Lockdown # 4

Again an old article from May that was not PEER REVIEWED

Abstract

The current epidemic of COVID-19 is unparalleled in recent history as are the social distancing interventions that have led to a significant halt on the economic and social life of so many countries.

 However, there is very little empirical evidence about which social distancing measures have the most impact. 

We report a quasi-experimental study of the impact of various interventions for control of the outbreak.

 Data on case numbers and deaths were taken from the daily published figures by the European Centre for Disease Control and dates of initiation of various control strategies from the Institute of Health Metrics and Evaluation website and published sources. 

Our primary analyses were modelled in R using Bayesian generalised additive mixed models (GAMM). 

We found that closure of education facilities, prohibiting mass gatherings and closure of some non-essential businesses were associated with reduced incidence 

whereas stay at home orders, closure of all non-businesses and requiring the wearing of facemasks or coverings in public was not associated with any independent additional impact. Our results could help inform strategies for lockdown

Discussion

We have undertaken a quasi-experimental study of the impact of various forms of social distancing interventions on the epidemics of COVID-19 infection in 30 different European countries. 

Our analyses confirm that the imposition of non-pharmaceutical control measures have been effective in controlling epidemics in each country. 

However, we were unable to demonstrate a strong impact from every intervention.

 Closure of educational facilities, banning mass gatherings and early closure of some but not necessarily all commercial businesses were all associated with reduction of the spread of infection. 

Widespread closure of all non-essential businesses and stay at home orders seem not to have had much if any value.

 The results on face coverings are too preliminary to be reliable but what results are available do not support their widespread use in the community.

 Spatiotemporal hierarchical models, like the one presented here, have the advantage of being able to explicitly quantify the probability that an epidemic may or may not occur at a specific time or location. 

We have previously used similar models to account for spatiotemporal effects in health outcomes.11

Public health officials may be more inclined to deploy interventions if the probability of an epidemic exceeds a certain value.

 Ideally, public health decision-makers should agree on the specific epidemic thresholds (i.e. the incidence above which the disease requires imposition of control measures) to make model predictions meaningful. 

Whether or not school closures are likely to have been important in controlling the spread of epidemic disease is an issue of some debate in both the scientific and lay media. 

There has been uncertainty about how beneficial the closing of educational establishments can be on respiratory disease transmission 12-14. The value of school closures is particularly uncertain for COVID-19 given the observation that children have only mild or no symptoms 15. 

Decline in the infectiousness of the SARS-1 outbreak in Hong Kong in 2003 was also a time when many interventions were implemented, including school closures 16, making it hard to disentangle contributions of each individual measure. 

In the current pandemic, Hong Kong managed to substantially reduce the transmissibility of COVID-19 fairly early in the outbreak by a limited number of interventions one of which was keeping schools closed 17.

 However, there were also substantial behavioural change in the population at the time coincident with these interventions. 

Viner and colleagues state “Data from the SARS outbreak in mainland China, Hong Kong, and Singapore suggest that school closures did not contribute to the control of the epidemic”.

 However, this is not a valid argument against school closures as peak shedding of virus in SARS-CoV was around day 10 whereas peak shedding of SARS-CoV-2 is much earlier and possibly before symptoms develop 18-20.

 In contrast to COVID-19, SARS was primarily infectious after onset of symptoms when most cases would have been hospitalised or at least quarantined.

 Throat swabs from children have shown similar viral load to those in adults21, yet a review of contact tracing studies failed to find incidents where transmission occurred from children to adults 21. 

This is the problem of Ivor cherry picking old articles

Here is current information

We cannot resolve the lack of consensus in these lines of evidence, about whether children can pass SARS-COV-2 to adults.

 What our study also does not do is identify which level of school closure has the most benefit whether it is primary, junior, senior school or even higher education. 

This will need to be the focus of further research.

 Note that the results presented here are based on total closure.

 It is possible that partial school closures such as three-day weekends could have worthwhile impacts on the spread of infection 22.

 The second greatest impact on the epidemiology of the European COVID-19 was from banning mass gatherings (which could be of any size), both public and private gatherings. 

A 2018 review of the impact of mass gatherings on outbreaks of infectious disease 23 found that most evidence was linked to the Islamic Hajj pilgrimage, where most infections were respiratory, mainly rhinovirus, human coronaviruses and influenza A virus. 

The evidence for outbreaks arising from other mass gatherings such as music festivals or sporting events is less established, but not absent. 

Several outbreaks of respiratory infectious disease have been linked to open air festivals and other music festivals 23, 24. 

For instance, during the 2009 influenza season pandemic influenza A(H1N1)pdm09 outbreaks were recorded at three of Europe’s six largest music festivals, while some 40% of pandemic flu cases that season in Serbia were linked with the Exit music festival.

 The link with business closures is particularly interesting. 

The two variables relating to business closures are when the first closures occurred and when all non-essential closures were enforced. 

For many countries but not all, these two events occurred at the same time.

 So, there may be substantial collinearity which would underestimate the impact of one or both. 

Nevertheless, it is worthy of note that the strongest association was with the initial closures. 

Given that those initial closures were mostly directed at business where people congregate (i.e. the hospitality industry), this would suggest that these businesses are where the most impact may be had.

 Although outbreaks of food poisoning are frequently linked with restaurants, outbreaks of respiratory infections are much more rarely so. 

One exception was an outbreak of SARS at a restaurant where live palm civets were caged close to customer seating 25. 

Our findings on facemasks or coverings are perhaps counterintuitive especially given the strong debate on their use. 

In a recent systematic review we concluded that the evidence in favour of face mask use outside of hospital was weak.26

On the other hand a recent modelling study concluded that community facemask use could reduce the spread of COVID-19.27

Our results on face coverings should be considered to be preliminary because the use of coverings was recommended or required only relatively late in the epidemics in each European country. The results for face covering are too preliminary to inform policy but indicates that face covering as an intervention merits close monitoring

Limitations

We acknowledge that lack of direct observation of these variations may have biased our results.