Wednesday COVID-19 Briefing


Top news, reports and insights for today:

  1. Curated headline summaries for Wednesday:
  • New study finds since mid-May, U.S. has the highest COVID-19 death rate of any country with a large outbreak. Since June 7, U.S. COVID-19 deaths per 100,000 were more than twice the next highest countries (Israel and Sweden) and 6 times higher than Canada (JAMA, see Figure A below)
  • Sweden’s experiment to avoid shutting down is a “disaster” and should not be a model for the world (Time)
  • For the first time, the U.S. has now hired more than 50,000 contact tracers according to Johns Hopkins data. However, the push to grow a COVID-19 workforce is still less than half way to what experts say we need (NPR)
  • Harvard study finds that patients who survive severe COVID-19 infection retain high antibody levels for up to 4 months. This is as close as we can now come to President Trump’s unsubstantiated claim that he now has lifetime “immunity” to the virus (Daily Mail)
  • A 25-year-old man is now the first in North America to have a documented second infection with SARS-CoV-2. His second infection identified in June was more severe than the first found in April. The true extent of COVID-19 reinfection is not currently known (CBS News)
Figure A: Figure created by me based on https://jamanetwork.com/journals/jama/fullarticle/2771841
  1. U.S. daily cases continue to track upward, surges seen across all regions. Three Midwest states remain white hot and getting hotter
     As we move into a new week, the daily U.S. case data shows a continuing trend toward rising transmission intensity with the 7-day moving average lurching over 45,000 daily cases (Figure B). In the last week 340,470 cases were reported, the first time we have seen more than 300,000 weekly cases since mid-August. At my weekend briefing, cases were rising in all but 8 states. The numbers now look worse (Figure C); all but 5 states are now on the incline. There are numerous states in all regions where cases rose 20% or more week-over-week with the largest number in the Midwest: Iowa (+23%), Illinois (+41%), Indiana (+42%), Kansas (+26%), Michigan (+27%), Minnesota (+21%), North Dakota (+31%), Ohio (+26%), and South Dakota (+33%). The situation in the Dakotas remains especially alarming as things have gone from bad to worse. After weeks of very high transmission, both states saw new cases rise by more than a third. North Dakota added another 3,881 new cases raising its overall incidence rate to 3,706 per 100,000, second only to Louisiana. The largest weekly jumps in new cases actually happened in the West where Colorado (+41%), Montana (+63%), and New Mexico (+55%) are potential new hot spots.
     Switching to the rate of new infections, similar states stand out (Figure D). A week ago, North Dakota had 56 new daily cases per 100,000, now they are at 73. South Dakota, then at 47 is now 62. Montana jumped from 35 to 57 in a week. These are big increases in states that have been hot for some time. All but two Midwest states are now over 20 new daily cases per 100,000 and none are below the 5 benchmark. In the South, only Arkansas, Mississippi and Tennessee remain “hot”. The situation is holding steady in the Northeast although new cases are creeping higher in Rhode Island, now at 22 per day.
     The bottom line: The overall picture is toward rising cases across the board and in all regions. The Dakotas remain the hottest of the hot spots, followed by Wisconsin, and Montana. Thankfully, deaths have not yet started to rise appreciably but it is only a matter of time.
Figure B
Figure C
Figure D
