The science of Sundance: Digging into a theory the coronavirus was spreading early in Utah

Editor’s note: The Salt Lake Tribune is providing free access to critical stories about the coronavirus. Sign up for our Top Stories newsletter, sent to your inbox every weekday morning. To support journalism like this, please donate or become a subscriber.

Have you noticed how many people think they’ve already had the coronavirus?

It’s a common refrain in Facebook messages, among Salt Lake Tribune commenters and more: A significant number of people think that the bug they faced in December, January or February was COVID-19.

These suspicions exploded after we got news of early cases. For example, Santa Clara, California, did some backtesting and found a Feb. 6 death from the coronavirus. A man in France was found to have had it in December. And quite frankly, given the lack of testing and knowledge we had about the disease then, that suspicion is fair.

Now comes this story from The Hollywood Reporter that hypothesizes that the Sundance Film Festival in Utah was an early super-spreading coronavirus event. The magazine spoke to more than a dozen people who said that they had significant illnesses after attending the festival, where people pack into movie theaters and afterparties.

If some of these sick people really did have COVID-19, it would significantly alter Utah’s timeline with the disease. Sundance went from Jan. 23 to Feb. 2, while Utah’s first case wasn’t announced until March 6.

Utah state epidemiologist Dr. Angela Dunn was asked about this Wednesday, and said “It is definitely possible that COVID-19 was circulating at Sundance.”

Let’s say you were an epidemiologist like Dunn trying to figure out whether or not COVID-19 was at Sundance. Because coronavirus testing wasn’t really available then, you’ll need to be clever about how you do your research. What might you look at?

The obvious first step is to conduct antibody tests on those people who got sick after attending Sundance. Antibody tests aren’t perfect, with up to 10% failure rates, but they’d give you a good idea if someone had the disease a month ago.

So far, I’ve found two people who took the tests and then commented on social media about their results. One, a commenter on The Hollywood Reporter’s Facebook page, said “I live in Park City. I too had these symptoms on January 24. Went to urgent treatment. Negative for flu & strep. Had the antibody test this week for COVID19 bc I was convinced I had had it - test was negative.”

Another social media account, from Scott Dougan on Twitter, reported similar results. “I was at Sundance, came down with the described cold — cough, aches, worst I’ve ever had — and recently was tested for antibodies. I tested NEGATIVE. That’s one data point that indicates this cold was not COVID-19.”

But of course, two people posting their results on social media isn’t science — it’s certainly the flimsiest bit of evidence I’ve included in an article on COVID-19 so far. We’re looking for something more.

Another breadcrumb might be the number of deaths in Utah. Given that about 64% of Sundance festivalgoers are Utahns, not from out of state, you’d expect that an outbreak at Sundance would expand to the local community. If people were catching this at Sundance — especially at the rates discussed in the Hollywood Reporter article — then spreading it to loved ones, you’d expect a greater number of Utahns to die by the end of February or beginning of March than we normally see.

But in February, fewer people died in Utah than usual, according to numbers from the Centers for Disease Control and Prevention. Nor were there big numbers of deaths from flu or flu-like symptoms.

Again, that’s just a hint, though: it really still could go either way.

What we really want is something more certain. Enter genomic epidemiology. I realize I’ve lost some of you with this two word phrase, but I’m asking you to give me a chance, because it's extremely cool stuff.

Like everything in the world, the coronavirus has a genetic code, right? That code handles all aspects of the virus: how it looks, how it operates, how it replicates, etc. Every time a new bit of virus is created, that genetic code is copied.

The copying process isn’t perfect, though. It’s 99.999…% perfect, but occasionally, it messes up and puts something wrong in the code. This is called a mutation. Most of the time, the mistakes made are harmless, like a typo in an article that still gets its point across.

But these typos remain as the virus is copied into new people. That means they have an upside: if we look at the genetic code of the virus found in different people at different times, we can track where they likely got it based on what mutations appear.

One mutation that has received a lot of press recently is called D614G. You may have read about it in the The Los Angeles Times, with a headline that reads “A mutant coronavirus has emerged, even more contagious than the usual, study says.” That’s a scary headline! Let’s break it down.

Essentially, that D614G means “At position number 614 (of the spike protein) of the coronavirus, replace the letter D with a G.” This particular miscopying didn’t really start appearing until the disease reached Europe, when scientists found it in earnest beginning around Feb. 20 or so. So it’s really easy to see a split: those in Asia and some of the virus-carriers on the U.S. West Coast mostly have the letter D in that spot, whereas Europe and the rest of the U.S. has a G. (D is green and G is yellow in the map below.)

NextStrain visualization

What does this have to do with Sundance? Well, if all of the Utah cases had a G in that spot, you’d basically be able to rule out any Sundance spread of the coronavirus. After all, the G strain didn’t even exist until late February, after the festival ended.

Of the 141 samples uploaded by the Utah Public Health Lab available on NextStrain, over 85% of them have the G strain — the new one. But there are a small percentage of instances of the old D strain in Utah, and they seem to be clustered earlier than the G ones, too.

NextStrain visualization

What does this mean? Well, we can’t rule out Sundance spread with this mutation: enough of the D strain exists that it could be lingering from Sundance. Nor can we confirm it, though. It could be that someone from Asia, Washington state or California visited Utah in February or March, and started the D strain here that way.

In other words, from this limited analysis alone, Dr. Dunn is right: It is definitely possible. But given the small number of D strain cases found in Utah, I don’t think it’s particularly likely that Sundance was the site of a super-spreading event. If it were, wouldn’t it have spread more commonly into the Utah community?

Clearly, later someone else visited from the Eastern United States or Europe and started the more dominant G strain in Utah. Experts are unsure on whether the G strain is actually more contagious than the D strain, or if it has taken over as a matter of circumstance. It does not appear to be more deadly to those who catch it, though, nor should it markedly affect vaccine development.

By the way, there are lots of these sorts of mutations, including some with local ties. For example, while playing around on NextStrain, I found one at spot No. 146 on the spike protein that turns a H into a Y. This mutation has almost only been found in Utah tests. It was first discovered in a sample tested here March 20, then scattered around Utah tests through early April.

NextStrain visualization

Is this cause for alarm? No. It’s just another random mutation. And quite honestly, if it were dangerous, we would have seen some evidence in Utah’s coronavirus spread or death count, and that just hasn’t happened.

In general, far fewer people have had the coronavirus here in Utah than think they’ve had the coronavirus. Maybe the best piece of evidence for that is our positive testing rates: even early on when most people needed both symptoms of the disease AND close contact with a positive case to get tested, those people tested positive at roughly a 5% rate. That means, 95% of them had something else.

I know: it sure would be nice to have had the virus earlier than everyone else. As Andrew Noymer, an associate professor of public health at the University of California Irvine told The New York Times, “Everyone desperately wants to be immune to this thing, and they’re projecting their hope onto the data.”

But the numbers show... that nasty bug you got early in the year? It may have been coronavirus — but it probably wasn’t.

Andy Larsen is a Tribune sports reporter who covers the Utah Jazz. During this crisis, he has been assigned to dig into the numbers surrounding the coronavirus. You can reach Andy at alarsen@sltrib.com or on Twitter at @andyblarsen.