Shoddy statistics and false claims: Dr. Erickson dangerously misled the public on coronavirus

By now, you have likely seen the viral video of two doctors in Bakersfield, California (Dan Erickson and Artin Massihi) holding their own press briefing in which they argued that COVID19 is no deadlier than the flu, shelter in place orders are doing more harm than good, and schools and businesses should re-open. Clips of the press briefing have rapidly been latched onto by many people for a variety of reasons ranging from political leanings to desperation for hope. Unfortunately, these doctors have no clue what they are talking about, badly blundered the statistics, made numerous false claims, and have enormous financial conflicts of interest (i.e., re-opening businesses would be tremendously financially beneficial for them). In short, they are emergency doctors, not virologists, microbiologists, immunologists, or epidemiologists, and they should leave the statistics to people who are properly trained to analyse them. To demonstrate that, I’m going to go through their nonsense point by point, starting with what I consider to be the core issues. Please note that I will only be discussing the science and logic behind sheltering in place, not the politics.

Note 1. I based this post on the entire 1+ hour interview, not the short 2–5 minute segments.

Note 2. It’s worth mentioning that the American College of Emergency Physicians condemned their statements.

Bad statistics

Incorrect mortality rates

Their entire argument rests on the notion that the mortality rate from COVID is actually very low, even less than 0.1% (roughly the typical mortality rate from the flu). Actual studies have found that the mortality rate varies from  3.6% (Baud et al. 2020) to 1.4% (Wu et al. 2020). I have yet to see an estimate based on confirmed cases that was anywhere near the number these emergency doctors came up with (see Note 3). So how did they get such a low number? Easy: they’re bad at statistics.

To get their numbers for a given county, state, or even country, they took a series of simple steps. First, they took the number of tests that had been conducted and calculated the percentage of positive results. Next, they “extrapolated” that by applying that percentage to the entire population of the geographic region in question to calculate the total number of positive cases. Finally, they divided the number of deaths by their calculated number of cases, and lo and behold, the death rates were low, way lower than actual epidemiologists have calculated (see example in Note 4). Why is that? Anytime you see one or two “experts” present a value that is vastly different from what all the other experts have arrived at, you should be suspicious, especially if they announce their findings in a blog, press conference, etc. rather than the peer-reviewed literature (where real scientists present their findings).

In this case, there are two glaring problems with their analyses. First, you simply cannot extrapolate the percent of positive tests to the entire population because it’s not a random sample. Imagine, for example, that we have a bag with 1,000 marbles some of which are black and some of which are dark blue. We don’t know how many of each there are, so we reach in and pull out several random handfuls and count them, and we find 50% black marbles and 50% dark blue marbles. From this, we’d conclude that there are roughly 500 black marbles and 500 dark blue marbles in the bag. That would be a fine extrapolation, because we took a random sample. Now, however, let’s say we can see partially into the bag. It’s a bit dark so we can’t always tell the color of the marble for sure, but we deliberately select the marbles we think are dark blue. From this, we find that 20% are black and 80% are dark blue. Can we conclude that 800 of the marbles in the bag are dark blue? Obviously not! We clearly took a biased sample, which means we can’t extrapolate from it. This is experimental design 101.

Coronavirus testing thus far has been a very biased sample. It has not been truly random sampling. Rather, it has been heavily biased towards people who had symptoms, people who were in contact with someone who developed COVID19, people at high risk, etc. In other words, the percentage of positive cases in the testing is probably much higher than the actual state or country-wide percentages, just as our estimate of dark blue marbles was unrealistically high. This means that our intrepid doctors overestimated the total number of cases, thus vastly underestimating the mortality rate. They calculated mortality by dividing known deaths by their estimated cases, which means the higher the number of estimated cases, the lower the death rate.

