Antigen tests are cheap, easy to read, widely available, and importantly, sensitive when viral loads are high.

Somehow, it’s already August. And that means soon millions of children across the country will be heading back to school.

We thought the days of broad school closures were behind us. Though there was a concern last year that reopening schools would lead to a bunch of adorable little viral vectors running around, driving community spread, no such link was found. Schools, especially those that did a good job with masking and ventilation – appeared to pose a relatively low risk to public health.

This school year, things have changed. On the plus side, teachers – who are at substantially higher risk of serious complications from COVID-19 than their students– are able to be vaccinated. In some cases, they have even been mandated to be vaccinated.

But something else has changed. The delta variant has turned out to be a transmission machine – with a highly-cited, and still not yet peer reviewed study suggesting 1000-fold higher viral loads in the noses of those with delta compared to other variants, and an internal CDC document likening the infectiousness of delta to chickenpox – pegging the R0 (the average number of people an infected person infects) as high as 8, which is crazy.

So yes, schools were safe last year.  Will they be safe this year?  Does good ventilation really help when you’ve got 1000-fold the virus floating in the air?

Fortunately, the kids should be alright. Outcomes in children infected with Delta don’t seem to be any worse than other variants which is to say – the disease is pretty mild in most kids. But kids have a pesky habit of living with adults and if there is substantial transmission in schools, we might see worse community spread. How can we keep schools open?

One answer, tossed around since the beginning of the pandemic is testing. And here we have advantages we didn’t have last year. Last year, before school started, I was on the phone with my local superintendent trying to hash out a pool testing system for our local public schools. They were interested in the idea – but at the time I couldn’t find a lab that had the capacity to do it and get results in a timely fashion.

Now we have a new option – rapid antigen tests that cut the lab out of the picture entirely.

There are over 30 rapid antigen tests that have received emergency authorization from the FDA and many of them can be administered and read by minimally trained personnel. They are also substantially cheaper than PCR tests.

Of course, as you’ve heard, these antigen tests are less sensitive than the PCR-based tests we’re all used to. That might not be a bad thing though.

Looking at the antigen test literature, a theme quickly emerges. 

This study in PLOS One performed PCR and antigen tests (in this case the BINAXNow test – results in 15 minutes) on 783 kids who came to a walkup testing site in Los Angeles.

226 had positive PCR tests – that’s the gold standard. 127, or 56% also tested positive on the antigen test. That may seem disheartening, right? The antigen tests missed more than 40% of positive kids.

But PCR tests aren’t actually just positive or negative. They can also give a window into viral load. It’s not perfect, but lower cycle thresholds on PCR imply more virus in the nose. There were 16 kids with cycle thresholds less than 25 in the study – think of them as really infectious kids. The antigen cards found 15 of them. Importantly, false-positives on the antigen cards were super rare.

A similar study out of Boston looked at the Binax test in a drive-through clinic – this time with nearly 1000 kids.

134 of them were PCR positive, and the antigen card caught 92 of them – about 70%. But when you looked at those kids with low cycle thresholds, high viral loads, that sensitivity shot way up – to almost 99%.  This is captured in this graph – the red dots are the false negatives – the kids and adults the antigen cards missed – they all occur at those really high cycle thresholds, ie people with low viral loads.

Liverpool, same design, different antigen test – this time an Innova version.

 Same finding – as the viral load goes up, so too does the sensitivity of the antigen test.

The point is that antigen tests aren’t great at picking up all COVID infections, but they might be fine for picking up the infections that matter. Given their low cost, and lack of need for a lab to run the assays, could they be the solution to keeping schools open this fall?  

It could work. Massachusetts tried this in a clever way in fact.

Across 582 schools, they used pooled testing of PCR swabs to identify cohorts where someone was positive. Then, they used antigen cards to figure out who the infected person was. They don’t give hard numbers but note that provided pool positivity was low, the costs were low and the idea is broadly scalable.

The main drivers of antigen screening costs will be how often you screen and how many kids you screen. There are 56 million schoolchildren in the US. Screening each of them once a week for the 40-week school year would cost the federal government $11 Billion, not including some economies of scale. Not testing vaccinated kids (that number should grow with authorization of vaccination for younger kids) would bring those costs down further. Still, it’s a lot of money, but given the amount committed already to coronavirus relief, it’s a line item worth some serious consideration.

A version of this commentary first appeared on medscape.com