Thursday, July 30Comparing apples to apples.“…When you’re testing to that extent, you’re going to find more people, you’re going to find more cases…” — President Donald TrumpIf you took two identical states, with identical breakouts, and ran twice as many tests in one state as the other, the state with more tests would have more confirmed cases. Multiple unconfirmed cases exist for each symptomatic and tested case.How can we compare the outbreak in one state to that in another? My data science work contributed to the answer to this problem, as I detailed in an earlier blog post. With the ability to estimate the number of people infected from the number of confirmed cases and the test positivity, we can make fair comparisons across states. We simply divide infections by the population.In April, I created plots of Rt vs. the percent of people infected in each state. The percent infected was a 10-day sum of the people infected, because in patients more than nine days after symptom onset, replication competent virus cannot be recovered.Over time, I added different context to the plot. A red zone was added if your current plot position predicted your hospitals would be overrun in two weeks. A safe green zone was added at the suggestion of a friend.When I post such things on social media, I try to remember that a little bit of irreverence is useful to remind people this is a social communication channel and not an academic one. I call the safe zone the O-zone. These O-zone plots allow you to compare outbreaks across states fairly, and also let you see, from the Rt position, whether things are getting better or worse. Here is today’s O-zone plot.
Just a quick glance at this plot, and you know that Maine and Vermont have their outbreaks in much better shape than Georgia and Florida. The faint version of each state is its position three days ago. Georgia was, in late May, headed for the safe O-zone with an Rt under 1 and a percent infected close to 0.2 percent. Today, you can see Georgia has 10 times as many infections, with a percent infected close to 2 percent.The sensible public health approach was taken by most of Europe. You determine your percent infected and Rt. If community transmission is low enough and Rt under 1.0, you devise plans to allow more community activities with plans for mitigation and surveillance. Then you watch and wait. If transmission stays under control, you open up more community activities.Ultimately, you determine that as long as COVID-19 is around, some activities have to be performed with more mitigation and surveillance than before, but are allowable. Some activities are simply unsafe. And we remember from yesterday that control of community transmission is required to control transmission everywhere else.The opening of activities in the state of Georgia paid no attention to Rt. On June 13, when bars were opened and restaurants allowed at 100-percent capacity, Rt was already higher than one. The state had inadequate plans for mitigation and no plan for surveillance, and we predictably moved to the red zone for community transmission.We can re-take control over community transmission whenever we like. It simply requires halting community activities that generate a lot of infections, getting Rt under 1, getting the infection density to one tenth its current level, and then carefully re-opening activities with realistic mitigation and surveillance plans guided by infectious disease epidemiology.This approach embodies the principles of public health.Safety at schoolA group of dedicated infectious disease epidemiologists are out there. One that I follow on Twitter, Dr. Saskia Popescu of the University of Arizona (@SaskiaPopescu), has been helping create graphics to be used for guidelines about risk in different types of activities. We have local primary and secondary schools opening Monday, so I thought I would share one graphic from her twitter feed.The credits for creating this graphic are indicated in the image. This graphic can and should be carefully studied by those planning on opening schools, because it shows which activities require more mitigation and surveillance, which ones require less, and how you can reduce risk.
David Blake is a professor in the Department of Neuroscience and Regenerative Medicine at Augusta University. His Ph.D. is from Johns Hopkins University.Contact Dave Blake.