Positivity vs. Prevalence: Are You Sure You Know the Difference?

The two terms are easy to mix up, but one is far more useful for judging risk

February 01, 2021

Since the start of the COVID-19 pandemic, we’ve all had to become armchair epidemiologists. Sometimes it feels like you need a Master’s degree just to figure out whether you can go to the grocery store or to visit your family over the holidays. How risky is this, really? How much riskier would it be to take the bus versus walking? How many risk points would be deducted if I’m careful to wear a mask and keep my distance? Everyday life has become a series of calculations that we’re forced to make—whether or not we feel qualified to make them.

For instance, a resident of Mobile, Alabama might get up this morning, make some coffee, and be confronted by today’s terrifying statistic. Alabama’s percent positive rate is a whopping 34.1%! But what does that mean, exactly?

You would be forgiven for assuming that means about one in three Alabamans have COVID-19. But, luckily for Alabamans, that’s not exactly right. The confusion stems from the difference between two terms: prevalence and positivity.

Prevalence means the proportion of people who are infected at a given time. So if the prevalence were 34.1%, that really would mean that about one in three Alabamans have COVID-19. The problem is, unless you test every single person in the state, you can’t really know the true prevalence. The only thing we can say for sure is: of the people who were actually tested, how many were positive? That’s percent positivity.

You can see how that number could be potentially misleading. Many people get tested because they already have a good reason to suspect they have COVID-19—they have symptoms or came in contact with someone who recently tested positive. If we only test those people, the percent of those tests that come up positive will likely be much higher than if we tested everyone at random.

So how can we learn the actual prevalence in a community? For that you need to get a baseline. But what’s a baseline? It’s when you test a large group, regardless of symptoms, to get a snapshot of community prevalence at one point in time.

In an ideal world, we’d design ongoing testing strategies that make prevalence and positivity match as closely as possible. That’s because whether you’re trying to decide if you can visit family over the holidays or if you can open restaurants at full capacity in your state, prevalence is the number you actually need to know. 

How do epidemiologists get a sense of prevalence?

There are two important things to look for when trying to find prevalence:

First: more data is better when it comes to getting closer to true prevalence. The more data you have, the less likely the number you get will be influenced by random chance. It’s like the difference between flipping a coin once or twice and flipping it a hundred times. If you only flip it twice, it could easily come up heads both times. If that’s all you know, you could conclude that flipping a coin gives you heads 100% of the time, which is obviously wrong. The more times you flip the coin, the closer and closer you’d get to the right answer.

Second: the more random data, the better. As Daniel Westreich, an epidemiologist at the University of North Carolina, remarked last year, “We just haven’t tested enough people yet. If you were doing random screening of the whole population, we just don’t know what you’d see.” If we really want to know what proportion of people are positive in a state, we can’t just rely on those who go to the doctor to get tested. We also need to test regardless of symptoms or known exposure, as was the case at colleges with strong testing programs last semester. 

An additional benefit of that strategy is it gave them information about local prevalence. That’s important because the prevalence in a particular community might be radically different from the state-wide average—especially at schools that go to great lengths to prevent outbreaks.

Staying Positive

To look at the positive side of all of us becoming armchair epidemiologists, we’re gaining a lot of practical knowledge about what works and what doesn’t to contain an epidemic. We all hope it will be a long time before what we’re going through happens again. But if there’s anything 2020 has taught us, it’s that once-in-a-lifetime events have a way of happening back-to-back. We should take this chance to learn from our mistakes, which means investing in public health and building our national testing capacity. We have an opportunity to make sure we’re not caught unaware again. Let’s take it.