  1. Is the current surge in U.S. cases due to flu? Not yet!
      I want to talk about influenza. There has been a lot of discussion lately about what we should expect given the impending rise of flu cases and what that will mean for the COVID-19 pandemic both in the U.S. and throughout the Northern hemisphere. Some people are suggesting that the kickoff of the flu season may be behind the recent surge in cases.
     I decided to go back and look historically at the time course of influenza in the U.S. in the fall over the last three years. Thankfully, the CDC has a great tool for this called interactive FLUView. It shows trends in influenza-like illness (ILI) across the U.S. by tracking patient visits to doctors with flu-like symptoms each week. Figure E below is my best effort to show you what the U.S. maps look like every two weeks from the middle of October till early January over the last three years. My goal was to answer a very simple question. When does the flu season really start to intensify? If it is late October, it might be true we are already seeing the impact of flu on COVID-19 numbers.
     Have a look at Figure E. There are at least 3 main lessons I take from this picture. First, it is exceptionally unlikely that the surge we are seeing now is the result of early flu activity. Over the last three years, the intensity coming in January through March won’t actually get started until somewhere between late November (last year) and late December (2018-19). The second thing that caught my eye was that flu-like illness intensifies first geographically in the Southeast and works its way north. I don’t know why but it is an interesting epidemiologic puzzle. I will be looking carefully at Louisiana, Texas, Alabama and Georgia as the states most likely to show early flu intensity. Third, last year was a high-intensity flu season months before COVID-19 was on the radar screen.
     The bottom line: It is hoped that with good vaccine uptake, mask wearing and social distancing, this year’s flu season will be less severe. If it is, COVID-19 will be more manageable. If, however, we drop the ball on flu prevention, the co-occurence of influenza and COVID-19 could create a double whammy of tremendous death and suffering.
Figure E
  1. Quirky Qorner: Cyberchondria! Social media and internet browsing are driving COVID-19 fear and anxiety
     I started reading this week about cyberchondria, defined on Wikipedia as the unfounded escalation of concerns about common symptomology based on review of search results and literature online. It’s a very pretty word for a not-so-pretty facet of the COVID-19 pandemic. We all like to believe that having access to more information at our fingertips is a good thing. More and more we are realizing that this is often not true. Researchers find that google search trends can actually predict where outbreaks are heating up. Now, a team from the Netherlands has published a study in the Journal of Anxiety Disorders. The study, based on a survey of 439 adults conducted in March, found that the coronavirus outbreak is causing fear and worry to increase. That’s hardly surprising. The interesting part is that even after controlling for other factors, time spent regularly looking up coronavirus information online, through social media and through regular media increased fear. You would think it is the opposite. Regular intake of social and traditional media heightened people’s fears rather than empowering them. Of course, that can’t apply to my blog. Right? Well in their analysis, looking up information on “professional” websites was not associated with increased fear! Whew!

