We’re starting to get enough data to draw some conclusions. TLDR: COVID is dangerous – 4.5 times more deadly than the 2017 Influenza strain, which was a bad one. With the caveat that the long term studies are still underway for a lot of at-risk populations, COVID itself is about 215 times more deadly than the vaccine. The COVID vaccine isn’t really significantly more dangerous than the Influenza vaccine. Here’s a bit more detail and context …
There is a lot of argument and discussion over the relative risk of COVID vaccines, especially in Europe with the reports of the AstraZeneca/Oxford vaccine potentially causing blood clots in some people, and the Polyethylene Glycol (PEG) in the mRNA vaccines causing anaphylaxis (allergic reactions) here in the US. Both are concerning – and there is an urgent need to figure out why certain people are more vulnerable to adverse reactions than others. Certainly those with known allergies should be very careful to check the components of each vaccine before receiving it – the CDC publishes guidelines for this, and if you have sensitivities check with your Doctor before getting a shot (or any) procedure. This is the dilemma of vaccination: it’s best for the vast majority of people, but can be dangerous for a few. But care must be taken not to blow that true statement out of proportion.
Chances of dying from Influenza (2017 H5N1 strain): 1 in 740
Chances of dying from Influenza Vaccine: 1 in 100,000
Chances of dying in any Accident: 1 in 1,350
Chances of dying from Gun Violence (you are a criminal): 1 in 3,000
Chances of dying from Gun Violence (you are not a criminal):1 in 220,000
Chances of dying from a Weather or Earthquake Hazard: 1 in 2 million or so
So in context, the vaccines are not risky compared to the disease – and certainly not compared to dying in a car accident (1 in 6000 or so). There has been some reports and talk that the COVID vaccines are significantly more dangerous than the Influenza vaccines. That’s a bit hard to judge. For one thing, the COVID vaccines are being scrutinized in a way the Influenza shots have not been. But even given that, the raw numbers show that the potentially associated mortality rate is about 2.8 times higher. It’s likely that difference would disappear if similar tracking were in place, but even if true isn’t bad. So the “50 times more side effects” stuff you see circulating is overblown.
In reading “news” stories lately, not to mention various comments in social media about topics ranging from politics to COVID vaccines, I was struck again by the power of binary thinking, as well as how perceptions are manipulated by asking (and answering) the wrong question. Another frequent related problem is making assertions that are perhaps true, but presented out of context in such a way as to create a false perception. This usually results in the two “sides” talking past one another and a shouting match ensues; there is no shared worldview to even begin a discussion.
Here’s a concrete example regarding vaccines: In skimming a discussion about mRNA vaccines it was said by one advocate that there is no evidence or “mechanism” they cause birth defects. The problem is, that’s “true” as far as it goes but also misleading. Pregnancy was a specifically excluded condition during the trails reported so far, and all of the documentation submitted to the FDA said it was not assessed. As for mechanism, there are in fact several potential mechanisms where something could go wrong, given the rapid and complex cell division that occurs during the early stages. Is it rare? Possible or impossible? Probable? Likely? We just don’t know – there is no evidence. Last time I looked at least 18 people had become pregnant during the trials and are being closely monitored, but that’s a very small sample size, and until the children are several years old, it can’t be said for sure that there were not problems. It was also said no long term side effects have been reported. That is true but highly misleading: the vaccines were only developed less than a year ago, so there hasn’t been enough time for any long term effects to develop or reach a statistical threshold. So therein lies the problem – saying “there is no evidence” when there have been very limited (or no) studies is absolutely not the same thing as saying “there have been detailed studies an no problem was found.” That’s a distinction that is lost on many people.
For the record on this subject, here is what CDC says as of 7 January 2021: Based on how mRNA vaccines work, experts believe they are unlikely to pose a specific risk for people who are pregnant. However, the actual risks of mRNA vaccines to the pregnant person and her fetus are unknown because these vaccines have not been studied in pregnant women. We know COVID19 presents risks to pregnant women, so if in a high risk group (like a health care provider) it might make sense to be vaccinated with an mRNA vaccine despite the unknowns. Work from home and sensible about social distancing, etc? Maybe best to wait. It’s not an easy call, based on an objective view of the available data.
Again, this isn’t to be anti-vaccine. There are rational risk-benefit arguments for some, and over time as more data is collected and if the early results hold up, increasingly large segments of the population to take these vaccines. What bothers me is that people present it as a binary, “no brainier” choice. It’s just not that straightforward and it is hubris to assert that it is.
Unfortunately there is no shortage of hubris, exaggeration, and binary thinking in order to sway opinions in our public dialogue these days. I could cite many examples, from election fraud (it probably didn’t impact the results, but that’s not the point: the US election system is broken, with deep structural flaws such that it doesn’t meet standards it imposes on other countries), to social debates like LGBTQ issues or abortion or climate change or …
In short, it takes objectivity and careful analysis to reach good conclusions. This is especially hard given the political parties benefit from a sharply divided electorate, advocates for various issues minimize or are even blind to potentially adverse consequences, and demand you “take a stand”, and of course the media industry profits from the noise and drama all that creates. Please don’t feed that process, and try to understand that many situations are not sound-byte simple.
