Dallas has released a thorough report of their COVID-19 situation.
And now I’m scared.
COVID-19 is an influenza-like illness (ILI). Flu season generally runs October through May. Thanks to regular disease surveillance, we know how many emergency room visits are due to ILI. This type of surveillance alerts public health departments to any irregularities which might be a new disease. In the first graph, we see ILI for the past four years. These could be due to flu, adenovirus, rhinovirus, or even COVID-19. In Dallas, ILI surveillance showed an early peak this year, about four to six weeks earlier than normal. Was that COVID-19? It’s impossible to know now. But what we do know is that new ILI visits to the Emergency Rooms are popping up in March due to COVID-19, driving that peak upwards when it should have been flattening.
This trend in early ILI was seen nationwide. The national baseline is 2.4% (black dotted line). While we’ve seen early peaks in the past, this season we crossed the threshold at Week 44. In the previous years, that threshold was crossed in weeks 47, 48, 49, and 52. The 2009–2010 season was the year of the H1N1 Pandemic which started in April 2009 and continued through the fall and late into 2010.
In Texas, we usually see low level flu through December, then an onslaught of cases in January and February. This year was different. In Dallas County, influenza hospitalizations started to pick up around mid-November (Week 47). Like the ER visits, this uptick was earlier than usual, though cases hit their peak in late January and early February. Flu cases usually start to drop off in March (Week 10).
But two weeks into tracking COVID-19, hospitalizations are already where we’d be well into the start of flu season. They’ve quadrupled in one week. While the numbers are still low, at that rate of increase, Dallas hospitals will surpass a normal flu season very soon.
And now for the scarier part. We can assume that a COVID-19 patient is hospitalized because their illness is bad enough that they can’t get better at home. How bad? That’s when we look at patients in Intensive Care Units (ICU). By the second week of tracking (CDC Week 12), the number of patients in ICU is dangerously close to peak flu season.
At this rate, Dallas will also surpass the number of patients in ICU during peak flu season by next week. Not only are more people being hospitalized, more people require high level intensive care.
I have no doubt Dallas’s hospital systems are working furiously to address capacity and equipment, but there is a limit on human capacity. Those health care workers need to sleep, and they’re high risk for exposures.
Shutting things down will limit contact, and hopefully limit the spread, but not on its own. With a two week incubation period, a household of four could take as long as 56 days to go through a round of infection. They may all be low risk and not need hospitalization, but imagine one of those four is 85 year old grandma or an immunocompromised child. We could have prevented those two hospitalizations by identifying and isolating other household members. If not, we send two more people to the hospital.
As described earlier, Germany, Australia, South Korea, and Norway have done extensive testing, tracing and tracking. While some have found a higher infection rate, they have a much lower case-fatality rate and are flattening the curve. Why? Because they isolate those infected individuals. Many are either asymptomatic or have mild symptoms and don’t need hospitalization, but they’re still infectious. By isolating them, they don’t infect anyone else. To know whom to isolate requires testing and contact tracing. These countries have flattened the curve without completely shutting down businesses.
I try to be the voice of calm, to explain the data to make sense of the story. From a data nerd perspective, Dallas’s data report is a beautiful thing. Their unfolding story, though, has made me nervous.
Edit: the original link to Dallas county’s pdf is not functional. Here is the website.