Saturday COVID-19 Briefing


Top news, reports and insights for today:

  1. Curated headline summaries for Saturday:
  • “Uncle Tony” Fauci warns 7 states to take extra precautions over the Labor day holiday to prevent COVID-19 surge (Huffington Post)
  • Up to two-thirds of Americans say they won’t get COVID-19 vaccine when it’s first available, new poll shows (USA Today)
  • After facing criticism for high cases and a go-it-alone strategy, now Sweden has one of the lowest COVID-19 death rates in Europe (CNN)
  • Widespread COVID-19 vaccines not expected till Mid-2021, WHO says (Reuters)
  • A new study (not yet peer reviewed) reports on experiments that show when SARS-CoV-2 was introduced to heart muscle cells, it resulted in “carnage” on the slides, providing clues to explain widespread evidence that COVID-19 does lasting harm to some hearts (STATNews)
  1. U.S. COVID-19 cases spike to 50,000 on Friday, total cases exceeds 6 million. Big spikes seen in the Midwest
    Things were looking good for a slow-down in new cases in the U.S. for about 10 days. I said on Wednesday it appeared the slow-down might be stalling. Like a cold slap in the face, Friday’s numbers leapt to over 50,000 for the first time since August 15 (Figure A). Driving this spike were large rises in several key states increasingly in the epidemic’s cross-hairs (Figure B). Of the nine states with a rate of new case growth of 20 or higher, six are in the Midwest where conditions continue to deteriorate. They include Iowa (+5,851 new weekly cases), Kansas (4,172), Missouri (+9,223), North Dakota (+1,863), Oklahoma (5,780), and South Dakota (2,079). Growth factors in that region showed increasing weekly cases in all states except Michigan and Iowa. The Northeast and West both remain relatively calm although notable surges in cases were reported in Hawaii. Although new cases fell for the week in California, that state still reported over 33,000 new cases in the last 7 days.
    What does it mean: We are still in deep whack-a-mole in the first wave of the U.S. outbreak. We get lulled into a false sense of progress when cases slow in one region, only to see big surges happen in another (in this case the Midwest). Figure C shows the big picture as we eclipse the 6 million case threshold. The most recent 500K cases were added in 12 days (compared to 11 days in the previous half million). The big picture remains largely stable as the U.S. outbreak enters a seventh month.
Figure A
Figure B
Figure C
  1. How is the U.S. doing compared to other nations in the Western hemisphere? Not great.
    As disease detectives, we seek to make apples-to-apples comparison to see how things are changing in different countries. The best way to do that (as I have explained before) is to use Log-log plots showing standardized growth trajectories of cases and deaths setting each country to a time metric indexed by days since a fixed number of cases (as opposed to calendar time). Figure D below does that for cases. The diagonal reference line shows the rate of growth if cases are doubling every week. Countries where cases are growing faster will be above that line, slower nations are below it. The U.S. is the pink line. This figure paints a particularly grim picture. Almost all the nations in our hemisphere saw rapid growth for the first 10,000 cases (on or above the 7-day doubling line). After that, every nation except the U.S. managed to react in a way that slowed the pace of new cases. After about 20,000 cases, every other country managed to get under the 7-day doubling rate. The U.S. didn’t get there until 500,000 cases. This tells us that the first half million cases were especially costly and put us substantially behind the 8-ball. The pace of the US epidemic slowed between 500K and 2 million. Then, things went very badly again: the rate of growth spiked severely between 2 and 4 million. Even compared to Brazil, the U.S. trajectory shows an inability to react early and to maintain epidemic control measures at critical points. While new case growth has been extensive in Argentina, Columbia, Peru and Mexico, none of those nations saw the prolonged unregulated growth seen in the U.S. Canada, which shares exposure to colder weather in the Northern hemisphere, is hidden here among a second cluster of countries that have dramatically better profiles.
    One might argue that the U.S. was testing more than these countries so perhaps cases are the wrong thing to look at. Plus, the U.S. has far more sophisticated health care systems so certainly we won’t see the same pattern for deaths. I’m afraid that picture is equally discouraging (Figure E). Between 100 and 100,000 deaths, even Brazil out-performed the U.S..
    Bottom line: The U.S. has done worse than any other nation (including Brazil) in the Western hemisphere in controlling the speed of the epidemic both in terms of cases and deaths.
Figure D: created by me using COVIDTrends website showing standardized rate of case growth among Western hemisphere nations as of Sept 4.
Figure E: created by me using COVIDTrends website showing standardized rate of growth in reported deaths among Western hemisphere nations as of Sept 4.

Daily COVID-19 Briefing: 5/12/20

Top news, reports and insights for today:

  1. Daily headline summaries for Tuesday:
  • Former FDA commissioner Dr. Scott Gottlieb says we are starting to see cases ramping up in states that have reopened like Texas, Alabama and South Dakota (CNBC)
  • Sweden has eschewed wide-spread “lockdown”. Now, that nation must adjust their strategy in response to skyrocketing deaths in nursing homes (Bloomberg)
  • Johns Hopkins University in collaboration with New York State now offering online training for contact tracers (Coursera)
  • New York City saw 24,172 more deaths than “normal” from March 11 to May 2 based on past trends. About 19,000 were tied to COVID-19, but an additional 5,293 excess deaths remain unexplained and might be coronavirus-related (Bloomberg)
  • Southern states are reopening, but patrons are not yet ready to go out for dinner. Data from OpenTable shows that reservations are still down 83-92% in states that have reopened restaurants (Slate, see graph below)
Screen grab from Article by Jordan Weissmann in Slate: https://slate.com/business/2020/05/south-reopening-restaurants-coronavirus-opentable.html
  1. The whack-a-mole effect: why you shouldn’t trust the U.S. plateau
    Everyday people are talking as if there is one big epidemic occurring in the U.S. (and every other country for that matter). This is reinforced by our preoccupation with the national numbers we look at each day. I’m as guilty of this as anyone. But, the truth is that like politics, all transmission dynamics are local. I struggle in this blog to convey deep truths about the nature of epidemics, many of which are contrary to our common sense. Here is another one: there is no one American epidemic. That’s just not how infectious outbreaks work. What we have is better thought of as a series of interconnected local outbreaks that rise and fall in traveling wavelets of infectious chain reactions. The coronavirus acts like a sneaky serial killer. It moves into a new area, spreads for a time at an exponential rate, kills a bunch of people, and then packs up and hitches a ride with a trucker to the next town where it starts it’s killing spree all over. Call it the Wuhan strangler! This creates a massively scaled whack-a-mole game, where the virus waxes and wanes in communities, rising in intensity here at the same time it is running out of easy targets there. When we see the national numbers plateau, we falsely believe we are bringing the disease to heal. But because there is not one giant wave, we fail to see that this is an illusion.
    I very much liked an opinion piece by Nathaniel Lash in the New York Times on May 6 that puts this issue into perspective. For weeks, the swell of cases from New York was driving the overall picture. As that hot spot started to burn out, it looked like the national epidemic was slowing. It wasn’t. instead, the serial killer was just shifting locations as the G-men were bearing down on it in the Big Apple. The 4 graphs below, taken from the Lash opinion piece show this clearly. Slide A shows the overall flattening of the national curve in the first week of April. Looks like good news right? Only if you are in New York. Slide B shows the same curve after removing the influence of New York City. It shows a different story; new cases elsewhere continue to rise. Figure C shows the national trend removing New York City, Detroit and New Orleans. There is no plateau. Slide D shows the whack-a-mole effect at the state level in Texas and Oklahoma. Texas looked like it plateaued (at least before early May), but actually, the initial hot spot in Houston was winding down, making it appear that the state was waning when the infectious vanguard had just shifted to other places. The same pattern is seen in Oklahoma. Both states have started reopening, resulting in new case surges both in the initial hot spots and the rest of the state.
    What this means: It’s a mistake to think that the U.S. epidemic is one thing. Transmission dynamics happen on smaller scale geographies. This epidemic, like all others, moves through populations in traveling wavelets that rise and fall and shift locations. Keep the whack-a-mole effect in mind whenever looking at trends in big geographies. We are a long way from herd immunity so don’t make the mistake of thinking we have this sneaky serial killer cornered.
  1. U.S. deaths fall to levels not seen since March 31, thirteen states saw deaths climb by a third last week
    On Monday, there were 836 reported deaths, a 1.1% rise, and the lowest total deaths seen since March 31. While this is good news, we are now familiar the drop in reporting seen consistently on Sunday and Monday; we will have to wait to see Tuesday’s numbers. If you read point 2 above, you now know that we must look beneath the overall trend and ask where the epidemic is now surging. The bottom graph, once again, shows change in deaths over the last week across states. New deaths are down in New York (+11%), Michigan (+11%), and Louisiana (+13%). But 13 states reported cumulative growth of 30% or more last week. That’s whack-a-mole. There were several states in each region where deaths were rising. In the West, Arizona, still on the bottom of the list in testing, leads the region with a rise of 54% in total deaths. New Mexico and Utah also saw concerning rises. Six mid-west states reported rising deaths, lead by South Dakota (+60%). Meat processing plants have played a key role in outbreaks in most of these states including South Dakota, Iowa, North Dakota, Missouri, Minnesota and Nebraska. In the South, Alabama and Mississippi saw accelerating growth in deaths. In the northeast, New Hampshire and Pennsylvania remain in the pandemic’s cross-hairs.
    Bottom line: Overall new deaths were down yesterday, a welcome sign, but caution dictates we wait for Tuesday rebound and stay focused on the shifting terrain of traveling wavelets.