Coronavirus #5: Why Opening Up Now Is a Mistake
From the very beginning of the COVID shutdown, it was clear that people would not stay in isolation as long as the science dictated. They were going to stay inside until they got tired of it.
It’s human nature. We like to think we make decisions based on reason, but this is rarely the case. We make decisions based on our gut emotions, and then we make up reasons to support what our gut wants to do. In the case of COVID-19, the rising voices of people who want to open up are coming from those who are concerned about the economy, and who also are simply tired of staying at home.
It is reasonable to want to get back to normal life. And there is nothing wrong with wanting to go out and earn a living. But the problem is, these are emotional decisions, and they have nothing to do with science. While many are weary of staying at home, the science is unable to say how far away from safely opening up we are.
The great scientific problem with COVID is that we do not know its prevalence. Prevalence is a statistical term that refers to the number of active cases of a disease in a given population at a moment in time. If we could take a massive photograph that included every person in the United States right now, and then take a black marker and put an X over every person in the picture who has coronavirus, the number of Xs would be the prevalence of coronavirus for the whole country.
But we do not have this information right now. And without knowing the prevalence of coronavirus, we cannot make an informed decision about when the economy can open up again.
Why don’t we know the prevalence? Because we don’t have enough tests, for one thing. Since we haven’t tested everyone in America, we don’t know for sure who has it. For another thing, COVID is an insidious disease. People can have it for as long as 10 days and be asymptomatic, meaning infected individuals can spread the disease without realizing they need to be tested.
Why is prevalence so important? Prevalence is the way doctors manage almost all diseases. When we know the prevalence of a disease, it is much easier to devise a strategy to deal with it. For example, if I am in a nursing home visiting someone and a man down the hall yells out that he is experiencing crushing chest pain, my first thought is that he is having a heart attack. Heart disease is very, very common in people over 65, and any elderly person who complains of chest pain should be considered to be having a heart attack until proven otherwise. On the other hand, if I am at an elementary school and a sixth grader complains of crushing chest pain, I am definitely not thinking heart attack. An anxiety attack, a piece of food stuck in the throat, heartburn, asthma, and a broken rib are all far more common in a pre-teen than coronary artery disease. My approach to a patient with chest pain depends on the prevalence of heart disease in the population I am working with.
Prevalence tells us what diseases to consider first when devising a treatment. Knowing prevalence saves a ton of diagnostic time, and it makes it easier to decide which patients need hospitalization and which can be sent home and observed. In the case of COVID, not knowing how often the virus appears in the population makes these decisions much more difficult. A person with shortness of breath in a doctor’s office might have coronavirus, or might not. The odds are heavily dependent on prevalence. With knowledge of a very low prevalence, the doctor could decide to send the patient home. Without it, she is probably stuck telling the patient to go the emergency room.
Other decisions, like when to open schools and shopping malls also depend heavily on prevalence. If one in a hundred people has COVID, it is probably too dangerous to open a movie theater. But a small restaurant with well-spaced seating might be just fine. If one in five people have it, even a small coffee shop with takeout only could be a risk.
Since we don’t know how many people have the disease, even talk about the death rate is complicated. Where I live, in Mississippi, as of this writing about 8,400 people have tested positive, and about 370 people have died. So if you take the deaths and divide them by the number of positive tests — 370/8400 — you get 4.4%. A 4.4% death rate is alarming, and implies that if all 3 million Mississippians got coronavirus, there could be over 120,000 deaths — more than we have had in the entire U.S. to date. But concerning as this appears, it is probably not an accurate picture. Many more thousands in Mississippi could have COVID and have not been tested. Most of these people will live, but some of them will die and may not be counted as COVID deaths. The real death rate in Mississippi could be lower than 4.4%, or it could be somewhat higher. Many people are guessing that the rate is lower, assuming that most of the truly sick people are coming to hospitals and being tested. But the problem is, we don’t know. Assume is a dangerous word in medicine. Any attempt to put the death rate higher or lower than 4.4% is nothing more than guesswork.
So what is the answer? Test, test, test. And if we don’t have the resources to test everyone, we need to do the next best thing — randomized testing. Testing the general population in a random manner should give us a rough idea of the prevalence of COVID. Doing so will require millions of tests nationwide, and the tests must be done in a short period of time, the shorter the better, because viral infection is a moving target. New infections occur every day, and older infections resolve and the patients return to normal again. To get a prevalence, you need a cross section of the population. Remember our giant photo? We need a photograph that picks up everyone at a set moment in time.
This leads us to another reason why the 4.4% death rate doesn’t tell us much. The death rate is an average over a period of many weeks, which is not what you want when you are trying to estimate prevalence. What we need to know is how many people are infected and can spread the virus right now. The overall death rate is an indirect indication of how many people may have had COVID since the pandemic began. While this is useful information, it is not a measure of prevalence.
Once again, in summary: In science, you cannot make assumptions when you don’t have enough data. Although it is common for people to make decisions in their own lives without complete information, it is not acceptable science. It is dumb for us to assume we know how many people will die from COVID, or if we are safe from it, without accurate prevalence data. That’s all there is to it.
If we don’t have the data, it is dangerous to just assume. Would you consent to have surgery if your surgeon told you, “This is a brand new procedure, and we are all curious to see if it works”? Would you take a medicine if the pharmacist told you he was “kind of sure” it would work, and even if it didn’t, he was “reasonably confident” there were no fatal side effects? No, you wouldn’t. You expect better from science. Science is, in fact, better than guesswork. That’s why we prefer science over rolling the dice when we build planes, manufacture medications, and design nuclear power plants.
This is a hard place to be in. People are tired of isolation, tired of losing money, sick of worrying about their parents and grandparents. They want things to be normal again. But we don’t have the facts. This, unfortunately, is where we are. We don’t know enough about this virus to know what the outcome will be if we loosen restrictions. We know for certain that people cannot catch the virus in isolation, but we don’t know how many infected people a person can expect to meet at the supermarket, at an office, or at a church picnic.
The best advice I and the medical establishment have right now is to continue to stay at home. If this is frustrating, and you want to blame someone, blame your government for not being able to make enough tests available to determine the prevalence of COVID. Although testing is complicated and difficult, I cannot help but think that if every politician had put 100% of his or her effort into solving this problem back in February we would be much further down this road than we are. That is no doctor’s fault — that is the fault of elected officials who were more concerned about how much COVID was going to cost stock market investors than about identifying the most important obstacle to fighting it, which was always testing, testing, testing.
To my knowledge, there is not a single state in America that knows the prevalence of COVID in its population. Some are closer than others — New York and California appear to have tested the most people and Rhode Island has tested the highest percentage of citizens — but none have tested enough for medical experts to clearly assess the risk of catching the disease.
You want to leave home? Make your government do its job and test enough people to fully assess community prevalence. This can be done through sheer numbers of tests or a carefully run randomized approach — either will do. But any attempt to open things up without knowing how many contagious people are out there is guesswork.
Don’t let politicians or angry citizens who are tired of staying at home gamble with your life. Because if you let them, they will.
