How to Examine and Understand COVID-19 Data

John Whyte, MD, MPH; F. Perry Wilson, MD, MSCE


July 14, 2020

Find the latest COVID-19 news and guidance in Medscape's Coronavirus Resource Center.

  • An important number to know is the positivity rate, or the percentage of COVID-19 tests that are positive. If you see the positivity rate of those tests stay stable or increase, it indicates more disease transmission.

  • According to the World Health Organization, if an area is testing enough, the positivity rate should be below 10%. A positivity rate of less than 5% is recommended before an area reopens.

  • There may be a lag between an initial reopening and death rates, because of rolling waves of people who continue to infect others over the next several weeks.

  • Because of false positives, antibody tests are not helpful for most individuals. Instead, antibody tests are useful primarily for public health officials to track COVID-19 progression.

  • The United States is a long way from herd immunity. On average, each person with COVID-19 infects two to three people. The herd immunity goal is 50%-75%, and our country is only at 5% at most.

This transcript has been edited for clarity.

John Whyte, MD, MPH: Hi, everyone. You're watching Coronavirus in Context. I'm Dr John Whyte, chief medical officer at WebMD.

Are you getting confused by all of the COVID-19 numbers, charts, and graphs? To help you understand what's going on, including in your local area, I've asked Dr F. Perry Wilson, associate professor of medicine at Yale, to join me.

Dr Wilson, thanks for taking time this morning.

F. Perry Wilson, MD, MSCE: Good to be with you again.

Whyte: There are a lot of numbers out there. There are a lot of graphs. What do you look to in terms of matrices to get a good handle on what's happening with COVID-19?

Wilson: There are a couple of things. Obviously, the easiest to wrap your head around is the number of cases — people who have tested positive for coronavirus. We've recently exceeded 50,000 cases a day, which is a record. But there's a lot tied up in that. We often wonder whether we're picking up more cases because of more testing. You do want to dig in a little bit more to figure out what's going on. One of the things that I look at is the percentage of tests that come back positive.

Whyte: The positivity rate.

Wilson: Exactly. If you increase the number of tests but there is little increase in disease transmission, then the positivity rate of those tests will go down. On the other hand, if there's more disease transmission, you might see the positivity rate of those tests stay stable or even go up. In fact, that is what we're seeing in large parts of the country right now. We're seeing positivity rates rising despite additional tests, and that really tells you that there is increased transmission.

Whyte: Is there a number that you look at in terms of absolute positivity, meaning just what the number is vs relative? I've seen numbers that have said it should be below 10%. But then I've also seen other numbers that say it should be less than 5%. Or do you simply look at what the trend is?

Wilson: I think the trend is really important to get a sense of which direction we're moving in with the epidemic. The World Health Organization does say that if you are testing enough, then that number should be less than 10%. For example, we had a number of states like Florida and Arizona where the positivity rate was above 20%. I mean, just a crazy amount, suggesting that a lot more testing needs to go on.

The WHO also recommended that your positivity rate should be less than 5% before you have a reopening of society. We're really not there yet, and I think the critical trend is to understand how the pandemic is evolving.

Whyte: Has the accuracy of testing changed? Do we have different tests than we did a month ago?

Wilson: We have a number of different companies making the nasal swab test, the main test that people are familiar with. They've always been relatively accurate when they're positive. We know there's a decent false-negative rate, which hasn't changed very much and might be due to the fact that it's such a hard test to administer. You really have to dig into the nose there. The test hasn't changed too much, but the positives really are positive. We also have these new antibody tests, and the data there are a lot more mixed.

Whyte: We're going to come to that. You mentioned number of cases, the positivity, and the number of deaths. We always talk about how we're 10-14 days behind. I want to point out a data point: For the past 10 weeks in the United States, COVID deaths have been decreasing to less than 600 a day vs a couple thousand in March. Even if we take into account the lag of 10-14 days, how do you factor that deaths have been decreasing for 10 weeks?

Wilson: The lag between infection and death is not exactly how to think about this. When you have a discrete event like a reopening — a date when all of a sudden the restaurants are open — you get these rolling waves of infection. You get a group of people who are initially infected, and then they infect more people.

You can think of this almost as a generational thing, where every several weeks you get a successive generation of infections as the disease spreads. The lag might be longer than the 2-week period, which is the average time between infection and death.

The other thing is that the people infected now tend to be a bit younger than those who were getting infected before. We know the death rate is lower for the younger generation, so we might not see as strong a signal. However, I will point out that death rates are now rising in Arizona, which was one of the first states to show a significant case-related spike. I think we're going to see death rates climb, and that's bad news.

Whyte: This also points to the fact that one should look at data in their local area. Is that correct? Much of the data presented on the news is national, and it may be different in New Haven or Washington vs for someone living in New York or Los Angeles. Would you agree that people should focus on what's happening in their local area, as well as perhaps surrounding counties and cities?

