Jump to comment:

• In reply to Pimenta
  John P.A. Ioannidis
  Published on: 22 February 2021
• A successful demonstration of why 'precision shielding' is impossible with high community transmission
  Dominic R Pimenta
  Published on: 15 February 2021

• Published on: 22 February 2021

  In reply to Pimenta

  • John P.A. Ioannidis, Professor Stanford University

  Many thanks to Dominic Pimenta for the interesting comment. I respect Pimenta’s well-intentioned activism during the COVID-19 crisis, but here he polarizes the discussion between two schools of thought by using a strawman argument, i.e. that one of the two schools wants to promote high community transmission. Such strawman arguments are prevalent in social media and the blogosphere, but they do not serve scientific discourse for resolution of major questions. I have signed neither the Great Barrington Declaration nor the John Snow Memorandum, so I cannot become an insider apologist for either (1). However, my reading of both documents suggests that neither of them advocates to promote high community transmission. If they do, this is certainly not what I would personally advocate.

  Pimenta draws a correlation from 10 observations on the data that I present on nursing home shielding factors (my Table 3) reaching the conclusion that precision shielding is impossible under high community transmission. This is a precarious exercise with rushed conclusions. These are ecological, whole-country data including only 4 observations with high community transmission. Drawing firm causal inferences from an ecological regression with effective sample size of n=4 is impossible. Sadly, over-confident, stretched causal inferences are common during the COVID-19 crisis. A similar look at the age-stratified data on Table 2 shows that shielding of the elderly was achieved in countries w...

  Show More

  Many thanks to Dominic Pimenta for the interesting comment. I respect Pimenta’s well-intentioned activism during the COVID-19 crisis, but here he polarizes the discussion between two schools of thought by using a strawman argument, i.e. that one of the two schools wants to promote high community transmission. Such strawman arguments are prevalent in social media and the blogosphere, but they do not serve scientific discourse for resolution of major questions. I have signed neither the Great Barrington Declaration nor the John Snow Memorandum, so I cannot become an insider apologist for either (1). However, my reading of both documents suggests that neither of them advocates to promote high community transmission. If they do, this is certainly not what I would personally advocate.

  Pimenta draws a correlation from 10 observations on the data that I present on nursing home shielding factors (my Table 3) reaching the conclusion that precision shielding is impossible under high community transmission. This is a precarious exercise with rushed conclusions. These are ecological, whole-country data including only 4 observations with high community transmission. Drawing firm causal inferences from an ecological regression with effective sample size of n=4 is impossible. Sadly, over-confident, stretched causal inferences are common during the COVID-19 crisis. A similar look at the age-stratified data on Table 2 shows that shielding of the elderly was achieved in countries with high community transmission (e.g. UK and USA), while it was probably not achieved in Hungary, despite low community transmission. China exemplifies significant inverse protection, even though it managed to practically bring community transmission to almost zero.

  For the 4 countries with high S values in nursing homes and high community transmission (Spain, UK, Belgium, and Sweden), there are many other explanations for high S values, as I explain in my article. With limited testing of staff, suboptimal infection control, poor living conditions, and with staff making shifts across multiple nursing homes, disaster is expected. Not only nursing home residents, but also their staff should be handled as high-risk individuals. Unfortunately, the opposite happened, perhaps because of flawed over-confidence that general population measures would suffice. Rational efforts to reduce community transmission are most welcome, provided that they do not do more harm on several other aspects of health and societal well-being (2). However, it is naïve to expect that simply taking community measures will salvage nursing homes without additional, draconian, focused measures. To use an analogy, when a known serial murderer has declared his clear intention to kill someone, the specific candidate victim should be immediately, actively, specifically protected. Dispatching more policemen at random locations around the country or preaching to a community congregation to love each other will not save his life.

