Better in their words…

Recently, I had the pleasure to chat with Dr. John PA Ioannidis about the pandemic. Dr Ioannidis is Professor of Medicine at Stanford University. He is the author of over 1000 peer reviewed articles, and is routinely featured on lists of the most cited scientists.

Previously, I have spoke to John in a series of interviews; Oct 2, 2020 – Plenary Session; Feb 2021 – Plenary Session; Dec 9, 2021 – Youtube interview

Now, in mid-2022, How does Prof Ioannidis take stock of where we are?

—Vinay Prasad, MD MPH

VP: We now enter the third year of the pandemic, and there are a few policy questions that remain. It appears that the vaccine, while still retaining benefits against severe disease and hospitalization, has proven that it is unable to suppress milder infections. It seems unable to halt transmission.  Do you share this assessment? Did you find it surprising? Or could it have been predicted with what we know about coronavirus?

JPAI: The mechanism of action of SARS-CoV-2 vaccines suggested upfront that they would probably not be very effective to halt transmission. The rapid development of vaccines that were apparently very effective for decreasing the risk of serious disease was an amazing success and it could have been a wonderful opportunity to showcase the power of science and to build more trust in public health that had suffered over the years from attacks from the anti-vax movement.

Unfortunately, this opportunity was lost, to a large extent by trying to push an inflated narrative that COVID-19 vaccines are perfect, the ideal silver bullet to put an end to epidemic waves, and having no side effects at all.

In December 2020 I wrote a paper where, based on mathematical modeling, I showed that once people started to increase their exposures again, vaccines with modest effectiveness in halting transmission would probably even lose all of their effectiveness for this outcome. medRxiv declined to post my paper as a preprint claiming it was dealing with a sensitive public health issue. arXiv also declined to post my paper as a preprint, the message I received was stunning: “Our moderators have determined that your submission is not of sufficient interest for inclusion within arXiv. The moderators have rejected your submission after examination, having determined that your article does not contain sufficient original or substantive scholarly research.”

I sent an appeal and the reply was even more stunning: arXiv, a preprint server, offered to post my paper as a preprint only AFTER it had been published by a conventional peer-reviewed journal! I submitted the paper to SSRN and then it got published in npj Vaccines, the vaccines journal of Nature, several months later.

By that time many public health authorities in many countries had fallen headlong in the trap of believing that people who get vaccinated will not transmit and vaccines all alone were enough to halt the epidemic waves. The consequences were grave.  In most developed countries, despite vaccination in 2021, excess deaths were higher in 2021 than in 2020.

VP: Does this fact have implications for mandates? A prerequisites to a mandate in my mind is that it confers benefit to a third party, that is no longer the case, ergo mandate should fall. What are your thoughts?

JPAI: The argument that vaccines can abort epidemic waves was a center piece of the push for aggressive mandates of all sorts that would have been previously unimaginable. This resulted in a superimposed blow against public health and its credibility, disruption to social cohesiveness and marginalization of many vulnerable people. Mandates may have increased uptake modestly in some circumstances, but they did damage to public health efforts in a wider perspective and may leave a nasty legacy behind.

 

VP: As we have moved from adult vaccination to kids vaccination, the evidence base is sparse. We are vaccinating children between the ages of 6 months and 4 years with a vaccine targeting the original Wuhan strain. Available evidence merely shows non-inferior antibody titer generation. Could we have gathered better evidence? In the absence of better evidence, should this be mandated?

JPAI: I should refer to a paper I published 9 months ago on vaccination in children. The main considerations remain the same and the evidence that has accrued since then is useful and largely reassuring, but not conclusive. I am very sensitive to child health and I wish we can make the best decisions, even if one more life can be saved.

However, we should also acknowledge that both benefits and risks in children are very small and thus they are probably beyond the full resolution capacity of the epidemiological microscope. I would have hoped to see larger randomized trials performed, with longer-term follow-up, including active surveillance of potential adverse events.

Still, one has to recognize that event rates in children are likely to be extremely low and residual uncertainty would remain even with substantially larger trials. This means that one can be reassuring that we have not seen major frequent problems with these vaccines in children and that we have also seen some supportive surrogate efficacy data. We can present the available data and the rough risk estimates of potential toxicity to parents, who can then decide – but not impose any mandates.

Of course, we should also tell parents what the official recommendations are, but then the problem is that often the official recommendations differ in countries say like the USA, Sweden, and Finland.

VP: MIS-C appears to a fallen in children independent of vaccination, due to the prevailing strain. Do you agree? Is this surprising?

JPAI: Data on Mis-C remain confusing to me. These data are collected with very weak designs, and they are likely to depend on ascertainment intensity and diagnosis bias. There are large differences across countries, even across highly developed countries, e.g. USA and Germany.

