Responses to “On testing.”

My spouse posted my previous essay on social media, and I’d like to address some of people’s comments. There were some excellent points!

My apologies if I failed to address everything that people said, but I tried my best.

Scroll to find my responses to:

A shutdown could have prevented the Covid-19 epidemic.
We know that the current shutdown is either delaying or preventing deaths due to Covid-19.
Ending this epidemic with a vaccine would be ideal.
Ending the shutdown while requesting that at-risk people continue to self-isolate would save lives.
Why is it urgent to end the shutdown soon?
Why might more people die of Covid-19 just because we are slowing the spread of the virus?
How is the shutdown causing harm?
What about the rate at which people get sick? Isn’t the shutdown worthwhile, despite the risks described above, if it keeps our hospitals from being overwhelmed?
Don’t the antibody tests have a lot of false positives?
What about the political ramifications of ending the shutdown?

1: “A shutdown could have prevented the Covid-19 epidemic.”

If we’d acted early enough, we could have isolated all cases of Covid-19 and prevented this whole debacle.

But we didn’t.

Covid-19 is highly infectious, and we made no effort toward containment or quarantine until the virus was already widespread. We took action in March, but we already had community transmission of Covid-19 by January. Given where we are now, current models predict that the epidemic will continue until the level of immunity reaches somewhere near 70%.

2: “We know that the current shutdown is either delaying or preventing deaths due to Covid-19.”

To date, the data suggests that the virus has only reached saturation inside a few closed environments, such as prisons. In Italy, both the timecourse of mortality and the results of antibody studies suggest that infections were still rising at the time of their lockdown.

Among the passengers of the Diamond Princess cruise ship, deaths peaked 21 days after infections peaked – if the virus had already reached saturation in Italy, we’d expect to see deaths peak sooner than 21 days after the lockdown began. They did not.

So, again, this much is clear: worldwide, there was a significant new cause of death. When we look at mortality data, we see the curves suddenly rise in many locations. Some researchers, such as John Ioannidis, have speculated that Covid-19 causes death primarily in people with low life expectancy, in which case we would expect to see these mortality curves drop to lower-than-average levels after the epidemic ends. But even then, it’s unprecedented to see a number of deaths that would usually occur over the course of a year all within a matter of weeks.

Covid-19 is killing people, and the shutdown is either delaying or preventing people’s death from Covid-19.

For the shutdown to actually prevent death, one of the following needs to happen:

1.) We create a vaccine, allowing our population to reach 70% immunity without as many people contracting the illness.

2.) We take action to change which segment of the population is exposed to the virus, allowing us to reach 70% immunity without as many at-risk people being exposed.

See #3 and #4, below.

3: “Ending this epidemic with a vaccine would be ideal.”

Vaccination is great science. Both my spouse and I love teaching about vaccines, in part because teaching the history of vaccine use is a good component of anti-racist science class.

Developing vaccines often takes a long time. I’ve read predictions of a year or two; my father, an infectious disease doctor, epidemiologist, research physician who runs vaccine trials, and co-developer of Merck’s HPV vaccine, guesses that it will take about five years.

And then, for the vaccine to end this epidemic, enough people will need to choose to be vaccinated that we reach approximately 70% immunity.

The reason it’s worthwhile to compare Covid-19 to seasonal influenza is that a vaccine will only end the epidemic if enough people choose to get it. Many people’s personal risk from Covid-19 is lower than their risk from seasonal influenza. Will those people choose to be vaccinated?

Obviously, I would be thrilled if the answer were “yes.” I’d love to live in a nation where people’s sense of altruism and civic duty compelled them to get vaccinated. My family is up-to-date on all of ours.

But many privileged families in the United States have elected to be freeloaders, declining the (well tested, quite safe) measles vaccine with the expectation that other people’s immunity will keep them safe. And, despite the well-documented dangers of influenza, only 40% of our population gets each year’s influenza vaccine.

Yes, the influenza vaccine tends to be less effective than many others – some years it gives as little as ten percent protection, other years about sixty percent protection. By way of comparison, the HPV vaccine has over 90% efficacy.

