When Studies Collide (COVID-19 Edition)

#15,187

Depending on which `study' and/or newspaper accounts of COVID-19 infection rates you read over the past 24 hours, you might believe that  Europe is either substantially on its way towards achieving herd immunity, or not even close.

Two examples - one from Germany (Spiegel) and the other from the UK (The Guardian) - follow:

One in seven could already be immune

How does the corona virus spread and how dangerous is it? The first interim results are available for the particularly affected NRW district of Heinsberg.

          April 9, 2020, 11:59 a.m.    

(Excerpt)    

According to study leader Hendrik Streeck, the first results suggest that the stringent requirements to contain the epidemic could be gradually relaxed if hygiene measures are continued.

Less than 1% of Austrians infected with coronavirus, study shows

Research appears to scotch hopes of countries being close to relying on ‘herd immunity’

          (Excerpt)

The co-founder of Sora, Christoph Hofinger, told a news conference: “Based on this study, we believe that 0.33% of the population in Austria was acutely infected in early April.” Given the margin of error, the figure was 95% likely to be between 0.12% and 0.76%.

In truth, neither of these media reports are based on robust, peer-reviewed, data.  

• The Austrian study - commissioned by their government - is based on polling data, which is highly subjective, and is not backed by laboratory tests.  

• The German study, is based on lab testing, but of only a small number (n=500) people in one town, and we have no data on the sensitivity or specificity of the test.  

Honestly, either one could be correct, or off by a country mile. Frustratingly, in the media's rush to `feed the beast' with updated news every 10 minutes, both are presented as `proof' of sorts - when at best - they are small snapshots of what may be happening in one location. 

Already I'm seeing pushback on the German study:

Criticism of Corona study from Heinsberg

Implausible figures and not transferable to the whole country: experts doubt the meaningfulness of the Heinsberg study. It gives no reason to loosen contact blocks.

Meanwhile the M.I.T. Technology Review opines that if the numbers are correct in the German study, they are still far lower than what would be needed to begin to roll back social distancing measures.  `Herd Immunity' would require at least a 5-fold increase in seropositivity. 

And that assumes that mild or asymptomatic infection actually conveys reasonably long lasting immunity.   We assume it does, but it isn't a proven fact. 

In April of 2016, in EID Journal: Antibody Response & Disease Severity In HCW MERS Survivors, we looked at another study that tested 9 Health care workers who were infected during the 2014 Jeddah outbreak (2 severe pneumonia, 3 milder pneumonia, 1 URTI, and 3 asymptomatic), that found only those with severe pneumonia still carried detectable levels of antibodies 18 months later.

Those who experienced a milder pneumonia had shorter lived antibody responses (1 out to 10 months, 2 out to 3 months), while the URTI and asymptomatic cases tested negative at 3 months post infection.

Will COVID-19 follow suit?  We simply don't know at this point.

Over the years we've looked at dozens of conflicting studies, and we can expect to see many more in the future.  It is a rare thing when a study produces a definitive answer on anything. `Facts' have a finite shelf life, and as science advances, their shelf life grows smaller every year.

That isn't a slam against researchers, just an acceptance that the pursuit of scientific knowledge is incremental, that findings can vary widely depending on the methods and data used, and there are sometimes going to be detours along the way.

For now, an extra dash of Caveat Lector on COVID-19 studies is probably warranted, given the  limited data available, and this novel virus's propensity to surprise.