Wednesday, April 1, 2020

The Effectiveness of Social Distancing on Pandemic Viral Transmission

Social distancing and self-quarantining are two concepts that the vast majority of people were not aware of prior to the current pandemic.  As you will see in this posting, apparently these ideas that are new to us are not particularly new to the academic community.  A recent early release Policy Review which is found on the Centers for Disease Control and Prevention website by Min W. Fong et al entitled "Nonpharmaceutical Measures for Pandemic Influenza in Nonhealthcare Settings - Social Distancing Measures" looks at the evidence-based research into the effectiveness of social distancing measures as a non pharmaceutical intervention (NPI) when battling an influenza pandemic.  What I found particularly fascinating about this research is that it took place just prior to the current COVID-19 pandemic and yet it is very pertinent to our current global health situation.

Here is the abstract of the Policy Review that appears in the Emerging Infectious Diseases Journal Volume 26, Number 5 - May 2020:

The authors open with this:

"Experiences from previous influenza pandemics, in particular the 2009–10 pandemic, have demonstrated that we cannot expect to contain geographically the next influenza pandemic in the location it emerges, nor can we expect to prevent international spread of infection for more than a short period. Vaccines are not expected to be available during the early stage of the next pandemic, and stockpiles of antiviral drugs will be limited, mostly reserved for treating more severe illnesses and for patients at higher risk for influenza complications. Therefore, nonpharmaceutical interventions (NPIs), such as social distancing, will be heavily relied on by health authorities to slow influenza transmission in the community..."

The authors note that influenza virus infections are transmitted by close human-to-hyman contact in schools, public places, workplaces, homes and other crowded settings.  Social distancing measures are implemented with the aim of reducing this contact and form a key part of current pandemic preparedness planning.  The purpose of the study was to review the evidence base for social distancing measures with the aim of uncovering evidence regarding the effectiveness of social distancing at reducing community influenza transmission.

The authors looked at six methods for reducing influenza transmission as follows (note that all bolds are mine):

1.) Isolating ill persons at home - since medical facilities are unlikely to be able to handle all infected persons, governments will require that infected persons self-quarantine in their homes.  Here is what the authors found from their study of scientific papers on isolation of ill persons:

"we found limited scientific evidence to support the effectiveness of this intervention in the community. The observational studies included in this review were conducted in atypical settings, and the effectiveness of isolation in these settings might not be generalizable to the community-at-large. Nonetheless, with the rationale discussed, and assuming that a high level of compliance with home isolation is possible for symptomatic persons, voluntary home isolation could be a preferable strategy to prevent onward transmission compared with other personal protective measures, which have not shown effectiveness in multiple randomized controlled trials.

One area in which there is a lack of evidence is the duration of infectivity, which has implications for the period of voluntary isolation. Current recommendations include voluntary isolation until cessation of fever or until 5–7 days after illness … Another area of uncertainty is the degree to which transmission occurs before illness onset (presymptomatic transmission) and the degree to which mild or asymptomatic cases are infectious. If there is a substantial fraction of asymptomatic transmission, this fraction would reduce the impact of isolation."

2.) Contact Tracing - the authors examined four simulation studies which looked at the impact of tracing persons who have come in contact with an infected person.  They found that contract tracing is effective when used with other interventions including isolation, quarantine, and prophylactic (i.e. preventative) treatment with antiviral drugs.  That said, they note the following:

"the addition of contact tracing to an existing combination of quarantine, isolation, and antiviral prophylaxis measures would only provide modest benefit, while increasing considerably the proportion of population in quarantine and the consequent costs.

Contact tracing requires substantial resources to sustain after the early phases of a pandemic because the number of case-patients and contacts grows exponentially within a short generation time. Therefore, there is no obvious rationale for the routine use of contact tracing in the general population for control of pandemic influenza." 