A vaccine with low efficacy will still offer better protection when more people get it. If a higher percentage of our population were vaccinated against influenza, then influenza transmission would drop, and so each person’s immunity, whether high or low, would be less likely to be challenged.

Also, the efficacy of influenza vaccines is measured in terms of the likelihood that vaccination prevents infection. The influenza vaccine is not great at keeping people from getting sick. But vaccination also tends to reduce the severity of your illness, even if you do catch influenza. Because you got sick, it seems as though the vaccine “failed,” but your case might have been far more severe if you hadn’t been vaccinated.

The influenza vaccine saves lives. In Italy, where fewer people choose to get vaccinated against influenza (about 15% compared to our 40% of the population), the death rate from influenza is higher. Although it’s worth noting that this comparison is complicated by the fact that our health care system is so bad, with poor people especially having limited access to health care. In the United States, people between the ages of 18 and 49 comprise a higher proportion of influenza deaths than anywhere in Europe. Either our obesity epidemic or limited access to health care is probably to blame; possibly a combination of both.

In summary, for this plan to help us save lives, we will need to develop an effective vaccine, and then people will have to get it.

I am quite confident that we can eventually develop a vaccine against Covid-19. The virus includes a proofreading enzyme, so it should mutate more slowly than most RNA viruses. We don’t know how long it will take, but we can do it.

I am unfortunately pessimistic that people will choose to get the vaccine. And, unfortunately, when a low-risk person chooses to forgo vaccination, they’re not just putting themselves in harm’s way, they are endangering others. Most vaccines elicit a weaker immune response in elderly or immunocompromised recipients – exactly the group most at risk from Covid-19 – which is why we spend so much time harping about herd immunity.

4: “Ending the shutdown while requesting that at-risk people continue to self-isolate would save lives.“

This plan has major downsides, too. Because we didn’t take action soon enough, every plan we have now is bad.

Low-risk people can still die of Covid-19. Even if they don’t die, Covid-19 can cause permanent health effects. Covid-19 reduces your ability to get oxygen to your body and brain. Even a “mild” case can leave your breathing labored for weeks – you’re not getting enough oxygen. Your muscles will ache. Your thoughts will be sluggish.

With a more severe case, people can be looking at heart damage. Renal failure. It would be cruel to look at all these long-term consequences and blithely call them “recovery.”

If our health care system were better, we’d treat people sooner. The earlier you intervene, helping to boost people’s oxygen levels, the better outcome you’ll have. There’s a great editorial from medical doctor Richard Levitan recommending that people monitor their health with a pulse oximeter during this epidemic.

If you notice your oxygen levels declining, get help right away. Early intervention can prevent organ damage. And you’ll be helping everyone else, too – the sooner you intervene, the less medical care you will need.

Because medical debt can derail lives, many people in this country delay treatment as long as possible, hoping that their problems will go away naturally. That’s why people are often so sick when they show up at the ER. I imagine that this is yet another reason – alongside air pollution, food deserts, sleep loss, and persistent stress exacerbated by racism – that poor communities have had such a high proportion of people with severe cases of Covid-19.

And I imagine – although we don’t yet have enough data to know – that financial insecurity caused by the shutdown is making this worse. It’s a rotten situation: you have a segment of population that has to continue working during the shutdown, which means they now have the highest likelihood to be exposed to the virus, and they’re now under more financial strain, which might increase the chance that they’ll delay treatment.

We know that early treatment saves lives, and not everyone is sufficiently privileged to access that.

All this sounds awful. And it is. But, if we took action to shift exposure away from high risk groups, the likelihood that any individual suffers severe consequences is lower.

And there is another caveat with this plan – some people may be at high risk of complications for Covid-19 and not even realize it. In the United States, a lot of people either have type 2 diabetes or are pre-diabetic and don’t yet realize. These people have elevated risk. Both smoking and air pollution elevate risk, but people don’t always know which airborn pollutants they’ve been exposed to. (Which, again, is why it’s particularly awful that our administration is weakening air quality standards during this epidemic.)

