Here’s another long post about COVID-19. Yes, I know it’s longer than a tweet but this is important so I hope you keep reading.
There are a lot of numbers being used to show you how fast the disease is spreading and how horrible it is. I’m not disagreeing that there are people getting sick and dying. I’m not even saying that that number isn’t growing. What I’m trying to do here is to show you that the information you’re being given isn’t necessarily the whole story. In some cases, things are better than reported and in other cases things are worse. Governments and the media are quick to throw around numbers without explaining that, taken out of context, they can be misleading.
Asymptomatic People: Asymptomatic simply means “without symptoms”. If you’re symptomatic, you likely have or had the virus and didn’t realize it. Also contained within this number are people who had mild symptoms but didn’t realize it was COVID-19-related. The estimated number of people who have the virus but don’t have any symptoms (or have mild symptoms) is estimated to be anywhere from 25% to 80%. That doesn’t mean that 25%-80% of all people have the virus. It means, of everyone who had the virus, 25%-80% of the people won’t even know it. The only way to determine this number is to measure them. But, since testing priority is being given to people who appear sick, there have been only a few attempts to make this measurement.
Daily Count of Positive Cases: I just saw a post from someone who shared a picture of a graph showing a “spike” in the number of new cases. Simply looking at the number of new cases doesn’t tell you much. Let’s say that each day 100 people are tested and 25% of them (25 people) typically test positive. If you suddenly test 1,000 people and 25% of them test positive(250 people). At first glance it looks like there’s a 10x increase in the number of new cases. But the increase is only because they TESTED 10 times as many people. In this example, the rate is, in fact, not increasing by a factor of 10 but is instead holding steady at 25%.
Another flaw with this measurement is that not all areas are testing the same way. Some are only testing people who have COVID-19-like symptoms. Other areas are allowing anyone, with or without symptoms, to be tested. If you are testing only people who are sick, you’re going to get more positive results than if you are testing a random group of people who may or may not have symptoms. If you know how they’re testing in your area you can potentially glean some useful information from that data. But, rolling up the data to a state/national level isn’t possible.
In our example above, let’s say that only sick people are being tested and the rate holds steady at 25%. Then one day they allow random testing and the rate drops to 10%, does that mean there’s been a drop in the infection rate? No, it means that the drop in cases was because we stopped testing only people who are sick. The results of the two tests can’t be compared because the type of people being tested has changed.
Total number of cases: When you see the number of cases reported, it really means the number of CONFIRMED cases. That is, this number only contains people who have tested positive. It does not contain people who are merely exhibiting symptoms, but haven’t been tested, or people who have no symptoms at all. The total number of CONFIRMED cases is technically an accurate number, but it’s misleading because the way it is typically presented implies that it is the TOTAL number of all cases (measured and unmeasured). Why is this important? Because the total number of CONFIRMED cases is being used to calculate other statistics which will be impacted by incorrect reporting.
The actual “total number of cases” is impossible to measure (I can explain if anyone is interested) but with the right data, we can make a fairly accurate estimate of what it is.
Total number of deaths:There is no global agreement on how to count any particular death as a COVID-related death. So, we can’t compare how the virus is impacting different parts of the globe. Some countries only count a COVID death as someone who died “from” the disease. That is, someone who died as a direct result of having the disease (for example, respiratory failure). Other countries, the U.S. included, count everyone who died “with” COVID-19 as having died “from” the disease. In other words, I could die from massive blood loss after an accident but, if I tested positive for the virus, it would be counted as a COVID-19 death. Another example is someone who is dying from cirrhosis of the liver because of a lifetime of drinking. While they might have died on Wednesday anyway because of liver failure, if they die on Tuesday because of a COVID-19 symptom, it still counts as dying “from” COVID-19. Reporting someone as having died “from” when they actually died “with” the disease is one of the reasons why it appears like the U.S. is being hit harder than some other countries.
In countries like the U.S. where someone who dies “with” the disease is counted as someone who dies “from” the disease, the total number of deaths is being reported higher than it actually is. This directly impacts any other calculations that use this number causing those results to be unreliable.
Percentage of Deaths: The percentage of deaths as reported has a lot of issues. What they’re really saying is “of everyone that we know had the disease, this percentage of them died”.
The percentage of deaths is calculated like this:(the total number of people who have died from the disease) divided by (the total number of cases) times 100.
As I mentioned earlier, the U.S. is incorrectly counting the number of people who died “from” the disease. So, using that number to calculate the percentage of deaths will give us an incorrect answer. If the U.S. was correctly counting people who died “from” the disease, we could consider this to be “Total VERIFIED percentage of death”. But, since they’re not, we can’t.
Further, the calculation typically uses the number of VERIFIED cases which isn’t the same thing as the total number of cases. But, as mentioned above, we don’t know the actual total number of cases. So, with the exception of the number 100 (which is used to turn this into a percentage) everything being used to calculate the percentage of death is inaccurate.
Why is this important? Let’s say that you live in a county that has 100,000 people. You’ve tested 10,000 people and found 1,000 people who tested positive. So, you’ve got an infection rate of 10%. Of those 1,000 people, 22 people die. That’s a death rate of 2.2%. But remember, having 1,000 people who test positive only means that those are the people you know of. It is unlikely that you were lucky enough to test only the people who had the virus. So, this means that there are other people that have the virus but weren’t tested. If you were testing random people, then it would be reasonable to believe that 10% of your total population is infected. So, the total number of people infected in your county is 10,000, not 1,000 and your death rate is actually .22%.
But, as mentioned earlier, most places are testing only people who are experiencing symptoms of COVID-19. If that’s the case, then the death percentage really means “of all the sick people that tested positive, this is the percentage of them that died.” Even so, the death rate is incorrectly being reported as worse than it actually is because they’re only considering a portion of the people who are infected.