3.) Quarantine of Exposed Persons - the authors examined 1 intervention study, 5 observational studies and 10 simulation studies and concluded the following about home quarantine:

"Overall, we found that the evidence base was weak for home quarantine. In general, the intervention is estimated to be effective. However, being able to identify case-patients and their close contacts in a timely manner can be challenging during the early phase of a pandemic, and impossible for health authorities after the early phase. Quarantine also raises major ethical concerns regarding freedom of movement because the evidence on the effectiveness is limited, providing no solid rationale for the intervention, in addition to restricting movement of some uninfected and noninfectious persons. The increased risks of infection among quarantined further exacerbate the ethical concerns. Therefore, voluntary/self-quarantine is likely to be preferred over mandatory quarantine in most scenarios….Prolonged quarantine can cause substantial burden to social services and working persons. Some measures can be taken to minimize the possible harms, such as pairing quarantine with antiviral prophylaxis provision for the household."

4.) School Dismissals or Closures -  A systematic review of studies on school closures found compelling evidence that school closures could reduce influenza transmission, particularly among school-aged children.  Preemptive school closures may also reduce mortality and transmission rates.  Unfortunately, while planned school holidays did reduce influenza transmission and delay the epidemic peak, in some instances, the transmission rate resurged after the schools reopened:

"…closing of schools after the epidemic peak showed little impact on the overall attack rate and none on the timing of the peak or the size of the epidemic peak because it has already passed. In other past epidemics, transmission resurges after schools reopen, so that the closures delayed the epidemic peak but might not necessarily have reduced the size of the epidemic peak or the overall attack rate."

5.) Workplace Measures and Closures - the authors examined 10 simulation studies, a systemic review of workplace measures and 3 recently published studies and concluded the following:

"there was evidence, albeit weak, to indicate that these measures could slow transmission, reduce overall attack rates or peak attack rates, and delay the epidemic peak…in general, the simulation studies predicted that workplace closures would be able to reduce transmission somewhat in the community, but probably would have a smaller effect on transmission than school closures.

We found limited evidence that workplace measures and closures would be effective in reducing influenza transmission…however, workplace measures and closures could have considerable economic consequences, and inclusion in pandemic plans would need careful deliberations over which workplaces might be suitable for application of interventions, whether to compensate employees or companies for any loss in income or productivity, and how to avoid social inequities in lower income workers, including persons working on an ad hoc basis.

6.) Avoiding Crowding - the authors reviewed 3 observational studies and found that timely bans on closure of public places including churches and theatres had a positive influenced on reducing the excess death rate during the 1918 Spanish Influenza pandemic:

"The evidence for avoiding crowding is limited. The implementation of measures to avoid crowding might require a large amount of resources (e.g., financial and trained personnel), which might be less feasible in low-income and middle-income countries."

To summarize, the authors found that social distancing measures are useful components of a public health response to a pandemic but that the overall quality of evidence proving the effectiveness of using social distancing is "relatively low" since most of the evidence is currently based on observational and simulation studies.  In order to prove the effectiveness of  social distancing measures, controlled studies of both single and combined interventions are needed to clarify the following:

1.) transmission dynamics of influenza

2.) role of presymptomatic contagiousness

3.) fraction of infections that are asymptomatic

4.) optimal timing and duration for the implementation of distancing measures

It has become increasingly apparent that health authorities and governments around the world are using social distancing and self-quarantine as the best solution to the COVID-19 pandemic when, in fact, this analysis shows that these two nonpharmaceutical interventions may be far less effective than the authorities want us to believe.  They are, however, very effective at destroying an economy and reducing our freedoms.


  1. Going from your summary, the Chinese, Taiwan and South Korea experiences suggest that they are likely wrong on 2. Contact Tracing, on Ghinese experience wrong on 5. Workplace Measures and Closures and probably right on home quarantine and School Closures.

    I wonder just how realistic those assumption are that went into the model?

  2. What we really need to know about the coronavirus is how it will affect us as individuals. A big problem with getting good information about how fast this virus is spreading comes from the fact we simply don't know, how long a person should be quarantined and how long the virus can live on different surfaces is still being debated. More about this subject in the article below.