TEN MILLION INFECTED, 140K DEAD, Getting the real numbers

There’s a lot of conflicting information concerning COVID-19, but it’s becoming increasingly clear that the information being released is significantly understating the scale of the epidemic. Many Governors are basing policy decisions on unrealistically rosy information. These decisions aren’t based in reality and are likely to come back and slap them in the face. Fortunately, there’s enough information out there now to put together a more realistic picture of what’s really going on. That’s the objective of this article. First, I’ll try to determine what is the true number of infections and fatalities? I’ll then attempt to establish if we’re seeing a real reduction in cases. The true state of the pandemic needs to be understood before making decisions about reopening of the economy.

Everything about Coronavirus statistics are controversial and nothing more so than the number of deaths from the virus. Social media is awash in right-wing memes claiming to have evidence of car crash fatalities being classified as COVID-19 victims. Missing in these memes is the actual evidence needed to verify the claims. There are also anecdotal accounts of people dying with clear Coronavirus symptoms not being recorded as a COVID-19 death because they weren’t tested. Fortunately, we have a way to tease out the truth.

The CDC is publishing the total number of deaths across the nation from all causes. Looking at the data these numbers tend to be pretty steady. There may be local spikes, but across the country the number of deaths on a given day is consistent from year to year. With this new virus it’s a straightforward process of comparing this year's deaths to the average and seeing if we have a bunch of extra bodies. Lo and behold, in mid-March as Coronavirus cases began increasing there was a significant increase in deaths. The nice thing about this data point is it’s a reliable statistic. When someone dies it triggers a death certificate which gets recorded and becomes a statistic. The cause of death can be compromised, but the existence of a body in the morgue is an unequivocal fact.

This was reported about a month ago, however, I wanted to look at the data myself. Plus, an additional month of data should result in more definitive results. The main limiting factor in this information is it can take one to five weeks for death certificates to be included in the data. On average after 10 days only about 60% of deaths have been recorded. As a result, you have to go back at least a month to get reliable numbers. The first week the total deaths began rising above normal was the week ending March 21st. At that point there were only about 500 extra deaths. However, the numbers continued to grow each week and by April 18th 58,164 more deaths had been recorded than usual. Officially at that time there were 39,332 recorded COVID-19 deaths meaning we had 18,832 unexplained additional deaths. Projecting that forward to the current date it seems likely the number is near 40,000 by now.

If this is accurate it means the official Coronavirus death toll of about 100,000 is more likely around 140,000. That’s a significant under count, but is it accurate? To put this method to the test I looked at a previous spike in deaths between mid-December 2017 and mid-February 2018. During this two-month period there were an extra 21,761 deaths. As it turns out, that was one of the worst flu seasons we’ve had in many years with approximately 20,000 more flu fatalities than normal. This is a good indication the total deaths are a good indicator of the real impact of an unusual epidemic.

The most compelling argument I found questioning this statistic is that people who had unrelated ailments or injuries may have died as a result of emergency rooms and intensive care units being filled with COVID-19 patients. On the one hand even if that is the case, they would still represent at least an indirect casualty of the pandemic. But even if that’s a stretch, it seems unlikely we’re talking about very many people who would fall into that category. There were few if any places that maxed out their hospital capacity. I doubt it’s possible to accurately quantify this, but it seems unlikely it significantly impacted the numbers. It’ll take months to get the full data from the first wave of the Coronavirus pandemic. However, indications are the death toll is close to 40% higher than the numbers being reported.

The total number of Coronavirus cases is the other statistic that is batted around a lot. This one is a lot more problematic. Unlike with the deaths, there’s no solid baseline to look at to determine the number of cases. From the beginning we’ve lacked the proper level of testing capacity, although it has improved. As a result, most accept that the official number of cases is understated, but by how much? New York state recently did a study concerning this by giving a random sampling of people the COVID-19 antibody test. They concluded that the infection rate was 12.3%, or more than 6 times the official statistic. New York certainly has a higher infection rate than the rest of the country. Yet if the rate of under reporting holds accurate, we can extrapolate that America has a little over ten million cases. That’s about 3.2% percent of the population. I had to make some assumptions that can’t be verified to get that number, but this seems more plausible than the official statistics. Also of interest, this team took their finding and combined it with the information from the excess deaths to calculate a fatality rate of 1.4%. This is in line with earlier estimates but higher than many conservatives have been willing to admit to.

In all this depressing news there is at least one encouraging new development. There have been numerous reports of people testing positive after recovering from the Coronavirus. The implications of this were pretty grim. If getting sick doesn’t provide future immunity, developing a vaccine may be impossible. The good news is several studies concluded the positive test results were the result of the original virus, not a new infection. None of the people who tested positive actually got sick a second time. Also, scientists were unable to grow the virus they shed in a laboratory. This means they were shedding dead viruses and were not contiguous. Researcher now believe based on how other Coronaviruses work people likely have immunity for a year or more after recovering from the virus. There’s also reason to believe there may be cross immunity with other Coronaviruses. This could explain why some people get much more severe symptoms than others, but this is speculative at this point.

Finally, and perhaps most importantly, is the first wave of this pandemic finally receding? The Trump administration is confidently proclaiming that it is, but once again the truth is murkier than they’re letting on. The first consideration is which matrix you’re going to base your answer on. The number of cases is so under reported that it’s a nearly useless statistic. However, that seems to be the most commonly used standard, so we’ll look at that first. The below chart does show a gradual decline in new cases since early April, but it’s only a slight decline. When you look at the most recent data on that chart, you’ll see that the decline has leveled off. Of course, there’s also more testing now which makes it hard to compare numbers. The most ridiculous argument I’ve heard some on the right make is, “we know the infection rate has gone down because the percentage of positive test results has gone down.” That logic is bananas. A month ago people with clear symptoms were the main group getting tested. The range of people getting tested has widened so obviously the percentage of positives is going to go down. That’s not an indication of anything.

The more reliable indication of the infection rate is the number of deaths. There are problems with this matrix as well. As previously discussed, it’s also under reported, although not by nearly as much. There’s also a greater lag time between being infected and dying than from being infected and testing positive. Finally, as doctors get better at treating the virus presumably more infected people will survive. So, a decrease in deaths doesn’t necessarily mean a decrease in infections, still, this is the best measure we have. The thing that immediately jumps out from the chart below is the rate in New York is orders of magnitude higher than the others. The other eye-catching item is the remarkably steady rate in the other states. Nationally there is a slight decline but that isn’t being reflected in most states. The total rate of COVID-19 deaths nationally has fallen from a high of 2200 per day in mid-April to 1260 now, a drop of 940 per day. That’s a significant decrease. However, in New York, New Jersey and Connecticut the combined number of daily deaths has dropped from 1240 to 290, a decline of 950. That’s right, the drop in those three states is more than the drop in the entire country. The vast majority of that drop comes from the New York City metropolitan area. In other words, when you exclude New York City the rest of the country is still at their peak deaths per day.

There’s still a lot of researchers don’t know about this virus, but they are making significant progress in their understanding. Despite internet memes to the contrary, we can say with confidence that COVID-19 is about fifteen times more lethal than the common flu. In just two months it has killed three times the typical annual number of flu fatalities. Early in the pandemic there was hope that with warmer weather in the spring and summer we’d see a natural decrease in the incidents of this virus. Thus far, outside of New York City, this doesn’t seem to be the case with the rate of COVID-19 related deaths remaining constant in the rest of the country. The real-world evidence is pointing to the continuance of this pandemic for months to come.