The other problem is that they are only using the people that have died thus far, but that number is going to keep going up, even if no one else ever becomes infected. In other words, some of the people who are infected with COVID right now are going to die. So, you can’t take the ongoing infection data and divide the number of deaths by the number of cases, because people are still dying. That number is going to keep going up. To illustrate, let’s say that we have 10,000 currently infected people, plus another 10,000 who have either died (100) or recovered (9900). It would be stupid to take those deaths (100) and divide by all those cases (20,000) and conclude that there is only a 0.5% death rate. We can’t do that because we still have 10,000 people who are infected, some of which will die. Do you see the point? Using these numbers midway (as they did) biases the results towards a lower death rate.

These very basic problems with their analyses completely nullify their results. The numbers they are basing their arguments on are invalid, which means that they have nothing to back up their claims.

Note 3: Both the number of confirmed cases and the number of confirmed deaths are almost certainly large underestimates, but since we don’t know those values, it’s hard to know what the true death rate is. That does not, however justify the type of shoddy statistics they used. Also, there have been several recent prevalence estimates based on antibody tests that argued for a much higher disease prevalence than is currently documented, but the estimates thus far have been riddled with problems (including non-random sampling) that are beyond the topic of this post.

Note 4: To work through the math of one of the examples they gave, at the time they collected their numbers, California had done (according to them) 280,000 tests, with 33,865 positives, giving a prevalence of roughly 12%. They then assumed that 12% applied to the entire state and multiplied 0.12 by 39.5 million (CA’s population), resulting in 4.7 million calculated cases (according to them; they didn’t round correctly). Now, if we just divide the current number of deaths (1,227) by that number and multiply by 100, we get a mortality rate of 0.03%. (I have not checked their numbers as far as number of cases and such; I’m just reporting their shoddy math).

Sweden vs Norway

A core piece of these doctors’ argument is that sheltering in place doesn’t work. They claim that there is peer-reviewed evidence to support this (I haven’t seen any), but the only “evidence” they present is a comparison between Norway (which shut things down) and Sweden (which did not shut things down officially, but still many work from home, engage in social distancing, etc.). The numbers, according to them (which seem at least close to correct for a few days ago) are as follows: Norway (shut down) has a population of 5.4 million and has 7,191 confirmed cases and 182 deaths. Sweden (not shut down) has a population of 10.4 million and has 15,322 cases and 1,765 deaths. According to them, those death rates are “statistically insignificant,” a term that they clearly don’t understand. You can’t just eyeball the numbers and assert that they aren’t significant. You have to actually do some statistical tests. Based on those numbers, Norway has had 34 deaths per 1 million people. In contrast, Sweden is all the way up at 170 deaths per million! If we do an actual statistical test (chi square), that difference is, in fact, highly significant (P < 0.0001; this means that there is less than a 0.01% chance that a difference this great or greater could arise by chance; more on stats here). Again, these guys are emergency doctors, not statisticians or epidemiologists. They are talking about things they do not understand.

Now, to be clear, that comparison I just made is not great either. I only made it to show the absurdity of their claim that those numbers aren’t different. The reality is that there are tons of differences between those countries that make it very difficult to make such a comparison. For example, not only did Norway lock down, it also has done substantially more testing than Sweden which can also have a huge effect (it’s also why the number of cases per million is similar between the countries even though the per capita mortality rate is so different). Any sort of country comparison like that is inherently problematic, particularly if you have a sample size of 2 countries. Also, a better approach is to look at trends, not snapshots. For that, I’d take a look at the graphs presented by the BBC (they are on a log scale, so the difference between Norway and Sweden is substantial). My point is simply that their analysis is totally bogus. Once again, the data aren’t on their side.

Shutting things down works

Another core piece of their argument is that shutting things down and sheltering in place aren’t effective. This is based on, as far as I can tell, utterly nothing. They claim there are studies to support their claim, but they don’t cite them and I can’t find them. They also try to use the Norway/Sweden comparison, but as I showed above, if anything that actually suggests that shutting down works. Finally, when pressed by a reporter for their evidence, one of them seemed frustrated and said, “I don’t need a double-blind clinically controlled trial to tell me if sheltering in place is appropriate. That is a college level understanding of microbiology.”

