Calculate Your School’s Estimated Prevalence
May 03 2021
A little while ago we told you about the difference between two terms that are very easy to confuse: positivity and prevalence. Since then, we’ve been fortunate to provide testing to hundreds of schools across the country, giving them the information they need to help reduce the spread of COVID-19. But, in doing so, we’ve introduced another number… pool positivity.
A major barrier to testing in K-12 schools is cost. Some COVID-19 tests cost upwards of $100 per person. Multiply that by tens of millions of students, plus teachers and staff, once or even twice a week, and you’ll quickly get to an astronomical sum. One solution to that problem is a strategy called pooled testing.
Reminder: In pooled testing, groups of people swab their noses and all of those swabs are tested as one pooled sample. If the group’s test is negative, you can assume that everyone in the group is negative—for the cost of just one test. If the test is positive, you can follow up with just that group, concentrating resources where they’re actually needed.
Hundreds of school communities are currently using pooled testing. And now each of these schools has a new data point to understand: the percentage of positive pools. But how do you interpret that number? And, how is it different from the positivity rate from individual testing?
What does “pool positivity” mean?
Let’s take Massachusetts as an example. They recently published numbers from all the schools participating in their state-funded testing program. In February and March, the percentage of pools that tested positive across the state was 0.76%. So, does that mean that 0.76% of the students and staff who participated were positive for COVID-19? In short: no! In fact, according to Massachusetts, that’s highly unlikely.
For 0.76% of the students and staff to be positive, every single person in every positive pool would have to be positive for COVID-19. According to their press release, Massachusetts is not aware of any positive pool that included more than one positive person, so it’s likely that the estimated prevalence rate is much less than 0.76%.
Of course, estimating prevalence from pool positivity has a lot of complicating factors. We have to keep in mind that no test is correct 100% of the time and that some students and staff choose not to participate in testing. But, using the data we have, 0.76% is a reasonable high estimate for prevalence.
So if 0.76% is a high estimate, how do we get a low estimate? It’s easy!
The best case scenario is if every positive pool includes just one positive person. We’ll also assume that there are no false positives or negatives, which isn’t necessarily the case, but lets us get to a decent estimate with simple math.
To do that calculation, you just need to divide the pool positivity rate (0.76%) by the average number of people per pool (7, according to the press release), giving you 0.11%. So, from the pool positivity of 0.76%, we can estimate that between 0.11% and 0.76% of students and staff who participated in testing were positive for COVID-19.
If you want to calculate your own low and high estimates, follow along with this worksheet:
If estimated prevalence rates are so low, why even bother testing?
A big benefit of this kind of testing is that it provides information about what’s going on in specific schools. That’s important because if schools go to great lengths to prevent transmission (which is frequently the case), the rate within the school could be radically different from the community rate. That’s true in Massachusetts, which has a statewide positivity rate of 2.49%—a very different number from the 0.11% to 0.76% estimated prevalence from the state’s school testing program.
Again, remember that comparing statewide positivity rates to data from routine testing is a little like comparing apples and oranges. Each strategy has its own biases. Statewide positivity rates may be skewed if people who have reasons to think they have COVID-19 are more likely to get tested. That’s less of a problem with routine testing, but data from routine testing can also be skewed, for instance if people with symptoms are excluded from participating. We can never completely escape bias, but routine testing in schools gets us closer to the data we need.
Although we know much more about COVID-19 than we did a year ago, we’re still in uncharted territory in many ways. Reasonable people have wildly different opinions about how schools should respond to the pandemic. With testing, schools don’t have to guess whether their strategies are working, and neither do parents or teachers. Rather than taking a best guess, schools can act based on hard data and have confidence that they’re being responsible.
The end is in sight, but it’s important that we not let our guard down before it’s actually over. Testing is an important way to help students and teachers safeguard their community as schools open to full capacity.