Even if we recommended continued self-isolation for only those people who know themselves to have high risk from Covid-19, though, we would be saving lives. The more we can protect people in this group from being exposed to the virus – not just now, but ever – the more lives we will save.

We won’t be able to do this perfectly. It’ll be a logistical nightmare trying to do it at all. People at high risk from Covid-19 needs goods and services just like everybody else. We might have to give daily Covid-19 PCR tests to anyone visiting their homes, like doctors, dentists, and even delivery workers.

At that point, the false negative rate from Covid-19 PCR tests becomes a much bigger problem – currently, these false negatives reduce the quality of our data (but who cares?) and delay treatment (which can be deadly). A false negative that causes inadvertent exposure could cost lives.

Stores will need to set aside morning hours for at-risk shoppers, and undertake rigorous cleaning at night. We know that infectious viral particles can persist for days on a variety of surfaces.

Some people will be unable to work, either because they or a close relative has high risk of Covid-19. Some children will be unable to go to school. We will need a plan to help these people.

We will have to work very hard to keep people safe even after the shutdown ends for some.

But, again, if everyone does the same thing, then the demographics of people infected with Covid-19 will reflect our population demographics. We can save lives by skewing the demographics of the subset of our population that is exposed to Covid-19 to include more low-risk individuals, which will require that we stratify our recommendations by risk (at least as well as we can assess it).

5: “Why is it urgent to end the shutdown soon?“

1.) By delaying Covid-19 deaths, we run to risk of causing more total people to die of Covid-19.

2.) The shutdown itself is causing harm.

See #6 and #7, below.

6: “Why might more people die of Covid-19 just because we are slowing the spread of the virus?“

[EDIT: I wrote a more careful explanation of the takeaways of the Harvard study. That’s here if you would like to take a look!]

This is due to the interplay between duration of immunity and duration of the epidemic. At one point in time, seasonal influenza was a novel zoogenic disease. Human behavior allowed the influenza virus to become a perpetual burden on our species. No one wants for humans to still be dying of Covid-19 in ten or twenty years. (Luckily, because the virus that causes Covid-19 seems to mutate more slowly than influenza, it should be easier to design a single vaccine that protects people.)

In the Harvard model, we can see that there are many scenarios in which a single, finite shutdown leads to more deaths from Covid-19 than if we’d done nothing. Note the scenarios for which the colored cumulative incidence curves (shown on the right) exceed the black line representing how many critical cases we’d have if we had done nothing.

Furthermore, their model does not account for people’s immunity potentially waning over time. Currently, we do not know how long people’s immunity to Covid-19 will last. We won’t know whether people’s immunity will last at least a year until a year from now. There’s no way to test this preemptively.

We’ve seen that immunity to other coronaviruses fades within a year. If immunity to Covid-19 is similar, we really don’t want to prolong the epidemic past a year.

If we could all go into stasis and simply not move for about a month, there’d be no new cases of Covid-19, and this virus would be gone forever. But people still need to eat during the shutdown. Many people are still working. So the virus is still spreading, and we have simply slowed the rate of transmission.

This seems good, because we’re slowing the rate at which people enter the hospital, but it’s actually bad if we’re increasing the number of people who will eventually enter the hospital.

Based on our research with other coronaviruses, we expect that re-infection will cause a person to experience symptoms less severe than their first case of Covid-19. But a re-infected person can still spread the disease to others. And we don’t know what will happen if a person’s risk factors – such as age, smoking status, diabetes status, etc. – have increased in the time since their last infection.

7: “How is the shutdown causing harm?“

If you turn on Fox News, I imagine you’d hear people talking about the damage we’re doing to our economy. They might discuss stock market numbers.

Who gives a shit? In my opinion, you’d have to be pretty callous to think that maintaining the Nasdaq bubble is more important than saving lives.

At the same time, I think you’d have to be pretty callous to not feel extremely concerned by the United Nations’ policy brief, “The impact of Covid-19 on children.”