Now that statement is interesting for a number of reasons. First, up until that point (it was late in the interview), they kept insisting over and over again that they were just following the science. They repeatedly claimed to be the ones objectively looking at facts. Yet when pressed for their evidence, they retorted by saying they didn’t need studies, because they just knew (a very common science-denier strategy). Further, the effectiveness of sheltering in place is clearly not something you’d test with a double-blind clinically controlled study, which makes me suspect that they know very little about experimental design.

In reality, despite their claims to the contrary, very basic math tells us that sheltering in place will work, and you can very clearly see the pattern across the world of countries shutting down, followed by flattening the curve and, if they stay shut long enough, levels dropping. Indeed, if you just think about this for a second, it should make perfect sense. You can carry COVID for roughly 2 weeks without symptoms. So, if you are out and about, you are spreading that everywhere. If you are at home, you aren’t spreading it. Further, if you are at home, then you aren’t being exposed to others who might spread it to you. Even a very, very basic understanding of epidemiology is enough to realize that the rate of viral spread in a population is strongly influenced by the number of interactions people have with other people. The more interactions, the greater the spread; the fewer interactions, the less the spread. I can’t believe I even have to explain that. This is why density is such a critical component of disease outbreaks. These guys can present themselves as experts all they want, but they clearly don’t know what they are talking about, which is why actual epidemiologists and health officials say they’re wrong.

Death rates are low because we took action. Also, the outbreak is ongoing

At several points, they criticized the early models that predicted hundreds of thousands or even millions of deaths. They cited them as evidence that people over-reacted and the disease is not much deadlier than the flu. This argument is, however, based on a poor understanding of the models. These sorts of models don’t show what will happen, rather they showed what could happen under a range of scenarios. We run them precisely so that we can change our behaviour and avoid the worst outcomes. That’s literally their purpose. We don’t run them for the fun of it. We run them so that we can learn how to save lives. The number of deaths is much lower than originally predicted because we shut down schools, implemented shelter in place orders, etc. The things that these doctors want to undo are the very things that prevented us from having millions of deaths. This is very much like anti-vaccers arguing that we don’t need the measles vaccine because measles deaths are rare. They are only rare because most of the population is vaccinated. Even so, the death rates are “low” because we implement measures to make them low.

The other thing to keep in mind is that the situation is very much ongoing. At the time I’m writing this, the US has nearly 57,000 deaths from COVID-19, and it is still adding well over 1,000 (often over 2,000) deaths daily. Erickson frequently cited annual US deaths from the flu as being 24,000-62,000 (at another point they said 37,000–67,000; it’s actually 12,000­–61,000). He used this as evidence that COVID is no worse than the flu, but stop and think about that for a second, we are already at nearly the highest end of that range, and given the rate at which people are dying, we are going to shoot well past it, and that is with things shut down. Just think how much worse this would be if things weren’t shut down! The only reason they are even able to make that comparison is because we shut things down. They are simply wrong that COVID isn’t deadlier than the flu.

Comorbidity doesn’t mean COVID isn’t responsible

They also spent a great deal of time arguing that COVID isn’t really the killer, it’s actually the other conditions like being immunocompromised, being a smoker, etc. This is a very stupid argument. Yes, most mortalities are associated with other factors, but that does not in any way, shape, or form change that fact that those people would not have died if they had not caught COVID. Indeed, that’s a big part of why keeping things shut down is so important: it protects the people who are the most vulnerable.