Plus large differences over time, as you say. I am not sure how much of that is bias and how much is genuine biology. As a researcher of bias, I am biased to believe that some of the reported high rates may be bias. However, prior vaccination and prior infection may also be important determinants in declining rates. I may well be wrong.

VP: Australia appears to be in the midst of a rise in cases and deaths. Does this reflect on their policy choices? Specifically lockdown?

JPAI: Australia is a tricky case in many regards. The calculation of the final pandemic endpoint, excess deaths, is subject to substantial variability depending on how pre-pandemic levels of mortality are modeled, in order to extrapolate expectations of deaths during 2020-2022.

Australia had a bad year (2019) just before the pandemic years. According to WHO and according to calculations that I published with Michael Levitt and Francesco Zonta, during 2020-2021 Australia had a substantial death deficit, while according to Economist calculations there was a smaller death deficit (if any) during these two years.

Then 2022 saw a steep increase in deaths and, if you take Economist, Australia has >15,000 excess deaths as of mid-July 2022 and steeply rising. If that trajectory continues it may soon catch up with most western European countries especially if one accounts for the fact that Australia has a lower proportion of elderly people (16% over 65, as opposed to 22-23% in Germany and Italy).

So, whatever they did or whatever the virus did, it delayed the massive epidemic waves and gave more time for vaccination. But when the waves hit, practically almost everyone was (or will be) infected. We should not forget that in past pandemics, different regions of the world probably had >50-fold differences on how badly they were hit (although our documentation was far more patchy). It would be superfluous to say that we know much on why that happened – but clearly lockdowns were not in fashion then.

VP: Paxlovid is being widely prescribed in America. The randomized control trial evidence supporting it is limited to a trial of high risk individuals who are also vaccinated. Pfizer has reported no statistical significant improvement in the vaccinated cohort in Epic-SR.  Is this evidence sufficient for mass administration of an antiviral? Might there be downsides?

JPAI: I am not terribly excited about the strength of the evidence regarding Paxlovid. On the other hand, it is nice to know that there are options available. If these options work, they may save lives. I am not convinced that Paxlovid will save many (or any) lives though considering the big, real-world picture. Clearly continuous reappraisal and careful cost-benefit analyses remain a good idea.

VP: The US is moving to an annual COVID-19 shot. Regulators have decided that the available evidence will be merely if the vaccine targets the pre-specified strain. Manufacturers will not be compelled to run randomized trials for clinical outcomes. Do you agree with this? What sorts of studies could we have?

JPAI: I think that public agencies should support the conduct of large randomized trials with hard clinical outcomes for new vaccination schedule questions. RECOVERY and SOLIDARITY were wonderful examples of how one can run efficiently large adaptive trials on drugs and biologics and get pretty reliable answers within a few months. I fully realize that the mass of the randomized evidence will be focused on immunological response (the same has happened with influenza). But this does not preclude running a few pivotal randomized trials with clinical outcomes. We may be (pleasantly or unpleasantly) surprised.

VP: A number of municipalities continue to threaten mask mandates for children. How do you judge the evidence for masking children? Given that zero prevalence is likely 90% or higher, does it make sense? What will be the stopping rules?

JPAI: I think that overall masks are efficacious, although in the real world their effectiveness is often dramatically reduced and can gravitate towards zero. The benefit is likely to be proportional to the risk of severe disease times risk of exposure. For children, the risk of severe disease is the tiniest, even more so now that practically almost all children have already been infected.

The risk of exposure can be substantive during a very active epidemic wave, but then the total benefit is likely to be extremely small, if any. People should take a deep breath, be thankful that children are far less likely to be seriously affected (compared with other age groups), and not strangle each other over mask debates.

We need to tone down excitement, panic, and partisanship, and give room for some normality in education and interaction for children. In most cases, this means no masks for children.

VP: For an adult individual who is vaccinated and boosted, and otherwise healthy, does it make sense to wear an n95?  They will eventually acquire COVID-19.  Is it worth it to delay it, assuming that the mask actually delays it.

JPAI: I think we should tell people what we know about mask effectiveness and leave it at that. There is increasingly a psychological factor in deciding whether to wear a mask or not. Goodness, I feel so confused myself with all these contradicting policies. I go to a scientific meeting and masks are required in the lecture sessions where people sit quietly 5 meters apart from each other, but masks are not requested in the reception where everyone has 50 other people in a radius of 5 meters and none wears a mask.

