[Editor’s note: Filling in our empty Friday slot is a guest post from Dr. Mark Burkard. Welcome!]
By all accounts the COVID-19 pandemic is spiraling out of control, particularly in the U.S. where daily cases top 150,000. The current patchwork of public health policies is insufficient to control the rapid rise, and many health systems and workers across the country are becoming rapidly overwhelmed.
On November 11 Dr. Howard Bauchner (the editor-in-chief of The Journal of the American Medical Association/JAMA, one of the preeminent U.S. medical journals) hosted a livestream video/podcast as part of a series, Conversations with Dr. Bauchner. The episode was titled “Herd Immunity as a Coronavirus Pandemic Strategy” and featured Jay Bhattacharya MD PhD (Stanford Professor of Medicine in Primary Care Outcomes and an expert in health economics), and Marc Lipsitch PhD (Harvard TH Chan School of Public Health, Professor in Epidemiology). The discussion was prompted by the conflicting views of Dr. Bhattacharya and Dr. Lipsitch.
Dr. Bhattacharya was an author and signatory of the Great Barrington Declaration (GBD), promoted by the libertarian Institute of Economic Research, previously covered on SBM. The GBD proposes ‘focused protection’, described as allowing “those who are at minimal risk of death to live their lives normally to build up immunity to the virus through natural infection, while better protecting those who are at highest risk”. At the same time, it proposes the vulnerable should be protected, for example, in nursing homes by using “staff with acquired immunity and perform frequent PCR testing of other staff and all visitors”. Although it proposes focused protection of the vulnerable, it also advocates allowing the vulnerable the freedom to risk their lives: “people who are more at risk may participate if they wish, while society as a whole enjoys the protection conferred upon the vulnerable by those who have built up herd immunity”.
This proposal was met with disdain by most epidemiologists and public health authorities as impractical and dangerous. WHO Director-General Tedros Adhanom Ghebreyesus characterized the proposal as “scientifically and ethically problematic”. Anthony Fauci, director of NIAID, described the proposal as “ridiculous” and “total nonsense”. Francis Collins, NIH Director, described the idea as a fringe view:
“What I worry about with this is it’s being presented as if it’s a major alternative view that’s held by large numbers of experts in the scientific community. That is not true,” National Institutes of Health Director Francis Collins told the Washington Post. “This is a fringe component of epidemiology. This is not mainstream science. It’s dangerous. It fits into the political views of certain parts of our confused political establishment,” he told the outlet.
Three Harvard medical students succinctly laid out the central problems with the proposal.
Yet the U.S. Government – which appears to be schizophrenically following this policy – held high-level meetings with the GBD authors, leading to promulgation of the idea and prompting mainstream epidemiologists to pen a counter-statement, the John Snow Memorandum. This document supports the prevailing consensus from public health authorities worldwide, that public health policies should seek to minimize community spread until a safe and effective vaccine is available, which will thereby minimize the exposure of high-risk populations. Despite the attempt to set the record straight, the John Snow Memorandum inadvertently created a false equivalence between the fringe approach and what is actually recommended by epidemiologists and public health officials.
Unfortunately, JAMA further bolstered this false equivalence through a mainstream-medical platform by publicly airing a joint discussion with Dr. Bhattacharya and Dr. Lipsitch for their countering points of view. Unfortunately, debates are poor vehicles for honest scientific discussions, as it is possible to misstate, mislead, and misdirect. Sure enough, Dr. Bhattacharya appeared to use these tactics to support ‘focused protection’, while the platform created the false impression that the GBD proposal is as legitimate alternative to public health efforts that seek to limit community spread.
Understated and misleading risks
In order for the focused protection proposed in GBD to be feasible, two things need to be done. First, it must be possible to easily identify low- versus high-risk individuals. Second, it needs to be possible to isolate high-risk individuals to enforce ‘focused protection’ in the setting of a highly infectious virus that is widespread in the low-risk individuals. Critics have pointed out that neither of these are feasible. How did Dr. Bhattacharya address this?
