What’s Up with the COVID Death Counts?

By /

The last time I blogged on all-cause mortality was back in December. Back then I said that I would wait until mid-January to get a sense of how the end-of-year data shook out. Well, we’re there, so this weekend I downloaded the CDC data again into Excel. What did I find? Something that suggests that we could have a serious misinformation problem.

What Is All-Cause Mortality?

As I have written before, there is evidence that early in the pandemic COVID-19 as a cause of death was being under-diagnosed. That makes sense as testing was scarce. On the flip side, late in the pandemic, it’s also clear that we have an over-attribution problem. Meaning that the actual number of people who died because of COVID-19 is being inflated.

Why? Look no further than a 20% Medicare hospital payment kicker for a COVID diagnosis. Or the peer-reviewed data showing that our PCR tests are being run way too many amplification cycles leading to finding dead-virus in the noses of our patients (aka false positives).

Regardless of what you personally believe, the great equalizer is all-cause mortality. This looks at all deaths, regardless of cause. This number is also hard to screw up as someone is either dead or they’re not. We also know approximately how many people are supposed to die each week. For example, about 50,000 people died each week in the US way before COVID got here. That weekly number goes up and peaks in January every year goes down in the summer every year like clockwork. So while the totals fluctuate year by year, if something is causing a lot more deaths, we would expect to see all-cause weekly deaths spike.

Request a Regenexx Appointment

My First Look

The CDC has many webpages with curated data as well as raw data you can download directly into Excel (1).

I first hit this graph at the CDC website of 2020 COVID deaths by month (1). I was at first a little taken aback. I see our first wave of COVID deaths peaking in April off to the left. However, where’s the massive peak in December that was all over the media over the holidays? The mid-December peak here is lower than the April peak (I drew the red dashed line from the April peak to the 2nd wave).

For this particular graph, the CDC is telling us here they have complete data through mid-December (solid line), but the dashed line after that represents incomplete data. Hence, complete death data lags about a month.

I next downloaded the CDC all-cause mortality data from this same page directly into Excel and this is what I got once it was graphed:

Huh? Now when I draw the line from the 1st wave to the second wave, it’s much smaller than the 1st wave. The first graph had 17K deaths at the worst week in April and 16K COVID deaths peaking in mid-December. So a net difference of 1,000 deaths at the peak. The all-cause graph is off by 10,000 deaths from 1st to 2nd waves (April to December). Why? These two should more closely match up if only COVID were causing most of our second wave.

Why would CDC reported COVID deaths at the peak weeks in April and December be off by 10K deaths on the all-cause graph? The single most likely reason is the over attribution of deaths to COVID. Meaning physicians in the field are tagging more people with a COVID diagnosis. That makes the number of 2nd wave COVID deaths on the first graph bigger. That isn’t to say that COVID isn’t spiking in a second wave right now, but the CDCs data tells a different story than we hear in the media.

There’s also another serious issue here. Looking at the peaks going back a decade, they always happen in January. Hence, a very strong argument could be made that all-cause deaths almost never peak in April (as they did in the first wave), hence the first wave deaths were actually due to COVID. However, the second wave deaths are peaking in January, right where we would expect to see a peak in deaths if COVID didn’t exist. Which makes it very tough to tell expected normal winter deaths from COVID deaths.

What’s Causing this Difference?

To try to answer that question let’s go back to the CDC website. For example, look at these death stats:

This table lists 329,935 deaths “involving” COVID. That simply means that a COVID-19 test was positive and they died. Given that just about every US citizen being admitted into the hospital or getting surgery right now is getting a COVID test and we have huge numbers of asymptomatic carriers, that number has loads of people who tested positive and died of other things. Specifically, that doesn’t mean they died of COVID-19. Then we have 319,102, which is people who died of pneumonia with or without a positive COVID19 test. Meaning some in this group didn’t have COVID-19. So how many died of pneumonia and had a positive COVID test? 152,620. That number is much less than half of the 395,851 listed on the Johns Hopkins site (2).

What Else Could Cause This Blunting of the Second Wave Peak?

