Ebola and Coronavirus Update: 17 Sep 2020
Coronavirus ArchiveEbola
–Almost a copy/paste from last week. Currently at 115 confirmed cases and rising. Contact tracing, as a percentage of contacts, has gotten worse. Travel screening numbers in terms of percentage of travellers through Ebola checkpoints has gotten worse. Response efforts remain constrained by insufficient resources, insufficient people, those they have are on strike due to non-payment of salaries, and resistance from the community to some of the containment measures. The only bright spots are that it has not left Equateur province (yet), and a couple spots within Equateur province have now gone 21+ days without a new case. Regardless, expect to see a very similar paragraph next week too, as odds remain better than not that this will get worse before it gets better.
Coronavirus
–First items of business. I cannot comment on the following press releases:
Thank you for your understanding.
–Second, ‘rona ’round the world. In the US, hospitalization rates for COVID nationally continue to crater–hence, you have not heard about hospitals near capacity in weeks.
The Rt is in the green in more states than last week, with a lot more hovering right around the 1.0 line (plateau). Another sea of mostly light green to light pink across the nation on the Johns Hopkins “Have States Flattened The Curve” map of rolling average of new cases…
…except for Wyoming. Yes… it’s happening! Wyoming turning a little pink this week and nudging that Rt over the 1.0 line… Now, I’m not sure yet if that is people tested or tumbleweeds, but if Delaware and Wyoming are two of your transmission hotspots on the week, again, the virus is having to go to greater lengths to find new susceptible hosts.
In Europe, the delayed second wave continues apace. Sweden, much like New York here stateside, remains pretty persistently flat in terms of new cases though. India is still climbing the first wave peak. Elsewhere in Asia, places like Japan and South Korea are coming down off either a true second wave, or pretty delayed first wave. That’s extra good news with flu season coming up here, and suggests that if there is a second wave of COVID-19, it was not as bad as the initial outbreak. In South America, Argentina looks like it has hit the end of the first wave.
–Last week there was a late breaking question from a reader about internet rumors that Fauci had made an important mistake at the beginning of the outbreak, and had taken the case fatality rate (CFR) for the infection fatality rate (IFR). As the CFR is generally a higher number, and a higher number is extra bad for an IFR, this was claimed to have led to an over-reaction by the entire world to COVID-19.
First, what the hell is the difference between the CFR and the IFR?
The short version is that CFR is the % of people with symptomatic disease (so clearly a case of COVID-19) who die. The IFR is the % of people who die of COVID-19, counting even those with mild or NO symptoms at all (basically, what percentage of the people exposed to an infectious dose die). As deaths are generally concentrated in people with symptoms, and the number of people with clear symptoms is less than the “I dunno, maybe it’s a COVID case?” that get lumped into IFR, the CFR is generally higher than the IFR. So if you have a large IFR, it implies a pretty significant CFR.
Make sense? I hope so.
Now, I have no idea what data Fauci had in front of him back in March. I do know that back then, testing around the world was not as plentiful as it is now. As a result, and as we discussed at the time, there were very tight restrictions on who qualified for the limited number of SARS-CoV-2 tests. In some cases, too tight. If you remember, we discussed how the restriction to travel history to or from China was too restrictive, and was likely missing cases in the US that were contracted from people traveling to and from South Korea and Italy, or in contact with people who did. You also needed to have symptoms of COVID. So what we were measuring, and largely ALL we were measuring back then, were bona fide cases of COVID-19. This was a pretty pure CFR measurement, and why we were discussing CFR at the time all of this was first starting. We only got the first true inkling of the IFR when the cruise ships got hit, and they tested everybody on board–finding many more SARS-CoV-2 infections than cases (more asymptomatic or mild symptoms at best positive cruise goers than people with all the classic COVID flu like symptoms). So all Fauci and the epidemiologists across the world had at the time was a CFR. They could make projections from that of the IFR, and did. All models are wrong; some are useful. Some projections assumed too many infections would become cases. As we have learned these past many months, a lot of people will have no or mild disease at best.
