence and the number of unreported cases. The authors call this metric MicroCovid.

In the following simplified table I have the corresponding calculations for the for two federal states and for two cities with low and high infection rates Personal risk understood:

It can be seen that the risk in Würzburg is 18 times higher than in Schwerin of meeting an infected person.

In the second step, the online calculator allows for certain Activities in everyday life to simulate the risk of infection. For example, if four people meet outside or inside for over two hours, they can do it differently: they can keep their distance and put on masks, or they can behave as if there was no corona epidemic.

For the Activity risk the online form provides these results:

The example shows that it is 16 times less risky to meet outside than inside, even if you do not comply with any corona rules at all. And inside you can keep your distance and wear a mask to reduce the risk by a factor of 5.

The third step is that Overall risk for this occasion calculated by taking into account the personal risk and the activity risk.

As Total risk of meeting 4 people The following results for the federal states of Mecklenburg-Western Pomerania and Bavaria, depending on the selected situation:

The rating as low or high risk is based on the authors' own one Infection risk of 1 percent over a whole year consider appropriate (for comparison: in the USA, more than 2% of all residents have been diagnosed with corona, although only 8 months of the pandemic have passed). One percent is equivalent to 10,000 MicroCovid. The authors consider this number to be their "budget", which they do not want to exceed within 12 months. Since every single activity carries a certain risk, all activities add up. For example, if you did not meet 4 people inside Bavaria, but instead 25 people inside “like before Covid”, there would be a total risk of 3,000 MicroCovid. With a single event one would have to be like this "Annual Covid Budget" Almost a third has already been used up. The rating would therefore be "dangerously high" in this case.

All of these numbers and calculations are of course based on Assumptions. How exactly these assumptions match reality is difficult to assess. The authors are also aware of this, and they themselves explain the limits of their method in detail. Since the code is exposed, other experts can help to further improve the calculation basis.

A alternative view it would be to say that a celebration with 25 people "like before Covid" is harmless even in Bavaria if the individual risk of infection is only 0.3%. Indeed, there are many risks in life, and some activities are dangerous. It is no coincidence that the authors based their MicroCovid unit on the unit of measurement Micromort, which indicates the probability of dying in a certain activity. A micromort is a one in a million risk of death. It is achieved, for example, when you drive 370 km or when you drink half a liter of wine. A parachute jump is 7 and a cardiac bypass operation is 16,000 micromorts.

In summary, the MikroCovid tool and the traffic light system allow one orientationto in everyday life more informed decisions for or against certain activities to meet with other people. What risk you want to take, what is important to you and what you feel comfortable with, everyone has to decide for themselves. Even when the overall risk of an activity is small, the many activities add up over time. And ultimately, definitely the behavior of millions of individual people the Infection process in the pandemic.


1 Jones NRJ, Qureshi ZU, Temple RJ, Larwood JPJ, Greenhalgh T, Bourouiba LB. Two meters or one: What is the evidence for physical distancing in covid-19? BMJ 2020; (370: m3223). On the Internet:

2 Ibasho community. microCOVID Project. San Francisco (September 13th, 2020). On the Internet: / paper / all; Status: 13.9.2020