In short, life is complex. Don’t fall into the trap of absolutes.
TLDR: The 27 March 2020 forecast all population symptomatic case fatality rate based on the Diamond Princess cruse ship data was 1.71%. The current (20 January 2021) US symptomatic case fatality rate is 1.67%. So while the public perceptions and actions have been shifting, the bottom line hasn’t changed that much, and while there was a lot of uncertainty back then, the early work wasn’t bad. Here’s some more background, including a rant about people who think people aren’t dying from this, and a look back at the mortality forecasts made early last year.
Our first solid look at COVID19 in a controlled environment was the cruse ship “Diamond Princess”. Almost exactly one year ago, on January 20, 2020, the ship departed Yokohama on a three hour tour, um, on a round trip tour of Southeast Asia timed to coincide with the Lunar New Year celebrations. A single passenger from China brought with him a hitchhiker: the SARS-COV-2 virus. Over the ensuing weeks (which included some dumb measures by the Japanese Government that made things worse), something like over a third of the passengers and crew are thought to have contracted the virus. Around 400 of the passengers and crew became sick enough to be classified as symptomatic, nearly 200 of those were hospitalized, and about 10 or 12 had a primary cause of death being from the disease which is now called COVID19.
So where do things stand? I again rant that real time death counters seen on “news” outlets are disgusting displays of much that is wrong with American “Journalism.” Let’s look at the National Center for Health Statistics data, which is probably the most authoritative/reliable source. These statistics are updated weekly; those of you with wonky tendencies may want to read the technical notes, but the bottom line is that mortality statistics take time to compile in the US. Nationally, only 60% of death records are submitted to NCHS within 10 DAYS of death! Lets ltake a look at the week of December 26th. As of the 14th of January (two weeks after), the number of reported deaths were 41,796. As of yesterday that number is up to 72,710 … now, that’s a worse than usual example due to the holidays, but it shows the dangers of relying on the “real time” data. The Johns Hopkins data sets that many in the media are using are quite good, but they are not definitive, and the sensationalist abuse of this data is not helping. Again, pandemics are “slow motion” disasters. They rarely evolve from hour to hour or day to day, it’s more of a week to week process, with the decisions of millions of individuals influencing the course of the outbreak. The hype is stressful, distracting, and given the politics, divisive, as it encourages people who think they are being railroaded to believe there isn’t a problem.
But there is. I’ve posted similar graphics before, but here it is again, updated through yesterday. The blue line is shows total reported deaths. The orange line is expected deaths based on the US population at the time. The yellow line is “non-COVID related deaths” as classified by the ICD–10 codes. The “expected” line is wavy because more people die in winter than summer – usually due to Influenza and Pneumonia. You can very clearly see that the 2017-18 influenza season was bad, and that 2018-19 was mild. The 2019-2020 season was pretty normal – until in early February something drastic happened. That “something” was the SARS-COV-2 virus. Even if the only data you had was the blue line – reported deaths – you’d know that there was a new disease stalking the land. Look carefully at the orange line – non-COVID deaths – and how high it is during the summer. Contrary to what some with an ax to grind are saying, it seems that we were more than likely under-counting COVID deaths rather than over counting them, although it is possible that some people died who would not have otherwise because they did not seek medical care out of fear, or some other indirectly related causes. In any event, reported deaths are obviously well above normal, there is obviously some new disease running through the population, and anyone who is saying otherwise is simply wrong.
Looking back at the early mortality forecasts, my own forecasts were off by a factor of two – the estimates from late March/Early April were on the order of 200,000 by now, but reported US fatalities recently topped 400,000. I had assumed average people would demonstrate more common sense than they did (yeah, you can rarely go wrong assuming people are idiots but despite being on doomwatch I try to be optimistic 😛 ). I also thought the initial reaction would be stronger nationally, and widespread masking start earlier, than how it played out. I never really thought the vaccines would be ready by now – and I’m still rather pessimistic it will have the impacts the vaccine chorus is singing. First, this is a corona virus: other beasties of this type are responsible for about 20-30% common cold cases, and they mutate so rapidly the immune system can’t keep up; it’s unrealistic to expect vaccines to keep up, although it’s also wrong to discount them, as like with influenza vaccines, they can provide some protection even against unrelated strains. Second, despite the PR deluge, the efficacy hasn’t really been statistically demonstrated to the usual standards, and the quality control and massive rollout have created problems that have harmed the process. Cheerleading and glossing over things like adverse reaction rates isn’t really great way to build confidence. As I have said, I am absolutely in favor of vaccines – but I’m also in favor of a careful, “first do no harm” approach to public policy that a rushed “do something NOW” process rarely allows. Appearances do matter, but the data matters more.
When I teach emergency management, the very first thing I try to get decision makers to understand is that no matter what they do, they are going to kill people. There is almost never such a thing as “erring on the side of caution” because all actions have consequences – and as I often point out, economic harm also causes physical harm, a fact that is often overlooked. Ultimately the trick is to figure out policies that will cause the least harm in the long run.
As the COVID19 pandemic shows, that’s a very difficult thing to do.