Interestingly, this argument is another one that they lifted straight out of the anti-vaccine playbook. Anti-vaccers frequently make the same argument claiming, for example, that measles doesn’t kill anyone; it’s the secondary infections that kill. This, of course, ignores the fact that those infections happen because of the measles. Even so, yes, COVID generally has help in killing patients, but that doesn’t negate its role. This argument is like talking about a gunshot victim and arguing that, “Bullets aren’t dangerous, because the bullet didn’t kill him; it was really the loss of blood.” It’s a very dumb argument.

 Sheltering in place won’t destroy your immune system

A final core thrust of their argument is the notion that sheltering in place will harm your immune system and make you sick. You see, according to them, viruses and bacteria are the “building blocks” of your immune system, and if you aren’t regularly exposed to them, they will all disappear somehow, they won’t protect you, and your immune system will be weakened. They also extended this to arguing that you shouldn’t disinfect things in your house, shouldn’t wear a mask in public, etc. They justified this by smugly saying, “we’re not wearing masks. Why is that? Because we understand microbiology, we understand immunology, and we want strong immune systems.”

To quote a famous meme, “That’s not how this works, that’s not how any of this works!”

There are a ton of issues here, but let me start by acknowledging the grain of truth in their silo of stupidity. It is true that you have a microbiome consisting of many bacteria, viruses, etc. and they do play important roles in your body, potentially including helping fight some diseases. Also, there is some evidence that exposure to microbes early in life helps to train the immune system to respond correctly, and a lack of those microbes results in autoimmune disorders and allergies (this is known as the “hygiene  hypothesis”).

None of that, however, supports their claims. First, there is no reason to think that staying home for a month or two is going to dramatically alter your microbiome. They acted as if all the bacteria living in and on you will die if you don’t go outside. That’s nuts. They will keep living and reproducing and doing bacterial things. You are their home. That’s where they live. To be clear, a change in your routine might shift the microbiome around slightly, but it is constantly shifting around slightly, and there is no scientific evidence that sheltering in place is going to shift your core microbiome in a detrimental direction, and it certainly isn’t going to deplete your body of bacteria.

Further, they acted as if your house is totally sterile (except when it was convenient for them to act otherwise; see later), which is insane. Even if you disinfect your counter (as they waxed on about), your house is crawling with bacteria. Do you live with other people? They have bacteria. Do you have a pet? They are coated in bacteria. Even inside, you are constantly exposed to bacteria other than the ones on and in you.

Additionally, microbes are not the “building blocks of the immune system.” They aren’t even part of the immune system. Sure, they train your immune system, but only in that it learns which microbes to attack (and how to attack them) and which ones not to attack. Many people (these guys included) talk about exposure to bacteria “strengthening your immune system” as if exposure to bacteria A will result in a general improvement in your immune system and ability to fight other bacteria, but that simply isn’t how it works. As I explained in detail here, exposure to bacteria A simply teaches your immune system how to kill bacteria A and whether it needs to. It doesn’t “strengthen” it against other bacteria/viruses.

Note 5: Unlike these two, I have actually published papers on host microbiomes.

They lied when they said experts agreed with them

At one point, they said that they had shown their results to local health officials, and those officials agreed with them and were just waiting on permission from the governor to re-open things. That was, however, a lie. A spokesperson for the Kern health department said “our director has not concurred with the statements that were made yesterday about the need to re-open at this time.” As a general rule, I don’t trust people who make such brazen lies.

They aren’t experts/don’t cherry-pick your doctors

This is a point that I have touched on repeatedly throughout, but it is worth stating again: these two are emergency care doctors, not microbiologists, not immunologists, not virologists, not epidemiologists. They are not experts on a topic like COVID. They are not people you should be treating as authorities. Having an MD does not make you an expert on all aspects related to medicine. They state early on that they have taken courses on these topics, which I’m sure they did back in pre-med/med school, but that doesn’t make them experts. I took courses on these topics, as well, as part of my training in biology, and, as part of my PhD, I even studied microbiomes and the effects of an emerging infectious disease on the ecology of wildlife populations, but that doesn’t make me an epidemiologist. The fact that I have a little bit of training and experience in that field does not make me an expert in it, and it certainly doesn’t put me on par with people like Fauci who have spent their entire lives studying these topics.