In one institution I visited recently masks were required on the corridors of the 4th floor, but not in the corridors of the 5th floor, because the 4th and 5th floors belonged to different departments of the same organization and they had different policies. I am struggling to put some logic to all of this mess and I am not able to find much logic. We need to move on with mutual respect for each other and for other people’s weaknesses – so that others may also respect our own weaknesses.

VP: You recently published an article on Long COVID in children. How would you judge the available evidence?

JPAI: Very weak. In some cases, it may be worse than having no evidence at all. We have a lot of extremely poor evidence. More details in the preprint.

VP:How can we process an entity like long COVID which the more we cover in news stories, particularly if they are sensational, the more we may encourage population to report such symptoms. How do we make sense of the nocebo effect?

JPAI: Interesting parallel. The nocebo effect for medications is known to be exacerbated by media storms, when specific adverse effects are highlighted by mass media, then more people say they have these adverse effects. We may have a similar situation here, where media dwell on horror stories surrounding COVID-19 long-term consequences. This does not mean that COVID-19 cannot cause long-term problems. It would be superfluous to make such a claim. But we need very carefully conducted, well-controlled studies to decipher how frequent and how severe these consequences are.

VP: A recent study from the NIH examined a few hundred long COVID patients and compared them to healthy controls who did not have COVID. They sent off a battery of tests, including tests of inflammation, hematologic and immunologic markers, echocardiograms, and PFTs. There were no difference in any objective marker. There was a difference in 6 minute walk distances.  The study was published in the Annals of Internal Medicine. What does the study mean for long COVID?

JPAI: The study is among the best ones – at least they tried to have some decent control. It shows that there are symptoms and they can be frequent, but there is no “objective” diagnostic footprint. This may just mean that the pathophysiology remains elusive and psychological factors cannot be excluded.

VP: Were conflicts of interest salient during the pandemic?  A number of proponents of massive repeated testing are consulting for testing companies. Do we need a sunshine act to unearth the conflicts of the pandemic?

JPAI: I would be very much in favor of a COVID-19 Sunshine act. It would have to include different types of financial arrangements than what is captured in the case of Big Pharma in the Sunshine act.

There were too many powerful players and stakeholders who had a lot to gain and a lot to lose. E.g. companies like Netflix, Amazon, Twitter, Meta and media giants (often entangled in intricate relationships among themselves) mattered as much as or more than Big Pharma in these potential conflicts. It was not any sort of conspiracy, simply powerful players have tons of smart people working for them and these smart people can find ways to defend and champion their powerful interests.

Science and scientists seemed miniscule ants in this war of titans. At times, I feel totally depressed at how people and entities with conflicts created narratives that showed them as heroes while others without conflicts were smeared, simply because they were blocking the way of conflicted agendas. Before the pandemic, people and entities with conflicts tried to hide. With the pandemic, they assumed some moral superiority over those who were not conflicted. People without conflicts now have to hide. It will be difficult to regain the lost moral ground.

VP: Schools were closed for a disproportionately long period of time, and increasingly that looks like a bad decision. You were always against school closure from the outset. What are your thoughts on this issue now?

JPAI: School closures were and are a bad decision. I believe very few people would not acknowledge that this was an error. But I always fear that the temptation to close schools again might appear – and may be followed in some places. The continued Chinese lockdowns make me very worried.

VP: You have published more than 50 peer review articles on COVID-19. What lessons have you learned? How have you kept up such extraordinary productivity?

JPAI: I did my best and I hope I did not pollute the COVID-19 literature too much. As always, I tried to learn from discussing and working with hundreds of great colleagues. Overall, I feel humbled – no amount of humility is enough to convey how often we have been surprised and how little we know in science in general, and this is true par excellence for the COVID-19 pandemic.

VP: Early in the pandemic you were viciously attacked for expressing a policy viewpoint and intuition that is different from others. In a number of ways your core thesis was: in an effort—well intentioned—to control the coronavirus, we may inflict great damage on ourselves. That lesson appears to be borne out in a number of historical and recent events. How do you judge you original intuition?

JPAI: I wish I had been wrong on this point. Unfortunately, I am afraid I might have been correct.

On March 17th, 2020, John wrote an article in STAT where he lamented the lack of empirical data on the coronavirus at the time, and cautioned that in a well intentioned effort to control the virus, we may inflict great damage on ourselves. That article received criticism, often based on uncharitable reading, but it was deeply prescient. John wrote, “One of the bottom lines is that we don’t know how long social distancing measures and lockdowns can be maintained without major consequences to the economy, society, and mental health. Unpredictable evolutions may ensue, including financial crisis, unrest, civil strife, war, and a meltdown of the social fabric.”

Vinay Prasad is Associate Professor of Epidemiology and Biostatistics at UCSF and is also the author of the Substack Vinay Prasad’s Observations and Thoughts.