My guess is every engineering student since the mid 1940’s has had to watch the video of the Tacoma Narrows Bridge collapse. Other professions have similar cautionary tales of hubris, short cuts, innocent mistakes, misunderstanding of nature, and things that went terribly wrong. They are often clear in hindsight, but lost in the immediate noise and pressure of crisis decision making. Often progress requires risk – but the two require a rational balance. Now that a number of potential vaccines are approaching approval and distribution, everyone from government officials to individuals are soon facing difficult decisions: who gets what vaccine, when or if to take it, and what level of persuasion (or even coercion) should be used to get people to take them. Yes, these are difficult decisions, and to be blunt anyone who says it’s a “no brainer,” simple, or obvious is either fostering an agenda, being disingenuous, or doesn’t know what they are talking about. These questions are even more difficult because several of the potential vaccines on the verge of distribution are using technologies that have never seen wide spread distribution and use. Compounding that are issues of politics, National pride, and commercialism. The details are complex, and most of the simplified explanations I’ve seen (and more than a few technical ones as well) are biased either towards “trust us; don’t worry” or “it could be a beaker full of death.”
As with most issues, the truth lies well within the extremes, but the the decisions are ultimately fairly straightforward. We have to weigh the consequences of COVID19 (to society and the economy as well as physical health) against the effectiveness and risks (known and unknown) of the various vaccines. Given the complexity, both advocates and detractors (some quite vocal) among the general public really don’t fully understand how any of the vaccines work or their implications. For the vaccines under development, there are four broad classes (link goes to Nature article with good graphics). The major vaccines that have been approved or are closest to certification are in three classes: killed virus (the Chinese Sinovac), viral vector, such as those based on the Human adenovirus (like the Gamaleya vaccine) or Chimpanzee adenovirus (AstraZeneca/Oxford); and those based on nucleic acid – the mRNA vaccines (Pfizer, Moderna). The killed virus approach is how most current vaccines work. The viral vector vaccines are fairly recent, but there are a few that have been approved and in use for over a decade. The mRNA vaccines are substantively different. And it is here we have a bit of a problem.
mRNA vaccines and related technology have only seen small scale experimental use, usually in the context of cancer or other deadly diseases, and have never before been certified. In theory they should be safe, perhaps even safer than traditional approaches, but there are some potential risks and more than a few unknowns. At least some viral vector based technology has been around for 17 years and Gameleya, for instance, has a number of vaccines (including an Ebola vaccine) that have been approved and in use for years. For those vaccines there have also been long term studies as to adverse reactions. So while the actual vaccine for SARS-COV-2 is new (since, obviously, the virus is new), the vaccine methodology itself isn’t – in fact, one was in development after the SARS-COV-1 scare back in 2008. For the mRNA vaccines there have been no long term trials or monitoring, and no previous vaccines approved for human use based on these technologies. Another factor is that the various studies are not using consistent criteria and methodologies. This is a distinction lost on many in the media, such as the NY Times article last Sunday discussing the commencement of distribution in China (Sinovac), Russia (Gamaleya’s SputnikV), versus Great Britain (the Pfizer vax). The FDA briefing materials for the Thursday Pfizer approval meeting also glosses over this issue, but is clear there are a lot of unknowns.
With all of the vaccines, there are unknowns, and always some side effects. How effective is it in the real world? And by “effective”, be very careful how that is defined – some of the criteria in the current Phase III studies seem like pretty low bars compared to past studies. Do they minimize transmission, protect the person inoculated from getting sick, minimize (but not prevent) symptoms? How long does immunity last – and how do the inevitable mutations impact effectiveness? But the ultimate question is do the advantages in reducing the consequences of COVID outweigh the potential side effects of the vaccine – especially given the need for mass inoculations of 80%+ of the population? There is little doubt that the benefit/cost radio of vaccination will be in favor of vaccination if – and it’s still a big IF – they work as the early results indicate, and adverse reaction rates are similar to other vaccines. But can we assume that? Most advocates are. With the mRNA vaccines there are additional unknowns as to the long term impact of the underlying delivery methodology. Again, in theory it should be as safe, and perhaps even safer than more traditional approaches using a live, killed approaches. But we just don’t know. While the short-term impacts seem low based on the early trials, they are just that: short term, relatively small scale studies (weeks to months, thousands of people). No one knows what, if any, the long term consequences of this kind of vaccine as a whole might be when applied to the general population over time. It’s never been done before. There is no reason at the moment to suspect there is some hidden gotcha – but the designer of the Tacoma Narrows Bridge didn’t think it would rip itself apart either.
Opinionand Conclusion: As I have said before, COVID-19 falls in a gray area. If it were causing wide scale deaths across a wide range of groups, and mRNA vaccines were the only viable option, then maybe we would need to “roll the bones” with a new technology. And yet, COVID is bad. But for all the disruption and pain COVID has caused, it’s not smallpox or the black death, much less the FGC-347601 virus Dr. McCoy had to deal with (and recall in that tale, the disease itself was actually an unforeseen side effect of a noble original objective).