In this report, they estimate that the shutdown we’ve had so far will cause hundreds of thousands of children to die, many from malnutrition and the other health impacts of poverty. The longer the shutdown continues, the more children will die.

That’s a worldwide number, and most of those children live outside the United States. But I’d like to think that their lives matter, too.

The report also discusses the lifelong harm that will be inflicted on children from five months (or more!) of school closure. Drop-outs, teen pregnancy, drug abuse, recruitment of child soldiers, and the myriad health consequences of low educational attainment.

I live in a wealthy college town, but even here there is a significant population of students who don’t have internet access. Students with special needs aren’t getting the services they deserve. Food insecurity is worse.

You’re lucky that privacy protections prevent me from sharing a story about what can happen to poor kids when all the dentists’ offices are closed. I felt ashamed that this was the best my country had to offer.

As the shutdown continues, domestic violence is rising. We can assume that child abuse is rising, also, but we won’t know until later, when we finally have a chance to save children from it. In the past, levels of child abuse have been correlated with the amount of time that children spend in the presence of their abusers (usually close family), and reporting tends to happen during tense in-person conversations at school.

We know that online sex work has increased during the shutdown. There is an increased supply of sex workers who are experiencing increasing financial insecurity. We don’t yet have data on this, but I’d be shocked if the shutdown hasn’t led many to feel pressured into riskier acts for lower amounts of money, including meeting clients in isolated (and therefore unsafe) spaces.

The shutdown has probably made our drug epidemic worse (and this was already killing about 70,000 people per year in the U.S.). When people are in recovery, one of the best strategies to stay sober is to spend a lot of time working, out of the house, and meeting with a supportive group in communal space. Luckily, many of the people I know who are in recovery have been categorized as essential workers.

But any slip can kill someone recovering from addiction. One of my friends froze to death last year.

A neighbor recently sent me a cartoon suggesting that the biggest harm caused by the shutdown is boredom. (I’m going to include it, below, but don’t worry: I won’t spend too much time rattling sabers with a straw man.) And, for privileged families like mine, it is. We’re safe, we’re healthy, we get to eat. My kids are still learning – we live in a house full of computers and books.

But many of the 75 million children in the United States don’t live in homes like mine, with the privilege we have. Many of our 50 million primary and secondary school students are not still learning academically during the shutdown.

Whether the shutdown is preventing or merely delaying the deaths of people at risk of serious complications from Covid-19, we have to remember that the benefit comes at a cost. What we’ve done already will negatively impact children for the rest of their lives. And the longer this goes on, the more we’re hurting them.

8: “What about the rate at which people get sick? Isn’t the shutdown worthwhile, despite the risks described above, if it keeps our hospitals from being overwhelmed?“

In writing this, I struggled with how best to organize the various responses. I hope it doesn’t seem too ingenuous to address this near the end, because slowing the rate of infection so that our hospitals don’t get overwhelmed is the BEST motivation for the shutdown. More than the hope that a delay will yield a new vaccine, or new therapies to treat severe cases, or even new diagnostics to catch people before they develop severe symptoms, we don’t want to overwhelm our hospitals.

If our physicians have to triage care, more people will die.

And I care a lot about what this epidemic will be like for our physicians. My father is a 67-year-old infectious disease doctor who just finished another week of clinical service treating Covid-19 patients at the low-income hospital in Indianapolis. My brother-in-law is an ER surgeon in Minneapolis. These cities have not yet had anything like the influx of severe cases in New York City – for demographic and environmental reasons, it’s possible they never will. But they might.

Based on the case fatality rate measured elsewhere, I’d estimate that only 10% of the population in Minneapolis has already been infected with Covid-19, so the epidemic may have a long way yet to go.

If we ended the shutdown today for everyone, with no recommendation that at-risk groups continue to isolate and no new measures to protect them, we would see a spike in severe cases.

If we ended the shutdown for low-risk groups, and did a better job of monitoring people’s health to catch Covid-19 at early, more-easily-treatable stages (through either PCR testing or oxygen levels), we can avoid overwhelming hospitals.