In his opening statement, Dr. Bhattacharya says that COVID-19 is deadly for “people who are older and for people who have certain chronic conditions”. He cites the survival as 95% for age 70 and older and 99.95% for those under 70. (It is clear he is referring to the infection fatality rate, or IFR, as he refers to seroprevalance studies). In general, the IFR for a disease is lower than the case-fatality rate, although both are difficult to estimate in real-time during an ongoing pandemic.
However, no estimates of risk sharply discriminate those 70 and older versus under age 70. For example, the CDC uses a range of estimates of the IFR for its planning scenarios, varying from 0.25-1% for age 50-69, and 2.8-9.3% for 70 years and older. Dr. Bhattacharya, it seems, has averaged IFR rates for younger age groups including 0-19 years and 20-49 years with those up to 69 years old in order to reach an estimated IFR of 0.05% under age 70. It seems unfair to average across all age groups and apply it to the 50-69 year old group, but this misleading argument is, in effect, necessary to claim that it is easy to discriminate high- versus low-risk individuals. Further, it is well known (and was acknowledged) that other medical conditions such as obesity and diabetes increase risk in younger individuals; it wasn’t clear whether these individuals were excluded from his ‘under 70 group’ but if so, again it is misleading to claim 99.95% survival for those under 70. This was not an auspicious start to the JAMA Conversation.
Even averaging the very young, the estimate of IFR of 0.05% for those < aged 70 doesn’t appear to fit with scientific sources. For example, the CDC estimates of the IFR are based on Hauser et al. Of note, one of the conclusions of Hauser et al. is that the IFR has major geographical differences that could arise from factors “including emergency preparedness and response and health service capacity”. When health systems are overwhelmed – as they are becoming in the U.S. today – the IFR is expected to escalate. Thus, it is odd that Dr. Bhattacharya can be so certain that there is 99.95% survival for those under 70.
A study recently published in The Lancet estimated the IFR of COVID-19 in spring 2020 in New York City. This study broke down IFR by age but the measures were much higher than those claimed by Dr. Bhattacharya – 14.2% aged 75 and older, versus 4.87% for aged 64-74, versus 0.939% aged 45-64. These are the results for all age ranges:
Here, the estimated IFR here is higher than 0.05% for all age groups except <25 years. It is hard to imagine that there are enough people <25 in this population such that the average for all those <70 years old would come out anywhere close to the 0.05% claimed by Dr. Bhattacharya, but even if that were the case, it would be very misleading to claim that everyone under 70 years old has very low risk, as clearly those close to 70 have a significantly higher risk. Perhaps Dr. Bhattacharya would argue that the data from NYC in the spring of 2020 are high estimates during a strained health crisis, and so the IFR would be lower now, given that health systems are not currently overwhelmed and that treatments have improved. Nevertheless, the claims of a 0.05% IFR for those aged less than 70 are at best misleading and at worst wrong.
So right in his opening statement, Dr. Bhattacharya has done three things. First, he has claimed a sharp discrimination of low- versus high-risk groups at age 70, which is misleading. Second, he has underestimated risk for people under age 70. Third, he inaccurately represents the IFR as a stable and well-known quantity. If it is necessary to mislead, misrepresent sources, and have greater certainty than warranted to support a point of view, perhaps that view should be reconsidered.
COVID-19 is about more than just deaths
We also haven’t discussed morbidity yet, the suffering caused by a disease that falls short of deaths. Dr. Bhattacharya dismisses concerns about the morbidity of COVID-19. The short-term morbidity for people who may not die includes hospitalizations and their complications. The long-term morbidity of COVID-19 is not yet clear. The topic of long haulers was brought up and then dismissed as having unknown incidence and duration. However, substantial long-term morbidity is at least possible as the SARS-Cov2 virus enters pulmonary and brain tissue, and causes cardiac dysfunction. Although the prevalence and extent of the long-term consequences are unclear, the clear short-term and probable long-term morbidity in a substantial fraction of infected individuals should give everyone pause. Dr. Bhattacharya appears completely dismissive of COVID-19 morbidity, and appears to argue that unknown risks should not impact policies, even when that risk will be incurred by tens of millions.