Make no mistake: mRNA therapies have the potential to be a major innovation in the treatment of all kinds of diseases from cancer to COVID. But while pressure often leads to rapid advances, we’ve also seen far too often in the history of technical advances that the temporal or economic pressure to do something now leads to catastrophe. Tacoma Narrows. Challenger. Mars Climate Orbiter. The pharmaceutical industry isn’t immune either: Thalidomide, or more recently Vioxx. Darvocet. That said, don’t fall into the anti-vaxxer trap of seeing every adverse reaction as evidence of a conspiracy. It’s painful to say it but a certain adverse reaction rate is acceptable in light of the impacts of the pandemic. And, yes, Big Pharma sees this as a huge money making opportunity, and that requires monitoring and regulation. But Pharma have done a lot of good work, and in the present system, being profitable is how things get done.
When you weigh all the factors, in my opinion it is simply too soon for the wide scale application of some of these vaccines in relation to the risk from COVID (again, not to minimize that risk). Much of the harm of this virus has been self inflicted – a coherent global response would have cut the economic impact five fold, and the death toll by a third in my estimation. If everyone would just behave responsibly, between mitigation and other measures we would have some time to sort this out. Let’s take a deep breath, proceed cautiously, roll out the various vaccines in a reasonable way and not get hung up on national pride (noting some vendors in China have an unfortunate reputation), or commercialism, while moving expeditiously to apply new technologies in parallel as they are validated. The rollout of the new, untested vaccines can and should be spaced over several years. Supply chain issues may force that in practice anyway, but that should have been the plan from the start. And foreign developed vaccines – properly vetted for safety, without nationalistic biases – should be allowed in as part of the mix. The Gamaleya vaccine is likely a prime example. But even the vaccines based on established methodologies need more testing.
I think it is reckless to push the wide spread distribution of novel vaccines on tens of millions of people until there is a longer safety and performance baseline. As noted above we have absolutely no idea what the medium term (2-3 years), much less long term (5-10 years) implications are with respect to adverse outcomes for some of these approaches. It is especially reckless where there are several candidate vaccines with more well understood risks. Maybe the mRNA based vaccines are fantastic, but not only do we not know, we don’t even have the data to know and won’t for several years. To coerce hundreds of millions of people to take these vaccines in an experiment of this magnitude is simply unethical. Should ever increasing numbers try it? Of course – with an appreciation of the risks, under careful supervision and long term monitoring. Those at highest risk? Absolutely – although I’d be careful with otherwise healthy members of the health care community upon which we depend. Tens or hundreds of Millions? It’s just too soon.
Addendum: I’m not a physician, but I do understand a lot of the issues surrounding this at a fairly detailed level More importantly, I am pretty knowledgeable (some would say expert) in emergency management decision making and how things go wrong in complex scientific and technical processes. To be absolutely clear, I’m not advising anyone not to get one of the new vaccines. I think a lot of people probably should get them. At the moment there is no reason to suspect there is anything wrong with them. In fact many if not most of the fears of the mRNA vaccines are way overblown. But that doesn’t mean it isn’t rational to have some concerns, and saying “we don’t have any reason to suspect there is anything wrong” is VERY VERY DIFFERENT from saying “here is a 5 year followup study that shows nothing is wrong”. Read the actual FDA briefing materials to see how often the word unknown is used. There are lots of competing blog posts and opinionating on all of this, and much of it lacks nuance. Don’t be stampeded into one position or another out of fear. Fear is the mind killer …
With the late season storms both here and in the West Pacific, and the developing catastrophe in Nicaragua/Honduras, haven’t formally checked in to see how the virus is doing until today … Yep, the virus is doing fine. Humans? Not so much. True, it’s not a Monty Python style dystopian “bring out your dead” kind of pandemic, but a lot of people are still passing away from this thing who would not have otherwise died. How do we know this? Forget the death counters popular on TV. As I have discussed before, the absolute numbers aren’t nearly as important as the concept of excess mortality – how many people are we losing who wouldn’t have died otherwise? For some more background on that take a look at this post. For those paying attention let’s jump right to the numbers. Here is the overall US chart for deviations in mortality over the last four years, as of the last week of October. Above average is above normal, below zero is below average. No, the numbers aren’t any more recent than the end of October. I’m so tired of ranting about the craptacular public health data reporting system in this stupid country, a system that is even worse than the stupid election system that can’t manage to count live ballots any better than it can dead bodies – the gallows humor there writes itself these days.
So it’s absolutely, unambiguously clear: something is killing ‘Muricans this year at greater numbers than past years, and it’s pretty clear it’s the SARS-COV-2 virus that causes COVID-19. And it’s not “just the flu”. “Just a bad flu” is what that spike in late 2017/early 2018 is. No, it’s not “Spanish Flu” bad, much less the Black Death, but it’s bad enough. Even correcting for the mild 2019 influenza season, (which is partly responsible for the early spike in COVID deaths – vulnerable people who would have died in 2019 lived into 2020 to fall victim to COVID instead), COVID has really distorted the mortality statistics.
What about the State of Georgia? Here’s that graph. Note with ha smaller sample size, it is “noisier”, but clearly the same story …
Before anybody says “oh, it’s getting better!” Remember these numbers are a couple of weeks old, and the lag between infection and death is around 4 weeks, so this is maybe 6 weeks behind the curve. The last few entries are certainly low as it can take four weeks or so to collect all the mortality data (insert primal scream here).