We said, at the time, that we expected the CFR to fall as more mild cases were tested and detected, and that our back of the envelope when all was said and done was “somewhat worse than a bad flu.”
Based on all available estimates since then, that is largely the case. The current best estimate of CFR in the US is a little over a seasonal flu–overall. Again, this is age and comorbidity dependent. This is barely a flu for those less than 50, especially with no co-morbidities. But for those 65 and over, it’s like multiple flus at once (CFR estimate 1.3%). Again, that is 1.3% of those 65 and older with definite COVID symptoms and proof of the virus (a “case”) expected to die of disease.
Estimates for the IFR, by contrast, in March by epidemiologists around the world that were projecting millions of US deaths if nothing was done ranged from about 0.8-0.9%. Again, that is IFR, so 0.8-0.9% of people exposed to an infectious dose expected to die.
Since March though, testing capability expanded. As we have discussed a lot, more mild cases and more asymptomatic cases got tested. In fact, right now, I would argue you are getting a better measurement of the IFR than the CFR with all of the “positive cases”–many of which are asymptomatic to mild at best, with PCR proof that the virus happens to be there. The very definition of IFR. Low and behold, the IFR is lower than the CFR. The IFR has varied by country somewhat, and studies have ranged from 0.02% to 1.3%. Overall, an IFR in the 0.4-0.6% range seems likely.
Yes, Virginia, that is about half the IFR projected in March, and why, fortunately, the dire death predictions at the time never came to pass.
But for comparison, the seasonal flu IFR runs in the 0.1% range. Again, the best analogy is that this is just the worst flu ever.
–And remember, and I cannot stress this enough, the threat of coronavirus, and the problem it DEFINITELY was in March/April/May, is NOT the direct mortality of the virus itself. It’s the ability to take up the next marginal hospital bed, and the “bed’s taken” effect on overall mortality.
–“But we shouldn’t have shut down!” I hear you. That’s hindsight though. The lockdowns, at the time, I think were justified to let testing capacity catch up, and disrupt transmission. And they definitely disrupted transmission (but did not stop it–that’s a key learning point going forward). Hence, all the “delayed second wave” action going on now. Most importantly, the lockdowns allowed the army we had at the time to become the army we have now, in terms of expanded hospital space and familiarity with how to triage and treat cases of COVID. Nowhere was that more evident than when Houston was getting bombed by the delayed second wave, at 100% hospital capacity, and the CEOs of its hospital systems were on the air saying “meh,” because they had used the time gained during the lockdown to plan for that exact contingency and were already executing that plan.
–“So how long did the lockdowns really need to be then?” If I were king of the world, a lot shorter than what we got in most places in the world. But that’s fundamentally a political question, so it’s just my opinion. I don’t have a good science based answer because, hey, we’re conducting this experiment on the fly as we speak. Is it harder to justify continuing some of those measures as hospital capacity has expanded, hospitalization rates have dropped, and more experience with treating severe COVID has been gained? I think so–but again, that’s my opinion on what has largely been a political decision.
–And perception of effectiveness is as good as being effective in politics.
–Of course, it certainly doesn’t help the argument of politicians when they are falsifying data to maintain the course they have chosen for political reasons. The most recent example of this is Nashville, where the city government has been accused of faking infections tied to bars, gyms etc. in order to excuse continued orders to keep those venues closed. As we said back in March, there is a fine line between the very real damage caused economically by closing borders and businesses and the risk to the health of the population. If true, the city of Nashville may be looking at a very expensive series of lawsuits from people whose livelihoods were wrecked by orders that may not have been justified based on the actual epidemiological data.
–That also makes me wonder about the data, its source, and accuracy underlying some of those infographics on “what activities are the riskiest” for transmission of COVID. For example, if data included Nashville’s reports on cases traced to exposure in bars, clubs and restaurants, they are suspect now and may overestimate the risks of acquiring coronavirus, even if being prudent in those establishments.