Wilson: There's a bit of a double-edged sword to that. There can be a sense that you are in a safe space if it's not being reported on the news — for example, if you're not in Dallas or Florida. That might lead people to engage in behaviors that aren't as safe, like not social distancing, going to large indoor gatherings, and not wearing masks.

We know there's a lot of transmission that is below the surface, so it's important to know what's going on in your immediate surroundings. It's important to keep your head on a swivel and not get a false sense of complacency.

Whyte: Where do you go to for your data sources?

Wilson: I go to a lot of different places.

Whyte: Give us a couple.

Wilson: Obviously, Johns Hopkins University is the gold-standard reporting source, which has a great graphical interface that you can easily find on the web. The COVID Tracking Project, which is run by The Atlantic, collates testing as well as positivity data from every state in the country and is great if you want to get a sense of those test positivity rates. They also show you hospitalization rates and death rates on a state-by-state level. So those are some really good resources.

Whyte: Let's talk about antibody testing. When we talked a couple of months ago, there was a different perspective on antibody testing. People thought we were going to be able to create a shield of immunity based on a test or have "passports" based on antibody status. Where are we today on the concept of antibody testing and its usefulness for an individual?

Wilson: Antibody testing has been a real trouble spot. Just to remind people, antibody testing is the idea and hope that you're past the infection; the infection's over and you confirm that you had it by showing you have these antibodies. What we've learned recently is that many of these tests have false-positive rates. They'll give you an impression that you might be protected when you're not.

We also have learned that there's a good number of people who, despite testing positive for antibodies in general, might not have protective antibodies, the kind that can prevent infection. So no, there are no antibody or immunity passports coming anytime soon. The place for antibody testing is for surveillance of how the disease is progressing.

Whyte: For public health officials primarily.

Wilson: Exactly. As broadly as possible, for public health surveys to see what percentage of the population in a discrete area has been exposed, because that helps so much in modeling how the disease can spread.

Whyte: What about if you're just curious? Is it worthwhile to get an antibody test? Is there any harm?

Wilson: To be honest, I think there is harm, because if you get a false-positive rate, you might be falsely reassured. Even if you do have the antibodies, the test does not tell you if they're the protective kind of antibodies.

If you had a classic COVID-like illness (severe fever and respiratory distress), a positive COVID nasal swab test, and the antibody test comes back positive, it might reassure you that you've been through this. But I think for the vast majority of people, it's not a good option to take an antibody test just to find out.

Whyte: Where are we on herd immunity? Classically, we think it's 70% to be effective. Other studies have recently suggested that it could be as low as 40% to have the same protection. There have been some preliminary data that show in New York City alone that herd immunity may be 20%-25%.

Wilson: It's a simple concept to figure out what the target percentage is. To give an example: If every person with COVID infects an average of two other people, and if 50% of people are immune, it means they can only infect one other person and we're flat, right? So the herd immunity percentage is directly related to how much on average someone infects someone else.

Because an average person with COVID infects about two to three people, we think that herd immunity would occur between 50%, 60%, to 75% of the population. This is the bad news: We are nowhere close to that. The Centers for Disease Control and Prevention seroprevalence data suggest that about 5% of the country has been exposed to coronavirus. This is about 10 times higher than the number of cases we've documented. It's only one twelfth the number of cases we need to think that herd immunity alone will stem the tide of the pandemic. Without taking other precautions, we're not even into the second inning of this pandemic and there have already been [133,000] deaths.

In my opinion, waiting for herd immunity to save us is an ineffective and a potentially disastrous approach.

Whyte: What's your response to the argument that if the number of cases is actually 10 times what is reported but the number of deaths is largely the same — although there are some data that recently came out that say it could be an underestimate — what does that say about the lethalness of COVID-19? Do we need to rethink that?

Wilson: We've been clear all along that the observed fatality rate is higher. We've always known that there are cases that aren't coming to our attention. I'll point out that the observed fatality rate in the United States right now is 5%. If that were the true fatality rate, it would be a disaster of apocalyptic proportions.

Even at 0.5%, it's still pretty lethal, especially when compared with other respiratory viruses like the flu. You don't need math to tell you that. You can just look at the fact that over [133,000] people have died already. We've blown every prior flu season, including the swine flu season, completely out of the water. So this is not benign, regardless of how many cases have gotten below the surface.

Whyte: Dr Wilson, I want to thank you for taking the time today to help us understand all these different numbers. And I hope we can come back to you in another month or so and check in and see where we are on COVID-19.

Wilson: Anytime. I hope we have some better news in a month.

Whyte: And thank you for watching Coronavirus in Context.

Follow Medscape on Facebook, Twitter, Instagram, and YouTube


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.