This graph looks like the normal cycling of a mostly out of control virus, where people notice it’s bad, react, it drops some, then they get complacent, and it rebounds, as well as the fact that we are seeing the virus move in to different populations in different areas. The other problem is that we are entering the normal respiratory virus season, and, flawed as they are, the other metrics – case counts, hospitalizations, positivity rates, and so forth – are all trending upwards. So it’s likely these numbers are about to trend higher.
Again, the problem with SARS-COV-2/COVID19 is that it’s bad – but not bad enough. It slots nicely into a place that scares some people in to overreacting, and others into under reacting, exacerbating existing fault lines in society depending on where you fall on the security/freedom and personal/collective responsibility prioritization scales.
So what do we do? Mostly it’s common sense. But that is in remarkably short supply. The problem is a critical mass of the population across the country (and even world)has to act responsibly. Otherwise the slow burn – punctuated with flare ups – will continue. And with flare-ups politicians will feel forced to “do something” dramatic, most likely things like shutdowns and restrictions which won’t work in the long run, but will further the social and political divisions, not to mention the incredibly fragile economic situation. An interesting question arises: If the mortality rate settles in to a new, higher value, say 20-30% above the previous average, will people ultimately just accept that and get on with life? It’s going to be interesting to watch the media coverage with respect to the statistics as the likely change in administration progresses. Will things actually be better next year, or will they just seem better with an (on the surface anyway) more coherent approach and a vaccine? When will the media give up on coverage and move on to other stories? Hard to say. Those are all issues just as important – maybe more so – as the biology and epidemiology of the virus itself.
I’m also very afraid that the vaccine won’t be the deus ex machina that people are hoping it will be. For starters, the 90%+ effectiveness reports are unlikely to be seen in widespread use. Those number always come down once things move in to general use, so there’s an expectations problem building. There’s also a fair enough chance one or more of several potentially unfavorable scenarios will come to pass – not the least of which will be that in the rush to get vaccines out, long term adverse reactions will start to crop up in six months or a year once widespread vaccination takes off. The other is potential risk is that immunity will decline rapidly and be seasonal at best. Great for the bottom line of Big Pharma, probably not so good for the rest of us.
Sense some frustration here? Yep. COVID19 long ago stopped being a mostly scientific problem, and after the behavior of both political parties in the US the last few years, only a hard core political activist affiliated with one of the tribes can be optimistic (aka delusional) about all this. Those of us in the real world will just have to continue to suffer through their shenanigans and try to keep out of the way …
As I frequently rant, various phenomena have their own pace, or temporal domain. With hurricanes, it’s about 12 to 24 hours. With a pandemic, it depends on the disease, but for COVID19 it’s about two weeks. You can’t force that into some other paradigm like a 24 hour news cycle, or political campaign, or whatever. It just doesn’t work, and leads to stress, confusion, and bad decisions. I’d add that each phenomena also has a natural spatial domain or geographic region that makes sense to analyze. Those geographies don’t usually match our arbitrary political geographies. For example, what sense does it make to have a hurricane warning that extends to the Savannah River, “forcing” the evacuation of Hilton Head, when you can stand on the north end of Tybee under only a tropical storm warning, with no evacuation in force, and look across the river/sound and wave at the folks fleeing in terror. OK, a little exaggeration, but not much sometimes. With the pandemic, given different approaches (masks vs not), people moving back and forth (Effingham being a “bedroom” community of Chatham as a local Savannah GA example), not to mention the inconsistencies in reporting, it’s hard to define a spatial domain that works. The point, in scientific terms, is that the temporal and spatial characteristics of some phenomena doesn’t always fit with our arbitrary temporal and spatial definitions. That leads to a lot of noise that – you guessed it – people with agendas can exploit to say “things are great!” or “things are terrible!” and, based on the data, they are both somewhat right and somewhat wrong.
But for those of us who care about reality, how are things? Well, the virus is doing fine. Which isn’t good news for us who are foodbags for the thing. I’ve basically given up on a lot of the statistics based on case counts and positive results. It’s just too noisy. The main thing I’m watching closely right now is the excess mortality numbers, and how that compares with both reported COVID mortality and the expected mortality. In other words, how many more people are dying than the longer term statistics suggest should be the case? Now, that doesn’t tell us that COVID is responsible for all of those deaths, it does tell us if something unusual is going on, and from other evidence it is clear COVID is in fact responsible for most of them (a fact that some folks just can’t seem to get a handle on).
Here I’m using the from the National Vital Statistics System. Death reports are subject to delays, so for example today the “latest” data is only as of the end of September, since it takes around ten working days for reports to filter up to CDC – longer for some jurisdictions (yeah, unbelievable, but there it is). Despite all the flaws, this is probably giving us the most complete overall picture of the impact of the virus. So the “temporal” domain for our analysis is going to be number of reported deaths per calendar week. That helps reduce errors caused by weekends, Mondays, and Fridays, as well as just normal day-to-day fluctuations. For the spatial domain, lets look at the whole US, and some selected states. Even that is a bit risky due to reporting differences, etc. but we can see some interesting things at the state level. So here is a plot of the ratio of the number of reported deaths divided by the expected (based on the previous 10 years) number of deaths, per week for the 39 weeks of 2020 thus far.
