Daily COVID-19 Briefing: 5/14/20

Top news, reports and insights for today:

  1. Daily COVID-19 headline summaries for Thursday:
  • Patrons pack Wisconsin bars immediately after the state’s Supreme Court struck down the Governor’s stay-at-home order; video goes “viral” (CBS News)
  • Virologists in Frankfurt Germany have identified a potential new way to treat COVID-19 patients by blocking viral replication. This could be done with a number of existing drugs. Clinical trials are now being launched (Futurism)
  • Virginia has begun to cook it’s COVID-19 books by combining viral and antibody testing, yielding a more rosy public picture and numbers that exaggerate the state’s testing capacity. The lack of national leadership and data sharing standards means more states are likely to cook their COVID-19 books (The Atlantic)
  • Scientists report new evidence supporting the efficacy of masks, finding that normal speech causes airborne virus transmission in confined environments for up to 14 minutes (Proceedings of the National Acadamy of Sciences)
  1. Which states have peaked? Which are still surging in cases?And what’s up with this weird zig-zagging?
    It’s been two weeks since I showed you the latest plots from #COVIDMonitor, a fantastic site run by Ralph Straumann. His plots show newly tested cases in the previous week vs. cumulative cases since the start of the outbreak using data from the New York Times. The plots are displayed on the log scale for both last week and cumulative, so the horizontal axis is not time. I like these so-called log-log plots because they show a clear downward hook when a given place has reached peak cases. This hook is the best way to detect the inflection point between accelerated (non-linear) growth and simple additive growth in cases. The graphic below is a screenshot from his site today.
    The first point to make is that thankfully, there are some states that clearly look to have peaked. That includes out West Alaska, Idaho, Montana, and Hawaii. In all those states, we see the obvious hook. Similarly, we see the peak signal in Vermont, New York, and Connecticut in the Northeast. In the South and Midwest, things look good for Michigan and Louisiana. That’s just 9 states where cases have peaked. The other 41 fall into 3 categories. First, there are twice as many states (16) where cases continue to rise non-linearly, meaning new cases are increasingly increasing. Most are in the middle of the country: Arizona, New Mexico, Kansas, Colorado, Nebraska, Iowa, Minnesota, Illinois, Indiana, Mississippi, Kentucky, Maryland, Virginia, Washington DC, North Carolina and Massachussetts fall in this box. What surprises me most (although perhaps surprise is not the right term) is the 12 states that display a zig-zag pattern, where the peak seemed to arrive over the last few weeks, then cases ramp up again. Instead of a clear downward hook, we see a jagged pattern of downturn followed by upturn. Take a close look at Washington, Oregon, Wyoming, South Dakota, Wisconsin, Tennessee, Missouri, Arkansas, Alabama, Ohio, West Virginia, and South Carolina. I believe this is an important observation. The states that are zig-zagging outnumber those that have peaked. A final group of states have turned flat in recent weeks, but the hook has not yet materialized. I would count California, North Dakota, Nevada, Utah, Indiana, Oklahoma, Texas, Pennsylvania, Georgia, Florida and South Carolina in this camp. Although there may be a faint glimmer of an upward zig-zag at the very tail in Texas and Oklahoma.
    The Bottom Line: It is crucial that we look for a key-like zig-zag pattern in these plots because that is where we are likely to see a rebound of new infections in states that reopen, as small transmission chain brush-fires ignite that have the potential to agglomerate and restart the whole exponential growth phase, sending states right back to where we were in March and April.
From #COVIDMonitor, United States, captured 5/14/20 at https://rastrau.shinyapps.io/covid-us/
  1. ORIGINAL ANALYSIS: The U.S. has a long way to go in testing: A look at test positivity rate (TPR), a vital sign of the health of a testing regime
    The testing talk hasn’t died down, nor is it likely to in the near term. Earlier this week, “Uncle Toni” Fauci warned congress in a surreal video testimony that the U.S. is taking a big risk by opening states too soon before we increase our capacity for testing and contact tracing. For many, this isn’t actionable enough. We need some concrete things we can track to see how well our testing is actually doing. In past briefings, I have talked about differences across states in testing rates (See May 7 briefing). Here’s something new I cooked up and I think it’s pretty interesting. Another key metric is the test positivity rate, or the percentage of tests that are performed that are positive. It’s a vital sign on the health of a testing regime. Early in an outbreak, testing resources are scarce and most tests are done only for people who are already sick. But almost 3 months into the epidemic, we expect to see the TPR come down. Way down. If it does, it tells us that we are testing a broader swath of the population beyond those who are already sick and have a high likelihood of being infected. The goal is to cast an increasingly large net, testing people with a lower probability of being positive. In an ideal scenario, when we could test everyone in the population, the TPR would equal the population prevalence. As long as it’s much larger, it means we are still testing selectively.
    As a nation, the U.S. is not doing well. The top graph below compares TPR by countries using data from FactCheck.org. While the number has been coming down, we are still seeing over 13% of our tests come back positive. The nations that have become models of effective testing and epidemic control, like Iceland, Denmark, New Zealand, South Korea, Germany and Taiwan, are all at or below 5%. That’s casting a broad net. Five percent is a good goal because the prevalence estimates from Europe have been around that range. The bottom graph shows how states are doing using data from the very handy COVID Tracking Project. Keep in mind this is one overall number for the entire period of the pandemic, so the states that had the biggest outbreaks first will have large TPRs. The take home messages here are three:
    1. Fifteen states have test positivity rates over 15%. Six of those, all in the northeast, are still above 20%. This tells us that the states at the front edge of the epidemic still have a lot of catching up to do.
    2. Good news: 11 states are below 5%, indicating wide and deep testing systems. Not surprisingly, these are mostly states that have not had huge outbreaks. Congratulations are especially due to West Virginia, Vermont and North Dakota, all get high marks across a range of testing measures.
    3. Several hot spot states have reopened despite rising rates of new cases. This is especially troubling in those states that also have persistently high TPRs, such as Maryland, Pennsylvania, Illinois, Virginia, Iowa, Georgia and Nebraska.
Test positivity rate by country, screen capture from FactCheck.org on May 13, 2020: https://www.factcheck.org/2020/05/testing-by-the-numbers/

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s