When pressed on why actual infectious disease experts fundamentally disagree with them, they first tried to dodge the question by going on rabbit trails about how Fauci’s actions weeks ago were justified because he didn’t have all the data (which completely ignores the actual question of why people like Fauci still disagree with them now). Then, they eventually argued that the disagreement was because people like Fauci have just been doing research from afar for years, whereas they are “in the weeds” seeing how things are on the ground. This is, of course, an insane argument. Beyond the fact that many (probably most) other healthcare workers who are “in the weeds” with them disagree (we’ve all seen the photos of nurses blocking protesters who are trying to open things up), treating COVID patients does not in any way shape or form make someone an expert on the factors and conditions that allow the virus to spread. Emergency patient care and epidemiology are two very different things and doing one does not qualify someone as an expert in the other.

Nevertheless, I’m sure there will be some who continue to insist that these two know what they are talking about because they are doctors, at which point my question becomes, “why trust them?” There are thousands of doctors with far more relevant experience who disagree, so why trust these two? Why cherry-pick them out of all the experts? What makes them more trustworthy than all the other MDs and PhDs? Is it possible that you are blindly believing them not because they have good data (they don’t) or because they are experts (they aren’t) but rather simply because they are saying the things you want to hear? You should carefully consider this possibility, because it is a very easy cognitive trap to fall into.

Massive conflicts of interest and probably biases

It’s always a good idea to see if people have something to gain from making public claims like this, and in this case, the conflicts of interest couldn’t be clearer. Erickson started off by saying that many hospitals have been furloughing staff, shutting down wings, etc., and he returned to that point frequently. Then, at the end, he pointed to the various news people in the room and asserted that if COVID had cost them their jobs and they weren’t getting a paycheck, they might have a different view of the situation.

I found that very interesting, because Erickson and Massihi aren’t simply doctors. Rather, they own a series of urgent care facilities (Accelerated Urgent Care), which, as they admitted, aren’t getting many customers right now. In other words, the shutdown is hurting them financially, and re-opening would be hugely beneficial to them, but I’m sure that had no influence, right? Never mind the fact that they literally said finances would influence people; I’m sure they are just after the public good (sarcasm).

There also seem to be some very strong political biases at play. Erickson hinted at a conspiracy throughout, with frequent statements about “something else going on.” For example, he asserted (with no evidence whatsoever) that ER doctors were being pressured to write “COVID” on death certificates for some political motive. “Why are we being pressured to add COVID? To maybe increase the numbers and make it look a little bit worse than it is? I think so.” This is a nonsense conspiracy that I have not been able to find any evidence to support. Further, he admitted that it would be administrators who would be pressuring doctors, but when asked why administrators (who are hurting financially from COVID) would try to artificially increase the death rate, he didn’t have an answer.

Things really went off the rails after the main chunk of the interview (and a related moment part way through) where Erickson started ranting about the word “safe.” In regards to official health recommendations, he said, “when they use the word safe, the word safe, if you listen to the word safe, that’s about controlling you.” Ah, the ubiquitous and amorphous “they.” Who is “they”? Who knows? Probably some evil government entity. Later he said (his emphasis), “who says what’s safe? Are you smart enough to know what’s safe for you or is the government gonna tell you what’s safe for you…They are using this to see how much of your freedom can they take from you.” This is, of course, pure pablum. I wonder if he also applies this line of reasoning when the government tells people how to be “safe” when a hurricane is coming. I also wonder if he has more trust for health officials when they determine the “safe” doses of drugs. It’s just insane to argue that health guidelines designed for public safety are really about public control, and, more importantly, it makes his biases abundantly clear. It is extremely obvious that he is motivated, at least partially, by a strong distrust of the government. Despite all of the claims to be just looking at the evidence, they gave a lot of indications of following political biases, not facts.

