Posted 14 June 2020
Apparently the Government is seeking an urgent review of the science behind the two metre rule.
Allow me.
As far as I can see from my own on-line searches, there have been four published systematic reviews of physical interventions to prevent or reduce the spread of infection from respiratory viruses. Two of these pre-date SARS CoV2 and the other two have been published during the current pandemic.
Essentially, a systematic review is a structured trawl through published research papers looking for studies that match specific search criteria. It is like performing a well defined Google search of all locations that are repositories for research papers. The search ‘questions’ and methods in such systematic reviews have to be well defined with the intention of allowing others to re-perform the same search queries if they so wish, to corroborate the findings. Other researchers can therefore check there has been no cherry-picking of papers.
Reviewing systematic reviews is therefore a little bit like cheating as others have already done the hard work scouring the research databases.
On a Sunday afternoon I do not mind piggy-backing on other peoples’ efforts.
The first paper found, published in 2011 and obviously therefore pre-COVID19, reported that it could find few studies looking at social distancing as a method of managing respiratory viral spread. The paper concluded that there was limited evidence that social distancing was effective as a method of reducing viral spread during epidemics. The paper rather concluded that physical barriers (masks), hygiene (hand washing) and isolation (of the infected and vulnerable) were more effective.
The next review was published in 2018. This looked at a number of modelling studies (computer simulations) and three population (epidemiological) studies. None of the studies looked at social distancing (keeping all people x metres apart in shops, restaurants, bars, public transport, parks etc) and so cannot give direct insights into the ‘2m rule’ being examined here. The epidemiological studies looked at workplace interventions (staggered mealtimes, small group working, working from home, temperature checking etc) – useful in the overall COVID19 debate but not relevant to the question being asked by the Government today.
The authors of this review considered the data from the three epidemiological studies to be subject to bias. The authors considered this bias to be ‘serious’ in two of the studies and ‘critical’ in the third. Therefore even if these studies had looked at 2 metre social distancing (which they did not), using the results would be potentially meaningless. Using the results from the remaining modelling studies would always be difficult as models are only as good as their assumptions as we saw with the Imperial College modelling.
Now I turn to the papers that have been published during the current pandemic.
The first of these papers appeared in May 2020. This paper looks at influenza transmission against a number of methods of social distancing. The methods investigated included: isolating ill persons, contact tracing, quarantining exposed persons, school closures, workplace interventions and crowd avoidance. This paper therefore did not assess 2 metre social distancing and so cannot contribute directly to the analysis here.
How much the reviewed studies in this paper could contribute to debates on contract tracing, school closures etc is open to question as again many of the included studies were simulations rather than epidemiological studies.
This leaves the final systematic review which was published in The Lancet on 1 June 2020.
Potentially this paper ticks lots of boxes for my analysis. It looks only at studies of coronaviruses – MERS, SARS and SARS CoV2 – rather than influenza. It also looks specifically at social distancing in the context I am considering here (ie the queue outside Sainsbury’s, the wide berth we are giving each other in the park and so on). The paper also looks at face masks and eye visors but I shall stick to the brief of looking at physical distancing.
In the Methods section of the paper the authors state that they were investigating the optimum distance for avoiding person-to-person virus transmission. Perfect.
This study is also what we call a ‘meta-study’ in that not only does it find studies that fit the search criteria, but it uses clever statistics to combine the results from those studies to draw pooled conclusions. This has the potential to be some very robust (possibly even, exciting) analysis.
The authors of this paper noted that there were no randomised control trials (RCTs). These are the only real studies that can show ‘cause and effect’. Putting aside the ethics of such studies, one could imagine that a RCT looking at physical distancing and viral transmission would place test subjects at measured distances from infected and non-infected people (the subjects would not know which) and then seeing if they contracted the infection. Those studies would be the only way to clearly show the effectiveness of various physical separations. Without such studies, we at best see correlations and no more.
The authors in this study found 172 studies in total but using clearly defined selection criteria they reduced this to 44 comparative studies. The authors come to the bold conclusion that “Transmission of [corona]viruses was lower with physical distancing of 1 metre or more, compared with a distance of less than 1 metre…….Protection was increased as distance was lengthened”. They state that the risk halved with every extra metre of physical separation.