Dr. Bhattacharya overstates harms of lockdowns
Dr. Bhattacharya claims “Lockdowns have absolutely catastrophic effects on physical and mental health of populations” and “For people under 70 (or 60) the lockdown harms are worse than COVID-19”. No evidence is provided here, but the evidence that does exist suggests this is inaccurate or highly misleading.
If we are to compare mortality of lockdowns with the mortality of COVID-19, there is no contest – COVID-19 is far worse. For example, mortality in the U.S. tracks with the surges of COVID-19, and at least 2/3 of the excess deaths are directly attributable to COVID-19. As an independent way of determining deaths from COVID-19 versus deaths from lockdowns, we could observe death rates in a location with a severe lockdown but relatively few cases – and we can make such an observation based on the experiences of Melbourne, Australia. From August through October in 2020, the city and surrounding state of Victoria had one of the most severe lockdowns in the world as they attempted to squelch a small outbreak before it exploded into the population. However, the severe lockdown was without a significant change in deaths compared to the previous year.
Critics of lockdowns claim they increase rates of suicide. Dr. Bhattacharya says “The CDC estimated in June that one in four young adults seriously considered suicide. One in four. You know, normally that’s something that’s on the order of 4% is now one in four”. This is misleading summary of the CDC MMWR report – the 4% appears to be the 2018 suicide ideation rate in all age groups, which increased to 10.7% – not 25% – in June 2020. The 25% cited apparently refers to the 18-24yo subgroup, which in 2018 was measured at 11.0% [pdf, Figure 58, p48]. Thus, the actual measured increase is 2.3-2.7 fold from 2018 baseline, much less than 6-fold escalation claimed by Dr. Bhattacharya. These data need to be interpreted with care due to the acknowledged methodologic differences in the 2018 and 2020 surveys. Nevertheless, the apparent increase in mental health impacts of the pandemic are of serious concern. But what caused this? June 2020 was notable for escalating COVID-19 cases and unrest across the country after prominent examples of racial injustice. In fact, stay-at-home orders had largely ended across the U.S. by June 2020. Thus Dr. Bhattacharya mis-cites the CDC report to claim a larger escalation in suicidal ideation than the data showed, and uncritically attributed the findings to lockdowns that largely ended prior to the survey timeframe. I am not arguing that we should ignore mental health issues – addressing them and escalating awareness can prevent adverse collateral impacts of the pandemic. However, Dr. Bhattacharya claimed the impact was worse than the data showed and attributed the outcome to lockdowns with undue certainty.
However, these data are about suicidal ideation; what is the impact of lockdown on actual suicide rates? Fortunately, during the pandemic there is no clear evidence that rates have been affected at all. For example, in the Victoria lockdown noted above, there was no reported increase in suicide rate [PDF]. In short, there is no evidence of higher suicide incidence due to lockdowns.
Another claimed mortality risk of lockdowns is the adverse impact on medical care. Dr. Bhattacharya said, “People stayed home. They’re more afraid of COVID than getting treated for heart attacks. That’s happened already with the lockdowns”. Again, Dr. Bhattacharya attributes failure to seek emergency care during the pandemic solely to lockdowns. However, hospitals were never closed to emergency care. People who failed to seek emergency care during lockdowns did so due to concerns about COVID-19 exposure and overwhelmed hospitals. These are likely to occur – and will be worse – if we permit widespread community spread of COVID-19 among the less vulnerable as he has proposed.
Morbidity of lockdowns is certainly likely. It can be distressing to restrict social activities, to mandate school closures, to cancel sports. However, the controversy stoked by the GBD is likely to fuel further restrictions and isolation for the many high-risk individuals who cannot safely enter community spaces. By proposing people ‘live their lives normally’, the number of cases and hospitalizations will definitely escalate even if the death rate remains low. They will particularly escalate in that the GBD proposes that high-risk individuals are allowed to take their chances if they wish to. The serious medical impacts of COVID-19 also contribute to distress. The serious medical issues, hospitalizations, visitor restrictions at nursing homes, all contribute to distress. It is not clear that lockdowns cause more morbidity and social disorder than an uncontrolled epidemic or the ‘focused protection’ that would disrupt millions of U.S. households.