And the shutdown itself is contributing toward chaos at hospitals. Despite being on the front lines of this epidemic, ER doctors in Minneapolis have received a 30% pay cut. I imagine my brother-in-law is not the only physician who could no longer afford day care for his children after the pay cut. (Because so many people are delaying care out of fear of Covid-19, hospitals are running out of money.) Precisely when we should be doing everything in our power to make physicians’ lives easier, we’re making things more stressful.

We could end the shutdown without even needing to evoke the horrible trolley-problem-esque calculations of triage. Arguments could be made that even if it led to triage it might be worthwhile to end the shutdown – the increase in mortality would be the percentage of triaged cases that could have survived if they’d been treated, and we as a nation might decide that this number was acceptable to prevent the harms described above – but with a careful plan, we need not come to that.

9: “Don’t the antibody tests have a lot of false positives?“

False positives are a big problem when a signal is small. I happen to like a lot of John Ioannidis’s work – I think his paper “Why Most Published Research Findings Are False” is an important contribution to the literature – but I agree that the Santa Clara study isn’t particularly convincing.

When I read the Santa Clara paper, I nodded and thought “That sounds about right,” but I knew my reaction was most likely confirmation bias at work.

Which is why, in the essay, I mostly discussed antibody studies that found high percentages of the population had been infected with Covid-19, like the study in Germany and the study in the Italian town of Robbio. In these studies, the signal was sufficiently high that false positives aren’t as worrisome.

In Santa Clara, when they reported a 2% infection rate, the real number might’ve been as low as zero. When researchers in Germany reported a 15% infection rate, the real number might’ve been anywhere in the range of 13% to 17% – or perhaps double that, if the particular chips they used had a false negative rate similar to the chips manufactured by Premier Biotech in Minneapolis.

I’m aware that German response to Covid-19 has been far superior to our bungled effort in the United States, but an antibody tests is just a basic ELISA. We’ve been doing these for years.

Luckily for us, we should soon have data from good antibody studies here in the United States. And I think it’s perfectly reasonable to want to see the results of those. I’m not a sociopath – I haven’t gone out and joined the gun-toting protesters.

But we’ll have this data in a matter of weeks, so that’s the time frame we should be talking about here. Not months. Not years. And I’ll be shocked if these antibody studies don’t show widespread past infection and recovery from Covid-19.

10: “What about the political ramifications of ending the shutdown?“

I am, by nature, an extremely cautious person. And I have a really dire fear.

I’m inclined to believe that ending the shutdown is the right thing to do. I’ve tried to explain why. I’ve tried to explain what I think would be the best way to do it.

But also, I’m a scientist. You’re not allowed to be a scientist unless you’re willing to be proven wrong.

So, yes. I might be wrong. New data might indicate that writing this essay was a horrible mistake.

Still, please bear with me for a moment. If ending the shutdown soon turns out to be the correct thing to do, and if only horrible right-wing fanatics have been saying that we should end the shutdown soon, won’t that help our current president get re-elected?

There is a very high probability that his re-election would cause even more deaths than Covid-19.

Failing to address climate change could kill billions. Immigration controls against migrants fleeing war zones could kill millions. Weakened EPA protections could kill hundreds of thousands. Reduced access to health care could kill tens of thousands.

And, yes, there are horrible developments that neither major political party in the United States has talked about, like the risk that our antibiotics stop working, but I think it’s difficult to argue that one political party isn’t more dangerous than the other, here.

I feel pretty confident about all the scientific data I’ve discussed above. Not as confident as I’d like, which would require more data, but pretty confident.

I feel extremely confident that we need to avoid a situation in which the far right takes ownership of an idea that turns out to have been correct. And it’ll be dumb luck, just a bad coincidence. The only “data” they’re looking at are stock market numbers, or maybe the revenue at Trump-owned hotels.

EDIT: I also wrote a more careful explanation of the takeaways of the Harvard study. That’s here if you would like to take a look!

Header image by Goran Paunovic.