Which professions are over-represented in Sweden’s coronavirus statistics?

The preliminary report assessed the risk of infection of people in different occupational groups, by measuring how many people from different groups had been diagnosed with coronavirus, compared to the size of that group in the general population.

Of the groups included in the study, taxi drivers were most over-represented in the statistics, with 83 confirmed cases and 12,475 people in the profession in Sweden according to Statistics Sweden.

They were followed by pizza makers (among whom there were 34 confirmed cases in Sweden, a relatively high proportion of the 5,521 people in this job category), bus and tram drivers (140 cases out of 23,319 people in this job), interpreters (16 out of 4,033), restaurant kitchen staff (23 out of 6,578), other service workers (the examples given by the agency for this category were staff responsible for filling vending machines or emptying parking metres, reading gas and water metres and checking TV licences) and firefighters (17 out of 5,658). 

All of the above professions were associated with at least double the rates of coronavirus compared to the population in general. 

• Coronavirus: Is Swedish trust in authorities starting to falter?
   
• People born overseas at higher risk from coronavirus in Sweden

Medical and care workers were excluded from the report, which the agency said was because this group has been tested to a greater extent from much earlier on in the pandemic, so it would be difficult to assess to what extent they were affected in comparison to other jobs. However, given that these jobs involve extended close contact with sick people, it's almost certain that these people are infected at a higher rate than the general population.

The professions which emerged as having a higher rate of infection than the general public, but not as much as double, were cleaners, caretakers and waiters.

The report also looked at the relative risk for staff working in schools and found that on the whole, teachers and teaching assistants were not over-represented in Sweden's confirmed coronavirus cases when compared to other occupational groups. 

For preschool teachers and upper secondary school teachers, the calculated relative risk was lower than average, while for assistants and secondary school teachers it was slightly higher. In the time period studied, 29 cases were confirmed among Sweden's 30,357 upper secondary school teachers, 68 cases among 68,605 preschool teachers, and 160 cases out of 105,418 secondary school teachers.

Where do these statistics come from and how reliable are they?

Included in the data are the confirmed cases of Covid-19 in Sweden which were diagnosed between March 13th and May 27th. During that period, Sweden's testing strategy has changed, with predominantly people requiring hospital care or belonging to a risk group tested in the first weeks, and testing expanding to more groups in society more recently. There are also regional variations in who can be offered a coronavirus test.

So it's important to note that there will have been many more cases than those which show up in the diagnoses. That means that when we say 29 cases were confirmed among 30,357 upper secondary school teachers, it doesn't mean that 30,328 did not catch coronavirus — just that they were not diagnosed with it.

The data on coronavirus diagnoses, limited to those aged between 25 and 65, was combined with Statistics Sweden's data on job classifications from 2018 in order to work out which jobs were most represented in the statistics in comparison to the number doing that job in the total population. For 25 percent of coronavirus patients, there was no information on their job, which adds further uncertainty to the final figures.

Only jobs where at least 15 people were confirmed to have caught coronavirus were included in the statistics, which meant this data is based on 3,188 people. 

Some of the factors which are known to affect the risk of catching or getting seriously ill from Covid-19 were not analysed, including health condition and lifestyle factors.

In other words, the limited data available, variations in testing, and factors not controlled for mean that the statistics can't be considered totally reliable, especially in terms of accurately calculating the differences in risk level for different jobs.

But they do give a general picture of the kind of professions that are most associated with an increased risk of catching the coronavirus, which can be useful for identifying patterns and informing strategies to prevent spread of infection. This was a preliminary analysis, and the agency said that more detailed follow-up analyses will come later.

You can read the full report in Swedish here.

• Calendar: These are the key dates to be aware of for changes in Sweden's coronavirus rules