  John P.A. Ioannidis, MD, DSc
  Stanford University, Stanford, CA, USA. E-mail: jioannid@stanford.edu

  1. Ioannidis JP. Scientific petitions and open letters in the era of covid-19. BMJ. 2020 Oct 26;371:m4048.
  2. Ioannidis JPA. Global perspective of COVID-19 epidemiology for a full-cycle pandemic. Eur J Clin Invest. 2020 Oct 7:e13423.

  Show Less

  Conflict of Interest:
  None declared.

  • Back to top
• Published on: 15 February 2021

  A successful demonstration of why 'precision shielding' is impossible with high community transmission

  • Dominic R Pimenta, Research Physician, Chairman Healthcare Workers' Foundation (Reg Charity 1189737)

  The article is predicated on the notion that “The ability to preferentially protect high-risk groups in COVID-19 is hotly debated.” This is a mischaracterisation of the debate. The ability to protect high-risk groups has never been questioned.
  The debate would be better characterised as the ability to focus non-pharmaceutical interventions solely on those most at risk of severe disease due to COVID-19, and removing restrictions and allowing widespread infection amongst the low risk groups. The debate has only ever been between those who advocated only applying NPIs to the highest risk cohorts ("focussed protection") and those that advocate that community wide measures are the only effective means to protect those at highest risk.
  The stated aim of this paper is to demonstrate whether ‘precision shielding’ was achieved in the first wave, and invents a metric to compare incidence and death in the high risk long-term care population vs the low risk younger population. The author uses seroprevalence data from the elderly and the young to estimate infection prevalence, but takes no account of the differing antibody response that would largely confound this approach.
  Using these metrics the author then goes on to ascribe a value judgement as “substantial shielding” or “substantial inverse protection.” Care homes and institutions are already known to be at risk environments for the spread of infectious disease, due to the close contact care work, the m...

  Show More

  The article is predicated on the notion that “The ability to preferentially protect high-risk groups in COVID-19 is hotly debated.” This is a mischaracterisation of the debate. The ability to protect high-risk groups has never been questioned.
  The debate would be better characterised as the ability to focus non-pharmaceutical interventions solely on those most at risk of severe disease due to COVID-19, and removing restrictions and allowing widespread infection amongst the low risk groups. The debate has only ever been between those who advocated only applying NPIs to the highest risk cohorts ("focussed protection") and those that advocate that community wide measures are the only effective means to protect those at highest risk.
  The stated aim of this paper is to demonstrate whether ‘precision shielding’ was achieved in the first wave, and invents a metric to compare incidence and death in the high risk long-term care population vs the low risk younger population. The author uses seroprevalence data from the elderly and the young to estimate infection prevalence, but takes no account of the differing antibody response that would largely confound this approach.
  Using these metrics the author then goes on to ascribe a value judgement as “substantial shielding” or “substantial inverse protection.” Care homes and institutions are already known to be at risk environments for the spread of infectious disease, due to the close contact care work, the multiple staff per patient, and the vulnerability and general immobility of the patient population as a whole. Terming this as “inverse protection” adds very little clarity to this area.
  On the basis of this analysis the author concludes 'precision shielding' is possible in 'real-life circumstances' but takes no account of the community-wide transmission which is the actual focus of the debate, and the basis for the statement that this approach is not feasible in reality. When correlated with the cumulative cases/million by the end of the first wave (ourworldindata.org) the shielding metric the author lays out for these 10 countries shows a very strong correlation: (Rsq of 0.82), showing that in the presence of high community transmission, 'precision shielding' isn't possible, while very effective community-wide suppression, such as in South Korea, made this much more feasible. An effective argument, amongst many others including economic impact, ethical considerations, multi-generational families, unknown risk in the community and long-term COVID morbidity, why 'focussed protection' or 'precision shielding' is unfeasible in the real-world.

  Figure 1:
  Can be viewed here: https://pbs.twimg.com/media/EtELbUkXAAA2X9m?format=jpg&name=large

  Show Less

  Conflict of Interest:
  None declared.

  • Back to top