So clearly, NYC had a Very Bad Spring with a lot of people dying who shouldn’t have. You can try to argue “well, it was older people, or in nursing homes, or people already sick” or some such excuses but the simple fact is that it wasn’t their time. Something (and we know it was COVID) was killing people over and above the usual causes of death. As you look at other states you can see different progressions – a mostly under control slow burn in Washington State, later peaks such as in the South East (GA/FL) and Arizona, etc. For the US as a whole, it’s the bigger population states driving things, but there was clearly the NY (and NJ) crash in the spring followed by a slow burn and second much smaller peak in mid summer. Lower population and rural states are starting to catch up a bit in the stats, but as you can see with a state like South Dakota it’s noisier.
What does it mean, and what does it indicate for the future? The virus continues to do a series of waves through different parts of the country. Now that fall/winter are here, we can expect another “wave” in the higher populated states since a) most people were not exposed to the first wave and b) compliance with mitigation measures is variable. Since states aren’t in isolation, and any “reservoir” of the virus will enable it to get back in to areas with it under control, that seems inevitable. One thing is that a lot of those vulnerable in, say, NYC were taken out by the first wave. However, people age, get sick with chronic diseases, etc., so there are now all new victims waiting to be infected. I suspect we will see “waves” of deaths rippling through the next year – probably no big spikes like the original NEUS spike, but lots of “slow burns” and increases in the rate between 10 and 30 percent above the normal mortality rate in most jurisdictions.
So what do you do? Same as it ever was – mask when encountering someone outside your household in an indoor environment for sure, or outside with prolonged contact; hand hygiene. Get a flu shot if it’s ok for you to do so – for one thing, we really want to not have any more confusion between flu and COVID than necessary, and it may give your immune system a little boost that might help. Try to take care of yourself and stay healthy. Because I doubt there is going to be a miracle vaccine for this thing, and it will just be a fact of life for the foreseeable future …
I would hope everyone is praying for a swift recovery of the President. If there is one thing this country doesn’t need right now it’s more uncertainty and disruption thrown into the Election. As for commenting on the situation, I think it illustrates some vital points to understand about the virus that causes COVID19 and how our society needs to try to react to it. Of course, everything is a mine field given how emotional and split everyone is over politics. So I’m going to try to discuss in a somewhat neutral way a vital issue: the limits of testing. That does mean pointing out how the White House made a mistake about testing. Please don’t try to read between the lines on this with respect to my politics – there is no agenda here. Of course I have an opinion, and I think I’ve decided which one I’m going to vote against, but I’m not happy about it, and to be clear I don’t support either side.
I’ve ranted a lot about the problems with testing here in the US, and what is happening in the White House seems to be a classic case of the misapplication of a technique designed for one thing and using it for something else. The first thing you need to understand is how COVID acts in a person. The New York Times has a surprisingly good article on the progression of the disease today. I think their COVID coverage is still free. They have a lot of simple graphics that do a nice job of (literally) charting the disease; the key graphic is this, with the danger zone of 2-3 days marked:
The problem is that the kind of test used by the White House to test visitors, the Abbott ID Now system, is designed to determine if someone who has symptoms has COVID19 or something else. The advantage of this (vs. the more sensitive PCR test) is speed: you get an answer in 15 minutes vs. overnight at best. But … if you look at the graph, the time of peak viral load (the therefore infectiousness) is generally before the onset of symptoms. Worse, at that stage (between the first pink vertical line and the when the green “symptoms” line starts), the test has a 30% miss rate. So someone could have the virus – and be spreading it – and the Abbott test will miss it about 1 in 3 times. Let’s say you have 200 people in the inner circle. If it’s like the rest of the population in areas of community spread were in the 10-20% exposure range. Taking the low number, that’s 20 people. With a one in three miss rate on the daily tests, and the period of infectious but not symptomatic, you would statistically over a period of a couple months have three to five people wandering around for at least a day spreading the virus. That’s why other measures – masks, hygiene – are essential even with routine testing and temperature checks. No one measure is a silver bullet – but like layers of Kevlar in a bullet proof jacket, taken together they can beat this thing.
I suspect what happened is that the White House didn’t fully appreciate this. They were testing everyone on the staff, felt comfortable with that, and it probably wasn’t clear that this system wasn’t designed as a wide scale preventative measure. It was designed to test people who were already sick. Whose fault was that? Did the manufacturer oversell the product? Did that technical detail get lost in the rush to put something in place? Did they just ignore advice and engage in wishful thinking? All of the above (I’d put my bet here)? To be clear, this mistake isn’t limited to the White House – I know of at least one medical facility that has been doing the same thing (treating the rapid test as exclusionary rather than confirmatory), so they aren’t alone. The bottom line is that it was statistically pretty much inevitable that there was going to be an outbreak within the White House.
It is absolutely vital to realize that in the short to medium term there is no single solution for dealing with this virus. Testing is useful – but has limits. A vaccine will be useful – but will likely have significant limits. Masks are useful – but have limits. Shutdowns at one time could have been useful – but (IMNSHO) are no longer given the cost/benefits. Social Distancing is useful but has limits. Hygiene is always a good idea. Immunity may or may not be a thing. It is utterly insane that the US political system has decided that some of these are “Democrat” solutions, and some are “Republican” solutions.
