Fiasco or mobilization?

Dr. John Ioannidis

Dr. John Ioannidis is well known as the author of the manifesto of reproducibility in research, Why Most Published Research Findings Are False. He published that in 2005. I learned of it much later when I attended an inspiring seminar by Edward Tufte.

Fast forward to now, and I’m on fire to apply data science to public health. When I turn my sights on Ioannidis again, I realize he is a professor of epidemiology and statistics. Those weren’t on my radar when I first encountered him, so it didn’t sink in back then.

Coronavirus and A fiasco in the making?

Dr. Ioannidis published an opinion piece on StatNews dated March 17, 2020 and entitled A fiasco in the making? As the coronavirus pandemic takes hold, we are making decisions without reliable data.

His thesis is we need to get better data in order to make informed policy decisions. He discusses his article in an interview for Journeyman Pictures.

His observations have provoked debate. I have my own concerns about his take on things, although I find the discussion healthy. Here’s an overview of his article.

Main points

  • We lack reliable evidence about the infection rate of SARS-CoV-2.
  • The world has responded with onerous policies. Ioannidis questions that we know enough to be able to assess benefits against costs.
  • There are shortcomings in data collection. For example, mortality rates are unreliable if we haven’t measured accurately the size of the population of infected people. We just haven’t counted them reliably.
  • Ioannidis attempts to generalize the infection and mortality rates of the Diamond Princess to the U.S. population by adjusting for age.
  • He comments on disruptive policies like social distancing and school closings, maintaining that we do not know if they work and we don’t know the consequences.
  • Ioannidis concludes by stating a worst case scenario and observing that in this case, the lives lost will come in great measure from a population soon to die.

Rebuttals

Hilda Bastian

Hilda is a member of the editorial board of the independent Drugs and Therapeutic Bulletin, and is the lead for the independent advisory group on the update of the Cochrane review on exercise therapy and ME/CFS. She is a member of the PLOS One Human Research Advisory Group, and a contributor to WIRED.

A Rebuttal to “A Fiasco in the Making?”.

I like the distinction Bastian makes between evidence-based medicine and public health interventions.

Marc Lipsitch

Marc Lipsitch, D.Phil., is professor of epidemiology at the Harvard T.H. Chan School of Public Health and director of Harvard’s Center for Communicable Disease Dynamics.

StatNews published Dr. Lipsitch’s response, We know enough now to act decisively against Covid-19. Social distancing is a good place to start.

We know enough to act; indeed, there is an imperative to act strongly and swiftly.

My two cents

I think everyone would agree that our coronavirus data have gaps and inaccuracies at this date and would agree that geting better data is beneficial. Dr. Ioannidis’s suggestions about what data to seek and how to get it sound helpful to me. In so far as his article being a call to arms to go do that, so far so good.

The part that makes me scratch my head is where he appears to dig in his heels at the level of the world’s policy response to the coronavirus threat. He appears to suggest more circumspection before acting.

My earlier career in insurance leads me to interpret the situation in terms of risk assessment and management. You don’t manage risk by waiting until you know what is going to happen. It’s precisely because you don’t know what will happen that it is called risk. You manage risk through prevention, mitigation, and risk transfer (insurance).

For example, even if you don’t know the value of the contents of your house or the likelihood of fire, when you see a bucket of oily rags sitting next to your furnace, you move it away. That’s mitigation of risk.

Changing subject for a moment, I just read a paper about opioid abuse in the U.S. written by my math instructor, Lawrence Fulton. (Geospatial-Temporal and Demand Models for Opioid Admissions, Implications for Policy. 2017. Lawrence Fulton, et al.) It pointed out that the U.S. Department of Health and Human Services estimates that 91,548 people died from opioids, synthetic opioids, and heroine in 2016. As I write in April of 2020, the estimate of U.S. deaths from COVID-19 stands at about 30,000. That’s a mere third of the deaths in a year from opioids. Why didn’t we lock down the country over opioids when the severity of opioid abuse measures three times that of COVID-19 right now?

There’s that word “risk” again. We just don’t know what the severity of infection by SARS-CoV-2 will be. It spreads rapidly, and it has killed a lot of people already. Ioannidis seems to suggest putting the skids on drastic policy measures until we know more. Okay, the time I’ve spent thinking about epidemiology lies closer to 10 minutes than to the time he’s spent thinking about it. Nevertheless, I’m skeptical of his skepticism.

The most uncomfortable part of his essay is the concluding section. Dr. Ioannidis paints a picture of a worst case scenario of an extreme number of deaths, and he says that he doesn’t even believe that estimate. Yet, he continues with observing that in that case, the majority of those who would perish would be people already close to death anyway.

He doesn’t state explicitly the point of making that observation, an observation that already appears counterfactual to me given the overwhelming mitigation the world has already undertaken. Instead, he kind of just lets it hang there, then concludes his article with,

If we decide to jump off the cliff, we need some data to inform us about the rationale of such an action and the chances of landing somewhere safe.

I have a feeling that if the guy sat in a room with me and we had a chance to talk this out, he could clarify his points and correct any of my misunderstanding about his article. We might find we even agree more than we disagree. Since that hasn’t happened, the apparent tone of his observations seems cringe-worthy to me. I can’t help but think of this scene from Monty Python and the Holy Grail.

This dark comedy tells of the Plague and the sketch here is about collecting the dead bodies. One of the bodies perks up and says, “I’m not dead yet!” The guy hauling him to the cart says, in effect, “Shut up, you will be soon.”

I get the critical care concept of triage, and I wouldn’t be surprised if you told me that, in public health, similar evaluations of priorities may be required at the level of a community or population. However, although I think Dr. Ioannidis’s zeal to acquire better data is appropriate, I don’t think he is prepared to undertake prudent risk management when the data at hand isn’t perfect, which is to say, always.