–Other questionable data is an economic focused report claiming the Sturgis bike rally was a super spreader event, released by four economists with San Diego State University. I had hoped to avoid having to go “Plandemic” on this, but no such luck. This report was released and not published. It did not go through peer review. The authors merely posted it publicly on the internet. That is your first red flag. Something like this should have gone through peer review and all the authors involved have, in fact, published in peer review journals before. Why not now? A lot had to do with their methodology, which was criticized late last week by Johns Hopkins. They used anonymized cell phone data to track people who went to Sturgis in the time frame of the rally, and then looked at the rate of COVID positives in their counties when they went back. Correlation is not causation. As Johns Hopkins noted, any increase seen may have been due to a higher likelihood to get tested for COVID because of Sturgis, detecting cases that were not due to spread in Sturgis. Hopkins also noted that to call a county “high”, they cherry picked the county comparisons. A county in Arizona was compared to a county in Maine, despite totally different demographics and social distancing practices and restrictions, to call one “high” and thus “Sturgis related.” There are 290 total cases identified among attendees of the event–and again, people looked. They also looked at Sturgis itself. As we mentioned, a huge swath of a small town got tested once the rally was over. There were more coronavirus cases than the baseline before, but there were other parts of South Dakota with much higher case rates.
If the rally was a super spreader event, especially associated with 260,000+ cases the economists in this “released” paper claim with $12 billion of health care costs, you should have been able to find more than 290 of them, and a lot higher positivity rate in Sturgis itself.
This is why the paper was “released” and not “published.” This would have never survived peer review. The claims are wildly overstated versus the actual documented cases when others looked for them–and looked hard.
Perhaps the best example of the method flaw of this paper is an anecdote from some of the folks I know who went to Sturgis. One of them works in a small office of about 7 people, in pretty close contact. This person went to the rally, and while there, got a call that 5 of the people in that office had all tested positive for COVID. The person who went to Sturgis did not get SARS-CoV-2; in fact, going to Sturgis may have kept them away from the outbreak back at work! But when they got back, their anonymized cell phone data now localized to their home county. Since those 5 cases in the office occurred within the window of the Sturgis rally, the method used in this paper would attribute all of them to the Sturgis rally. Although there is clearly no epidemiological link.
Do not rely on the lay media to know the difference between a paper “released” and not “peer review published” (because its method sucks)–especially if there is a good headline to be gained from NOT knowing the difference.
As a bonus, you get another nice headline from Hopkins having to come out against “overinterpreting” the report you breathlessly and uncritically covered just a few days before.
–Saw the claims by an alleged defecting Chinese scientist with evidence that SARS-CoV-2 was modified in the Wuhan laboratory, and may have been deliberately released. I have no idea. And most of it has been ripped off the internet already before I got a chance to take a look at it. Again, deliberate release seems unlikely–it got an awful lot of people sick in China (and probably more of them than we will ever know). It has not done much for China’s standing in the world. If anything, accidental release is just more probable. Did they modify coronaviruses as part of studying them? Quite possible.
I still question what really changes. Are we more suspicious of China if this is true, or at least strongly appears to be true? More companies are already cutting ties to China because of the multiple reasons to be a little suspicious of the Chinese Communist Party. H&M just cut off its Chinese manufacturer today and announced it will refuse to buy cotton sourced in Xinjiang province, where China has been quite credibly accused of serious abuse of the Uyghur minority there. Do we get the last 7 months back? SARS-CoV-2 goes away if true, and we get a complete do-over on 2020?
Don’t get me wrong. It’s an important story if true. If the defector has got a smoking gun, I laud her courage and resourcefulness getting it out of that country. But, after the Uyghurs, it’s really just another brick in the wall for that government, isn’t it?
–Finally, your chances of catching Ebola this week are equivalent to your chances that AI is really going to take over the world, annihilate the human race, and begin the Age of Artificial Life given the success of this: https://time.com/5497251/wildfires-artificial-intelligence/
–Your chances of catching SARS-CoV-2 are equivalent to your chances that AI is really going to take over the world, and is already acting on its plan to annihilate the human race, to begin the Age of Artificial Life, given the “success” of this: https://time.com/5497251/wildfires-artificial-intelligence/
<Paladin>