A friend of mine, sent me the following message:
"Thought you might like this piece. By John Ionnidis - Professor of medicine, of epidemiology and population health, of biomedical data science, and of statistics at Stanford University."
My (really long) response is below:
It’s an interesting piece. Ioannidis is well known for his (well-placed) skepticism of bad statistical analysis, and he brings up a number of good points.
I agree with about 90% of what he says. I am concerned by 9% and very troubled by 1%.
First the good:
His central premise, that we are flying blind right now, is spot on. It is a shame that the world has known about this virus for over 2 months and our vulnerability to the threat of a global pandemic for many years (see for example Bill Gates Ted talk from 2015), and yet we are so ill-prepared to monitor its spread.
He is right to point out that we don’t necessarily understand all the implications of school closures (for example, what does that do to parents who work in healthcare?) or how long the “flatten the curve” policies will need to be in place (is it sustainable for the amount of time required for success?). That is why I have tried to avoid commenting on whether the specific policies implemented here in the US are wise and instead focused on explaining the rationale behind them (if they are going to be implemented we might as well give them the best chance of succeeding). I simply do not have access to the kind of data that would be required to craft an informed policy proposal. I hope that those making these decisions do, but I am skeptical that anyone really knows how this is going to go.
He is absolutely right that we are almost certainly overestimating mortality rates right now since cases that end in death are far more likely to be documented than the mild and/or asymptomatic cases. The takeaway message is that long-term predictions have huge error bars.
One of the implications of this is that when experts start predicting numbers of cases and numbers of fatalities, you should take those predictions with a grain of salt. A model is only as good as the assumptions that it is based on and if you overestimate the mortality rate, you will overestimate the death totals. If you talk with an epidemiologist about their model, they will usually be very upfront about the assumptions and the fact that they recognize that there is a great deal of uncertainty about some of the parameters used to make predictions. Unfortunately, when results are reported by the media, those nuances are usually lost.
On to the concern:
It would be easy to misinterpret what he is saying. It would not surprise me if people were to use this article, to say “see, I told you everyone is overreacting”, but his point is more nuanced than that. If someone is claiming that millions of people in the US are going to die in a month unless we all lock ourselves in a bunker until May, then yes they are probably overreacting. However, his point is not that the current responses are inappropriate, but rather that we don’t know if they are appropriate. There is a big difference.
The problem is that inaction comes with just as many risks as action and we don’t have the luxury of waiting around to see how this plays out. Interventions become significantly less effective at mitigating an epidemic the later they are implemented. So, it is essential that some responses be implemented quickly (as they have been).
Lastly, the part that I find troubling:
Toward the end of the article he says the following:
“In the most pessimistic scenario, which I do not espouse, if the new coronavirus infects 60% of the global population and 1% of the infected people die, that will translate into more than 40 million deaths globally, matching the 1918 influenza pandemic.
The vast majority of this hecatomb would be people with limited life expectancies. That’s in contrast to 1918, when many young people died.”
I am not entirely sure what his point is with those last two sentences, but one possible reading is that 40 million deaths among the elderly is somehow a less significant tragedy than 40 million deaths among the young. I would be deeply uncomfortable about that line of reasoning. The idea that the lives of the elderly and the sick are somehow less valuable than those of the young and healthy is certainly unbiblical and dangerous. It would be irresponsible to downplay the severity of this crisis because those who are most likely to be affected were going to die soon anyway. I am not sure if that was his intent, but if it is, then I would strongly disagree.
So all that to say, for the most part I think the article offers a counterpoint to the chicken little’s of the world trying to persuade us that the sky is falling. However, it is important to safeguard against the opposite extreme as well: convincing people to ignore the threat that they face.