Editor’s note: The Salt Lake Tribune is providing readers free access to critical local stories about the coronavirus during this time of heightened concern. See more coverage here. To support journalism like this, please consider donating or become a subscriber.
What we as Utahns are doing is working.
You might be getting tired of staying at home or wearing masks or having your favorite restaurant limited to just take out, but the measures we’ve implemented to stop the spread of COVID-19 have been effective so far, and I can prove it mathematically.
Loyal readers may remember the concept of R0 from a column of mine that came two weeks ago. R0 is vital when it comes to understanding the coronavirus: it describes how many people the average sick person infects. If R0 is high, the disease will spread like wildfire through a community, potentially overwhelming hospitals and leading to a much higher death count.
That "0″ in R0 refers to the initial set of conditions — in other words, how our society acts normally. However, epidemiologists also want to see how the contagion rate R changes over time, especially as society adjusts to the crisis. We call the changing contagion rate Rt.
Stopping sports on one day would lower the Rt some, then closing schools the next day would lower it again. Instituting a stay-at-home order reduces it even more. That makes sense, right? But conversely, keeping a meat-packing plant open and encouraging sick workers to continue working would drive that Rt way up. More on that later.
Rt allows us to track how contagious the virus is in our state, with our conditions. The ultimate goal is to get Rt below one: If the average infected person spreads the disease to fewer than one person, the disease dies out. If Rt is above one, the disease still grows until it runs out of people to infect — though having it grow slowly has the benefit of not overwhelming the health care system.
Theoretically, by looking at the number of new Utahns infected on a day-by-day basis, and comparing it to how many people were infected the day before, we could easily estimate Rt to see how effective our coronavirus-stopping measures are.
In practice, it’s a little more difficult than that: there’s randomness to consider, changes in testing protocols, delays in reporting, and other factors. For example, around the nation, fewer coronavirus cases are reported on Mondays. That doesn’t mean the virus is spreading any differently on Mondays, it’s just a quirk of the reporting system.
But epidemiologists are clever, and have come up with ways to estimate Rt in real-time despite the challenges. One algorithm was described in a 2008 paper by Luís Bettencourt and Ruy Ribeiro — an algorithm data scientist Kevin Systrom implemented to describe the spread of COVID-19 in 2020. Essentially, the approach is this: rather than approaching the day-to-day counts as gospel, we consider them hints towards what’s actually going on. One good day or bad day informs rather than overwhelms the model.
Systrom published his algorithm, and created individual charts for each of the 50 U.S. states. Here is Utah’s:
The red line is the changing estimate of Rt, which takes into account the past seven days of data, while the gray shaded area around the line accounts for the potential error in the estimate. The algorithm’s best guess for Rt in Utah on March 15 is 3, but it could have been anywhere from 1.01 to 6.99 on that day. We just didn’t have a lot of info back then.
But every day, we got more information, and our community implemented more measures that slowed the spread locally. You can see that in the declining Rt estimate — around April 8, we got our Rt to below one. It appeared that daily case counts peaked.
But as case counts rose again in Utah over the past couple of days, the Rt estimate swung back above one. That rise is largely due to two factors: a change in testing protocols so that more Utahns are checked for the disease, and the outbreak in a South Salt Lake homeless resource center, which saw 94 positive cases out of just 205 men tested.
Right now, Utah’s Rt estimate is 1.12, but could realistically be anywhere from 0.57 to 1.72.
How does that compare to other states?
Utah ranks 20th right now in terms of Rt estimate, pretty middle-of-the-pack. It’s interesting to note that states with absolutely no lockdowns do seem to congregate at the bad side of the spectrum, but the states with partial shutdowns like ours don’t seem to fit any trend — partial, targeted shutdowns may well be enough.
I thought it was interesting how many relatively low-density states found themselves with high Rt numbers. You would think that it would be our most populous states that would have the hardest time, and we certainly see that in the death counts. But in terms of managing the ongoing spread, some low-density states have still run into trouble.
In particular, the example of South Dakota is fascinating. As of Friday, they’ve had 1,411 positive test results in a state of 884,000 residents. That’s a pretty high number, but it turns out 55% of cases are tied to one meat processing plant, Smithfield, in Sioux Falls. There, 777 people have tested positive: 634 workers at the plant, and 143 of their contacts. It is the largest coronavirus hotspot in the country.
It all unfolded pretty quickly, too. On March 20, Smithfield announced that they would continue operating normally to support the U.S. food supply. Makes sense: people need food. On March 26, though, they had their first employee test positive. At that point, employees asked for hazard pay, as they felt they were risking their health to come into work. Instead, Smithfield offered a $500 “responsibility bonus" to those who worked their full allotment of shifts in April.
That was an incredibly bad idea — it essentially was a bribe for people to continue to come to work, even when they were sick. And lo and behold, cases started to explode: there were 80 positive tests last week, and 700+ this week. They’ve since closed the plant, but it was too late.
I think it’s a case that has significant relevance to Utah’s situation. Right now, we know about the homeless resource center’s hotspot. We’ll need to track that carefully — and homeless people are a difficult group to track.
But in May, it seems we’re going to start opening some other workplaces. The particulars of how we go about doing that, especially in relatively large businesses, are going to be extremely important. In rural and suburban communities like those most Utahns live in, work is going to be the most likely location to catch and spread the coronavirus. If managed incorrectly, many workplaces can be the site of a super-spreader event.
By opening up, government is putting its trust in the hands of the private sector to prevent the spread. Businesses will have to be extremely careful about keeping facilities clean, employees separated and sick workers at home. If employers can’t manage to ensure their employees’ safety, we’re going to see Rt rise significantly once again, a second wave of the virus that will necessitate a second shutdown. That’s not what anybody wants.
So let’s keep an eye on Rt — the single number that will impact Utah’s coronavirus future most.
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 email@example.com or on Twitter at @andyblarsen.