Note 6: In cases like this, the general public does not know enough to know what is safe. To be clear, it’s about knowledge, not intelligence, but most people simply don’t have the training and experience to actually evaluate epidemiological data and determine what is safe. That’s why we should listen to experts like Fauci, not some random person on the street. This assault on expertise is yet another very common anti-science strategy.


At this point, I have covered the bulk of this nonsense and hit the most important points, but I want to touch on a few more minor issues.

Quarantining the healthy

At several points they asserted that quarantines have never applied to healthy people before. This is a lie. During the 1918 Spanish flu outbreak, things were shut down very much as they are today.

Coronavirus on plastic objects

At one point, they argued that because coronavirus can live on plastic for up to three days, sheltering in place is pointless because you are just bringing it into your house when you buy things. They even asserted that they can probably find COVID in your house. This is just stupid. The odds that a shovel from Home Depot (to use their example) has COVID on it are very low, and much, much lower than the odds that one of the many people you’d be near in Home Depot would give you COVID (seriously though, you probably don’t need to go to Home Depot right now). The fewer people you are in contact with, the lower your odds. It’s that simple. So yes, sheltering in place absolutely does minimize your risk.

Also, note that by claiming that things you bring into your home commonly harbor micro-organisms, they have just totally negated their own argument about staying at home preventing you from being exposed to bacteria and viruses. You can’t have it both ways.

The need for more testing

At one point, they accidentally gave up the game and said, “In order to re-open the economy, you have to have widespread testing, that’s #1.” That is actually something I agree with. If we had sufficiently wide-scale testing, we could start slowly re-opening because we could monitor cases and quarantine the sick. The problem is that we aren’t there. Not even close. We don’t have nearly enough testing to be able to do that. Thus, this admission defeats their entire argument.

When a reporter asked when they thought testing would be sufficient, they dodged the question and started arguing that people aren’t getting tested because they are just so scared to leave their home (never mind all the testing shortages).

More contagious than the flu

They also shot themselves in the foot by admitting that COVID19 is more contagious than the flu. Think about this with me. Their core argument is that COVID’s death rate (i.e., deaths among infected individuals) is the same as the flu’s. Even if that were true (it’s not), COVID would still be more dangerous because of the higher transmission rate. Being more contagious would mean more people get it, and, as a result, more people die.

Claimed New York ordered 30,000 ventilators and used 5

In another odd segment, Erickson went on a tangent about how ventilators aren’t saving anybody (which is untrue) and claimed that New York only used 5 of the 30,000 ventilators it ordered. I have not been able to find any evidence that NY over-ordered ventilators, and I certainly haven’t found evidence that only 5 were used.

Misuse of the word “theory”

At another point one of them said, “you’d better have a very good scientific reason and not just theory.” This is a minor issue, but I really hate the misuse of the word “theory.” In science, a theory is not an educated guess. It is an explanatory framework which has been rigorously tested and shown to have a high predictive power. Theories don’t graduate to become facts; rather, they explain facts. Saying that something is a theory does not indicate that we are uncertain about it. The very notion that viruses cause disease is a theory (i.e., the germ theory of disease).


In short, these doctors have no clue what they are talking about and seem to be motivated by money and politics, not science. Their statistics are bogus and their facts are faulty. ABC should be ashamed of itself for broadcasting their nonsense.

I think information should be shared and available. So you may re-blog, re-post, quote, etc. as you see fit, just please include a link/credit back to my page. Thanks

This entry was posted in Uncategorized and tagged , , . Bookmark the permalink.

1 Response to Shoddy statistics and false claims: Dr. Erickson dangerously misled the public on coronavirus

  1. Fallacy Man says:

    Due to the inevitable flood of political comments that will otherwise ensue, I have decided not to allow comments for this post.

    Liked by 6 people

Comments are closed.