You could argue that because a good portion of the country is locked down in one way or another, that’s reducing the normal spread of other infectious diseases. Hence, the normal expected deaths from influenza and pneumonia are down which is dragging down the impact of that second wave all-cause mortality peak. However, if we look at the CDC data for our 2018 peak for usual influenza and pneumonia deaths, even if those rates went to zero, that only removes 1,400 deaths. So we can’t make up a 10,000 difference that way.

In addition, given that COVID spreads at a similar or slightly higher rate than influenza, if one was down or up, you would expect it to go in the same direction as COVID, not the opposite (here COVID deaths are up as reported by CDC). Also, sitting at home and not seeing the doctor has been estimated to increase all-cause mortality for things like suicide, heart disease, and cancer due to delayed diagnosis and treatment. An October article in Health Affairs had this to say (5):

“Patients with or at-risk of cancer may also be disproportionately impacted by continued disruptions to care. The volume of preventive cancer screenings dropped sharply during the onset of the pandemic, and despite a recent rebound, remain well below pre-pandemic levels. Consistent with this rising cumulative number of missed screenings, the number of newly diagnosed cancers has also dropped significantly. These trends raise concern that many cancers may only be detected at a later, more advanced stage, which could limit treatment options or result in poorer prognosis. One study estimates that, over the next ten years, the pandemic’s negative impact on screening and treatment could lead to nearly 10,000 excess breast and colorectal cancer deaths.”

Join us for a free Regenexx webinar

Is This Our Real Second Wave Tragedy?

I’ve had a great deal of respect for the University of Washington’s IHME coronavirus model (3). Having said that, all COVID-19 models are tarot card reading with some complex math thrown in for good measure. A case in point is our Colorado hospital resource use. I’ve been watching that for two weeks and the IHME model predicted that we would slightly exceed our 555 ICU beds around January 12th. Today, once the model has been fed in some real-world data, Jan 12th shows that we only needed 262 ICU beds! Our peak will now be in late January with more than 200 ICU beds to spare. Again, Tarot card reading with math!

However, one drop-down choice over from Colorado on the IHME model is California. Here the IHME model is predicting that our most populous state will peak in hospital use around February 12th and need 9,241 ICU beds and only have 1,994. Let’s dig in here for a minute.

The IHME model states that California has 26,654 regular hospital beds. However, the state government website lists 73,906 total beds (4). That same document states that California has 7,451 ICU beds or about five thousand more than IHME says they have. Hence, as I have relayed before, much of our media frenzy over bed shortages is being caused by hospitals continuing to try to function during the second wave. Meaning in the first wave, most shut down elective procedures, opening 90% of the beds for COVID-19 patients. That’s no longer happening. Now, hospitals are reporting which beds are open to COVID patients, which is a number far smaller than their total beds.

So will California be a huge disaster? That could be, but if they temporarily shut down elective hospital procedures again, and enforce vaccination for all employees, the number of ICU beds can likely be expanded to closer to the predicted need.

Where Will This Second Wave Head?

We’re in the prime time of the usual respiratory bacterial and virus deaths right now. Every peak since 2014 has been around mid-January. The IHME US model has pushed our United States COVID death peak into late January. Hence, I won’t be able to get another good look at the all-cause mortality until late February. What will the data tell us? My prediction is that we’ll continue to have a COVID attribution issue. However, let’s let the data speak for itself.

The upshot? The CDC data tells a different story than the COVIgeddon media messaging. Having said that, nothing that I have written here is a license to stop hand washing, social distancing, and mask use, or to forego needed vaccines. There are definitely real COVID deaths out there and I would rather you not be one of the numbers in these stats come med-Feburay.