In conclusion, Dr. Bhattacharya overestimates the speculative mortality and morbidity of lockdowns, while minimizing the known mortality of an escalating pandemic. Worse, he ignores that the pandemic is infectious – and therefore grows exponentially. Even if the mortality and morbidity of COVID-19 and lockdowns were, say, equal, the potential of future harm scales differently. Essentially, the harms of lockdown would scale linearly over time, but COVID-19 in a susceptible population grows exponentially.
Dr Bhattacharya fails to provide viable approach for ‘focused protection’
Dr. Bhattacharya claims that the vulnerable can be protected by ‘focused protection’ without controlling community spread. When asked how it would be logistically possible to segregate the vulnerable from the low risk in a country of 360 million, Dr. Bhattacharya again overstates the risks of and imperfections of lockdowns, which he claims have failed. Ironically, he says “any public health intervention needs to take into account the environment it actually is going to be implemented in, not pretend like we’re going to have it work perfectly”. So what is his more realistic proposal?
First, he describes enhancing protections at nursing homes: deploying rapid antigen tests of visitors, personal protective equipment, segregating those who are ill, and limiting staff rotations. Nursing homes have already implement many of these policies and due to inadequate testing, often simply disallow visitors. Nevertheless, many nursing homes have had major outbreaks due to the high transmissibility and prevalence of COVID-19, and to the presence of asymptomatic and presymptomatic individuals. There are an estimated 15,600 nursing homes in the U.S. What would it take to protect them all? Consider that rapid antigen testing of visitors has limited sensitivity and has thus failed in arguably the most highly guarded building in the U.S. – the White House – where it infected the President, aged 74. But Dr. Bhattacharya believes this can be done in all 15,600 nursing homes, without substantially new policies that would materially increase protection, in the setting of wide community spread of COVID-19.
But nursing homes are just one piece of the puzzle. An estimated 20% of U.S. households and 64 million people live in multigenerational homes. How does Dr. Bhattacharya propose to protect elders in this population? Well first, he blames lockdowns for creating multigenerational homes (strangely ignoring that the risk is mitigated with lower community contact). But after this tangent, he proposes creating alternate living arrangements, like hotels for the homeless, when the younger group are infected. Of course, this assumes they know they are infected which requires symptoms or frequent asymptomatic testing which has cost, turnaround-time, and sensitivity issues. It would seem to require a rapid-turnaround, 100%-sensitive test that is done in millions of households almost daily. The vague proposal appears unclear, impractical, and ignores the sensitivity issues that led to the White House COVID-19 breach despite routine rapid antigen testing of visitors.
How would Dr. Bhattacharya propose to deal with high-risk workers, such as those with diabetes? He says that a 63-year-old diabetic Costco clerk should have reasonable accommodation to work from home. He wants to give disability insurance to many. I agree – lowering the exposure of the working vulnerable and providing social support is admirable. However, such significant social support is not available in the U.S. today. (Also, it is still unclear who he believes is at high risk, since we are now talking about people under 70 among whom he estimated the IFR of 0.05%).
Tragedy of the commons and libertarian viewpoint
As noted above, a pandemic spiraling out of control will exhaust the common medical resources, impacting all patients. Even if a small fraction of the infected die of COVID-19, the medical needs of the population will not be met adequately in the face of a COVID-19 surge. Others will die needlessly of cancer, heart attacks, strokes, trauma, and the inability to get care. Overwhelmed health systems can delay screenings and elective surgeries first, but this has its own impacts on future risk and quality of life. Even worse, the GBD proposes to allow high-risk individuals to ignore the limited ‘focused protection’ measures as it says “people who are more at risk may participate if they wish”. Since these groups will need more medical care, this policy would further serve to overwhelm the health care systems.