Numerous potential flashpoints of doom out there … but nothing as of this morning above the “that might get bad soon.”
Tropics: Typhoon Kujira is off of Japan, no threat to land. Tropical Depression 18-E is off the coast of Mexico, again no threat to land. Closer to home (well, mine 🙂 ) a system is moving across the Caribbean that the global models are showing spinning up in a few days as it approaches the Yucatan Peninsula. NHC gives this a 50% chance of forming something in the next five days. Some of the usual suspects are already flogging the potential for the system to spin up. Here is what the GFS model is showing for next Wednesday, a sort of organized depression/minimal storm approaching the Mexican coast, and a second thing trying to spin up behind it …
but … models don’t always do so great in this kind of situation. They are getting better, but 7-10 days just isn’t there yet for anything other than entertainment purposes. A couple of things to keep in mind – note there is no “X” on the NHC map, just a diffuse area where something might form. Second, no discrete model runs or INVEST area ID has been assigned yet. The Tropical Weather Outlook doesn’t have the majik words “interests in <name of some area> should monitor the progress of this system.” So unless you are a die hard weather junkie, you’ve got plenty of other stuff to worry about!
Like the debate tonight between the raging dumpster fire and the older well worn house that looks comforting from the outside but has bats in the attic, rats in the cellar, and an ax murderer living in the spare bedroom.
Or the continuing slow burn of the COVID-19 Pandemic. I posted on this yesterday, and nothing I’ve seen in the last month or so says there is any progress – or significant new threats. As I write this the talking head on the radio news said “we have hit 1 million deaths, one fifth of those in the US.” Which is total bullcrap for reasons I’ve discussed before (globally there is a huge undercount; the US is about 5% of global population and if you take in to account the horrible reporting in most of the world, is about 5% of deaths, not 20%). Guess he doesn’t read this blog. Sigh.
The economy continues to send up flares, red flags, warning lights, and Edvard Munch style screams. But Congress is deadlocked over the aforementioned election thingee, there is no coordinated plan to try to stabilize things, so the ongoing collapse of key aspects of the economy like small businesses continues. The wave of potential defaults is on the verge of becoming a tsunami, and when that hits the over-leveraged capital markets, Bad Things Will Happen.
In the geopolitical world, Donbass, Nagorno-Karabakh, Syria, Greece-Turkey, and Libya all continue to smolder. The situation in Nagorno-Karabakh is especially dangerous and tragic, given the involvement of Turkey in another potential attack on Armenians (which has a long and tragic history). It is one of many complex “frozen” conflict areas like Ukraine and the Balkans that were suppressed during Soviet times, but have flared up since. Why does this matter to you? The various tangle of alliances and obligations can rapidly drag outsiders in. Oh, did I mention oil? Because oil is involved as well … of course.
Oh, and Tampa Bay winning the Stanley Cup? Which sign of the apocalypse is that?
So we wait and see what happens. There’s always stuff to worry about, and it is best to be proactive when we can. But if you have a family emergency plan (always keep a week of emergency food, containers you can fill with water on short notice, and a contact plan), a weather radio, and are taking COVID precautions (masks when going to enclosed spaces, distance, good hand hygiene), you’ve got most of the bases covered, so enjoy life and don’t worry about all the might be’s until they become “probably”s …
It’s one thing for the media to have “death counters” and for talking heads to spend 15 minutes an hour talking about the COVID statistics. Sure, it’s overly dramatic, misleading, causing a lot of unnecessary FUD (Fear Uncertainty and Doubt). But … entertainers are going to entertain. However, an awful lot of people are making life and death decisions based on week to week (even day to day!) fluctuations in COVID19 statistics. Does that make sense? Let’s take a look …
For a start, lets take a look at US Death reports, comparing the estimated numbers with the “final” totals for the latest reporting period, the week of September 12th:
Oops. Looks like there is a problem. First, (looks at calendar), it’s the 28th. The data for the week of the 19th is so incomplete it isn’t even showing up yet, only 25% or so of jurisdictions nationally have submitted data. Let’s put this in bold: we don’t really have solid numbers ( say, no more than 2% missing) for SIX WEEKS. Not six days, much less six hours (or, for TV folks, six minutes!). Here is a CDC paper describing the lags in the reporting system.
Now, let’s look at the reported deaths vs. the expected deaths. For expected we are using the average deaths in the US over 2000-2015, adjusted for current population. It varies from week to week during the year, more people die in winter than summer (mostly due to Influenza and Pneumonia). Here is the plot since February 1st of this year …
Two things are obvious from this plot: first, there is drastic under reporting in the most six to eight recent weeks (but we knew that from the previous graph). The second is that something out there is killing lots of Americans. So those saying COVID-19 isn’t that big a deal, well, that’s not what the numbers are showing – that’s not really the point of this graph, which is to show that the rapid tail off in recent weeks is probably due to reporting issues. We won’t have a good idea of what the data is for this week until November. That’s insane for a so called developed country with computers and telephones and stuff, but there it is.
