It’s wrong to say that only 3% of Covid-19 cases are caused by restaurants and pubs

“PUBS and restaurants caused less than three per cent of coronavirus infections in the week before the 10pm curfew, official data shows.”

The Sun, the Daily Star and the Telegraph have all claimed that pubs and restaurants caused just 3% of Covid-19 infections in the week before the new 10pm curfew was announced. 

This is a misreading of data published by Public Health England (PHE), which does not cover all Covid-19 cases. 

What the data shows

The 10pm curfew for businesses selling food or drink - including pubs, bars and restaurants - came into force on 24 September in England.

PHE’s weekly surveillance report shows that in the week to 20 September, 772 new acute respiratory infection (ARI) incidents were reported in England, of which 22 incidents were from “food outlet/restaurant settings.”

The process for reporting ARI incidents varies by PHE centre, but incidents are reported based on situations where:

• two or more laboratory confirmed cases linked to a particular setting, or 
• an outbreak is suspected,

are reported to a PHE health protection team (HPT). 

PHE says: “All suspected outbreaks are further investigated by the HPT in liaison with local partners and a significant proportion do not meet the criteria of a confirmed outbreak. For example if suspected cases test negative for COVID19 or other respiratory pathogens, or cases are subsequently found not to have direct links to the setting.”

An ARI incident can be for any respiratory pathogen, including Covid-19 and the flu. 

Of the 772 ARI incidents reported in England in the week to 20 September, 532 had at least one linked case that tested positive for Covid-19. Of these, 17 (3%) were from food outlets and restaurant settings.   

Incidents in educational settings were most common, followed by care homes. 

The criteria for an ARI incident means that this data does not cover all Covid-19 cases. For example, if just one new Covid-19 infection was linked back to a pub setting and there was no suspected outbreak, this would not be classed as an ARI incident. 

Over this period there were 26,259 new cases of Covid-19 confirmed in England, compared to 532 “incidents”, though each incident can also include multiple cases. The PHE data does not show how many people were infected in each incident.

So rather than 3% of all Covid-19 cases coming from restaurants in the week before the 10pm curfew, as was reported, in actual fact it was 3% of reported incidents linked to Covid-19 investigated by PHE.

The PHE report also says that, since 10 August, all people who test positive for Covid-19 and are traced by the Test and Trace system in England are asked about “places they have been and activities they have done in the days before becoming unwell (in the 2-7 days prior to symptom onset).”

It adds: “Although this can’t say for certain if this is where someone picked up the infection, the information may be helpful to indicate possible places where transmission is happening.”

Between 10 August and 24 September, the most common event recorded was eating out, reported by 14.6% of people who tested positive, followed by shopping, reported by 13.4%. 

Cause and effect

Even just looking at confirmed ARI incidents, the phrasing used by the newspapers of certain activities having “caused” the spread of Covid-19 is not strictly accurate.

It's possible that Covid-19 picked up in restaurants and pubs may be recorded as an incident in another setting. 

For example, a child picking up Covid-19 at a restaurant and then spreading it in school could be recorded as a school outbreak, despite the initial case having been brought in from a restaurant. 

Alternatively, if two adults in the same workplace contracted Covid-19 at separate pubs, neither of which was identified as a location of an incident by PHE, then in some cases their workplace would be listed as the setting of this incident, despite the fact that both cases were contracted at pubs.