It is hard not to see Dr. Bhattacharya’s ideas as a fringe libertarian view, where the tragedy of the commons is not considered. First, Dr. Bhattacharya was a research fellow at the Hoover Institution, Stanford University’s conservative, at least partly libertarian, think tank. Second, Dr. Bhattacharya was lead investigator of the infamous Stanford seroprevalence study preprint, which erroneously overestimated the prevalence resulting in a minimized IFR estimate. Third, Dr. Bhattacharya followed up the seroprevalence preprint with an article in The Wall Street Journal, which used the conclusions to publicly disseminate his idea that risks of COVID-19 were overblown. Fourth, Dr. Bhattacharya went to the American Institute for Economic Research, a libertarian think tank in Great Barrington, MA where he co-authored the GBD. It appears that there may be a point of view guiding efforts to muster supportive data, rather than an effort to use data to guide policy.
Dr. Bhattacharya uses ambiguous definitions
At the end of the discussion, Dr. Bhattacharya and Dr. Lipsitch are asked what they mean by lockdown. Dr. Lipsitch indicates that the Stay-at-Home orders in many parts of the U.S. between March-May 2020 constituted lockdowns. Dr. Bhattacharya disagreed claiming that those were quarantines and that any restriction on activities – churches, restaurants, schools, arts – constitute lockdowns. However, this ambiguity serves mainly to obfuscate, since idiosyncratic definitions are not generally understood. Quarantine, in particular, has a medical definition in which individuals are isolated for a period after being known to have, or suspected of exposure to, an infectious disease. Lockdown is defined as confining individuals to a place due to an emergency, and is understood to be equivalent to stay-at-home orders. The ambiguity employed makes it difficult for interlocutors to pick up on the imprecision of applying harms measured during stay-at-home policies to the less-restrictive shutdown of high-risk public activities in parts of the country experiencing high COVID-19 activity. If one needs to ambiguate with idiosyncratic definitions to support a point of view, it is probably best to reconsider the point of view.
Real-world impact of the false and misleading claims from Stanford academics
Dr. Bhattacharya and the other GBD authors have failed to win over their scientific colleagues, infectious disease experts, and public health authorities. Because of a number of factual inaccuracies, false assumptions, ambiguity, and dismissal of practical matters discussed above, the ‘focused protection’ proposal is not considered serious by the bulk of knowledgeable experts who have examined it.
What, then, is the impact of publicly airing fringe views? If Dr. Bhattacharya and colleagues aired their views in scientific forums and peer-reviewed publications, these views could be vetted, examined, considered, criticized, and revised. Ultimately, portions of the view that are incorrect or unsupportable could be rejected, while accurate and promising facets – such as opening schools for young-aged children, or policies to keep at-risk workers safe – would gain traction and acceptance, influencing a cohesive public health policy. However, Dr. Bhattacharya and colleagues have sought the widest public dissemination of a wholesale fringe idea that does not convince the bulk of relevant experts. As above, the ideas appear to be based on misleading or inaccurate premises without a serious attempt to address the criticisms or practicality of ‘focused protection’. Unfortunately, this widespread promulgation of the GBD and related ideas has influenced the public, and allowed individuals to rationalize decisions to ignore public health rules and recommendations. Medical authorities such as Dr. Bhattacharya and JAMA publicly airing the GBD proposal undermines a united public health approach, and is potentially lethal – the public health equivalent of shouting ‘fire’ in a crowded theater.
Conclusion: A deadly false equivalence
JAMA has promoted the false equivalence of a fringe view and placed it publicly on social media. In a world where even Facebook and Twitter now fact check claims that impact public health, JAMA oddly failed to do so. As a result, a misleading and misplaced set of claims could lead many in the public to oppose or ignore public health policies, worsening the severity and impact of the pandemic. While it is of value to vigorously debate policies and interpretation of scientific evidence, JAMA has a responsibility to accurately portray scientific information on social media, to base discussions on facts and evidence, and to avoid allowing its platform to promulgate a false equivalence, particularly when doing so may adversely impact public health.