It would seem we have the evidence that the Government is looking for that 2m separation works.
Maybe. Maybe not. Let me play with the numbers and see for myself.
Firstly, have you noticed a discrepancy already? This paper specifically mentioned 1 metre. The UK Government have plumped for 2 metres. The WHO have stated 1 metre at least as the safe distance Germany went for one and a half metres. Sweden and Singapore went for 1 metre. Which is it? If there is no agreement at Governmental level then you have to start suspecting that the quantity is not backed by a strong single evidence base.
Skipping quickly over this potential stumbling point, let me focus instead on what this paper says to support or otherwise the UK Government’s 2 metre rule as applied to shops, pubs, restaurants, schools, gyms and so on. I shall therefore exclude the studies that looked at clinical settings and also exclude those that looked at face masks and eye protection. (If either of these physical barrier methods was used, this confounds any analysis of just the physical separation). This leaves just nine studies.
I shall not list the details of all nine as that will get a little long-winded but there is something I need to flag about how the results of these individual studies were used by the authors of this meta-study.
Let me look at just two of the studies, one looking at MERS and importantly, one of the three that relate specifically to SARS CoV2.
Firstly, a study that looked at the transmission of the MERS coronavirus in Saudi Arabia. This looked at occupants of a female dormitory with 828 residents. One woman initially with the infection transmitted the infection to 19 fellow residents. The study reported that 54 women had ‘direct’ contact with the original infection (assumed to be a zero metres separation) and 11 of these women contracted the infection. Of the remaining 774 residents ‘not in close contact’, 8 became infected. The meta-study then uses these data: 11/54 is the risk of infection at 0-1 metres and 8/774 is the risk of infection at more than 0-1 metres.
I cry foul here. How do we know these differences are due to the distance of separation? Is it not as much or more about contact with kitchens, toilets, bathrooms? Mugs, cutlery and plates, perhaps? What about duration of contact? A glancing 0 metres contact versus 3 metres separation all night in adjacent dorm beds? None of this is considered and instead it is all assumed to be about distance, and a pretty arbitrary defined one at that.
The second study I want to examine is from Taiwan and looked at 1,043 people who had been in contact with 32 individuals who were all laboratory confirmed cases of COVID19. Of the 1,043 contacts, 36 were from the same households and 47 from the extended families of the 32 infection cases. In total there were 12 transmissions in this 1,043 group: 7 from the households and 5 from the families.
For the meta-analysis, a key assumption was made. Those contacts in the households were at ‘zero’ metres separation from the infected people and those contacts in the families were at 1m or more separation from the infected people. (So, the meta-study ‘calculates’ that the risk of infection at 0 metres is 7/36 and the risk of infection at 1 metre is 5/47).
This is nonsense.
How do you know that the family members were never at less than 1m separation from the infected individuals? Why would you assume that all the household members spent the entire time at 0 metres separation from the infected individuals? How do you know that the remaining 960 contacts were at no point less that 1 metre away from the infected individuals? How do you know that there were only 960 other people 1 metre away from the infected individuals?
And yet these calculations (and 6 more sets like them from the other studies) are how this paper arrived at the conclusion that “Transmission …was lower with physical distancing of 1 metre or more…and .….protection was increased as distance was lengthened”.
All the papers really showed was that spouses, carers, people who shared the same bed, bathroom, dormitory etc were at more risk of infection than those who did not. Intimate contact versus not, seems to be the main factor. What the ‘statistical shenanigans’ of this paper did was to say that intimate contact equals a physical distancing of 0-1 metre; non-intimate contact was physical distancing of more than 1 metre.
So, when it comes to deciding if restaurants, bars, gyms, pubs, play areas, schools, shops should abide by the 2 metre rule just remember what the science says: If you hug, kiss, share a bed with someone then your risk increases. If you do not, the science says little else.
Of course, there will be people who are scared and will want to keep separation. There are those whose livelihoods have not been threatened by the social distancing and so they have nothing to lose by maintaining it. But there are millions of people financially and emotionally affected by it, and will be affected for years to come. The science has yet to show their sacrifice was worthwhile.