Essentially all of the numbers you are seeing reported on a daily basis are ESTIMATES – not actual data. Now, the CDC, NCHS, Johns Hopkins and others doing the estimates are trying to do their best, and everyone is trying to get COVID data expedited through the system, but that’s actually a problem because now COVID confirmed deaths are being treated differently than other cases. and due to inconsistent testing and reporting it’s clear that we are missing a lot of COVID related deaths. Why do I say that? Take a look at this:
In other words, either there is another respiratory virus out there killing folks (very unlikely), or we are under counting COVID-19 deaths by around 20% (5 or 6% of total deaths).
And remember anything since week 28 or 29 (Mid August) is incomplete … so don’t get duped by the apparent tailing off since week 29 or so in this graph.
I could grind through this on a state by state basis; some are doing better, others worse, but you get the picture: the data isn’t timely or accurate. This is why (much to the annoyance of some) I don’t get bogged down in what this or that article (or even specific credible study in isolation) is arguing, trying to use the COVID-19 statistics to prove masks don’t work, or do work for that matter, or if the mortality rates are going up or going down or reopening is or isn’t working. Because to be blunt, the data sucks and we just don’t really know other than generally or anecdotally. That’s not to say the data is worthless – certainly we can see trends, and professionals can extrapolate a good bit from incomplete data, but this obsession with the death statistics isn’t healthy. Cases? Forgetaboutit. That’s even worse due to testing, reporting, and societal issues.
All this noise is why you can find an “analysis” out there (some credible, some not) that supports just about any point of view you want to try to flog. But if you take a step back and aren’t trying to make some political point, the picture is relatively clear: the SARS-COV-2 virus that causes COVID-19 is killing a lot of people who wouldn’t have otherwise died, we aren’t counting everyone who is being killed by it, and it isn’t going away.
Of the four systems the US National Hurricane Center has on their outlook (link), only two are very interesting at the moment. The first, just offshore from the Southeastern US, they give a 70% of forming a tropical depression or greater in the next 2 days. The fringes might cause some winds and waves in the Northeast and Canadian Maritimes and Bermuda, but any storm that forms is forecast to stay offshore. The second system is in the far southern Caribbean. It two is tagged at 70% over the next 2 days, 80% by day eight. Most models dissipate it, but a few have it as strong as a tropical storm making landfall on the Nicaragua or Belize/Yucatan coastlines in 5 days. In the West Pacific, Typhoon Maysak is sideswiping Okinawa today, and is projected to make landfall dead center over South Korea. Here is a track overview:
JTWC has backed off on the intensity a lot since yesterday, if it continued as forecast, it would have been an $80 Billion storm; now that is down to $37 Billion. I suspect it will probably end up around $10 Billion in economic impacts – big enough for sure, certainly a risk to life from mudslides and flash flooding. The misery a storm causes is often not linked to the dollar value, something to keep in mind about Louisiana after Laura …
Like so many aspects of this situation, where you get your news probably governs what you think of the situation. So, did only 6%” of those listed in the U.S. coronavirus death toll actually die from COVID, or are only 5% people who would have died anyway? As usual, it’s complicated, and both points of view are both right and wrong and easily exploited by those who have an ax to grind. Most people who die of chronic diseases of some kind have multiple factors that would have killed them at some point, and these are related – if you have a heart condition, it makes diabetes more dangerous, and vice versa. Which one gets credit on the death certificate in any specific case is often subjective. In simple terms, the SARS-COV-2 virus that causes COVID-19 does in part is attack people in such as way that if there are any underlying problems it makes them worse. It also causes the immune system to go in to hyperdrive in some people, causing the bodies defense to turn on itself. What epidemiologists look for in a disease like this, or influenza for that matter, is excess mortality. How many people died who, despite the other conditions, would have made it through a given time period had they not contracted COVID19? That is where the Excess Deaths statistics are the place to start.
One problem with all this kind of analysis is it’s a moving target. To state the obvious, each year people are born, move through various stages of life (and therefore vulnerability to diseases like Influenza or SARS-COV-2), and die. The way the numbers are reported for COVID19 are really confusing and misleading on so many levels. We never do overall death counters for flu. It’s X died in a week, or in a season, and that is compared with the average mortality rate by cohort. Since this is a new virus the counter is starting from zero, but as we continue into the COVID19 outbreak, if we don’t reset the counter or start thinking in terms of excess mortality, it will exaggerate how bad it is. Over the last 20 years flu season numbers have ranged from 30k to 120k. This new virus this year will be high in year one because it is attacking a lot of people who were vulnerable, and the final number in the US will almost certainly be at least 300k, but in year two the vulnerable population will be smaller. Where will the COVID19 number settle? In the US, probably in the 200k range per year (absent a vaccine or better therapies of some kind).
So what does all this really mean? As noted yesterday, COVID19 seems to be three to four times as deadly as a bad influenza outbreak. It’s sneaky, because fewer average people get obviously sick, yet more vulnerable people die. The arguments over the details of this or that statistic in isolation generally miss that bigger picture. And don’t forget, while you may think you are in the less vulnerable group today (be aware lots of people have underlying conditions and don’t know it!), you might be in that group tomorrow, and certainly will be at some point in the future as you age. So what do we do? You know by now: mask up out of your bubble, good hygiene, take care of yourself healthwise. Duh. Off to do some treadmill …