(1) Centers for Disease Control and Prevention. National Center for Health Statistics. Weekly Updates by Select Demographic and Geographic Characteristics. https://www.cdc.gov/nchs/nvss/vsrr/covid_weekly/index.htm Accessed 1/16/21

(2) Johns Hopkins Unversity of Medicine. Coronavirus Resource Center. https://coronavirus.jhu.edu/ Accessed 1/17/21

(3) University of Washington. Institute for Health Metrics Evaluation. https://covid19.healthdata.org/united-states-of-america?view=daily-deaths&tab=trend Accessed 1/17/21

(4) CHHS Open Data. Licensed and Certified Healthcare Facility Bed Types and Counts. https://data.chhs.ca.gov/dataset/healthcare-facility-bed-types-and-counts Accessed 1/17/21

(5) Health Affairs. Spillover Effects Of The COVID-19 Pandemic Could Drive Long-Term Health Consequences For Non-COVID-19 Patients. https://www.healthaffairs.org/do/10.1377/hblog20201020.566558/full/ Accessed 1/17/21

This blog post provides general information to help the reader better understand regenerative medicine, musculoskeletal health, and related subjects. All content provided in this blog, website, or any linked materials, including text, graphics, images, patient profiles, outcomes, and information, are not intended and should not be considered or used as a substitute for medical advice, diagnosis, or treatment. Please always consult with a professional and certified healthcare provider to discuss if a treatment is right for you.

Get Blog Updates by Email

By submitting the form, you are agreeing that you read and consent to our Privacy Policy. We may also contact you via email, phone, and other electronic means to communicate information about our products and services. We do not sell, or share your information to third party vendors.

Category: Uncategorized

Leave a Reply

Your email address will not be published. Required fields are marked *

6 thoughts on “What’s Up with the COVID Death Counts?

  1. Daniel Green

    Dr. C,
    Calling COVID-19 models tarot card reading with some complex math thrown in for good measure is a bit insulting to me as a biostatistician. Infection modeling is a very complex process with a myriad variables and assumptions. It takes years of post graduate schooling to do this well. You are comparing a very sophisticated stochastic modeling to scienceless astrology and that’s not really fair. It would be more accurate to just say that even with complex modeling of very intelligent people, it is not easy to predict many of the outcomes associated with the pandemic. Here is a link providing information on many of the highly regarded
    modeling projects out there. https://www.aha.org/guidesreports/2020-04-09-compendium-models-predict-spread-covid-19
    Regards, Dr. D Green

    1. Chris Centeno, MD Post author

      Dan, it’s all about the outcomes. The SAGE model has been absurdly inaccurate. For example, predicting millions of US deaths in the first wave. IHME has been better, but still wildly inaccurate. So you can call it tarot card reading with math or “the complexities of predicting disease spread with biostatistics often means less than satisfactory accuracy for existing disease models”. So until we get to Hari Seldon’s level of sophistication (as Asimov described in the Foundation series), if anything teh pandemic has shown just how severely limited epidemiologic models can be.

  2. David Wieland

    It’s not demeaning Dr. Green’s abilities as a statistician to note that models of all kinds are approximations and/or proxies for reality (e.g., architectural models). Biological science has advanced enormously in the past century, but discoveries seem endless. Models are valuable research tools but are not capable of reliable prediction, especially for complexity with essentially unknowable elements (e.g., improvements in elder care or worsened shortage of PSWs). In addition, models make assumptions about the effect of parameters, e.g., changes in social distancing (from the page Dr. Green linked). Assumptions certainly affect the models, but their validity can only be judged with analysis of empirical data, and even then it may be impossible to disentangle confounding factors.

    I saw one model that reportedly showed the benefits of increased mask wearing. It assigned a specific effect to the percentage use of masks. The model output would have been the same if the parameter represented the use of headphones. As you said, we have to let the data speak for itself.

  3. Joe LaValle

    It seems one potential disconnect is the assumption that CDC all-cause mortality data is current and up to date through the end of December. Your assumptions presume it is, but the CDC states that all-cause mortality data reporting may be delayed by up to 8 weeks. If you download the state by state all-cause mortality tables that the CDC chart uses to construct the baseline chart you refer to and focus on state-by-state December 2020 data (which takes a bit of scrolling), an interesting data point jumps out — many states show a significant decline in reported all-cause mortality beginning the weeks ending 12/5 or 12/12 – with a major drop for the week ending 12/26. Some states with known high COVID-19 reported deaths in December show zero excess deaths to the CDC in these periods, which is a statistical anomaly worth pursuing. A logical conclusion as to why excess all-cause mortality in December doesn’t synch with reported COVID-19 deaths may simply be that states have reported incomplete December results. Indeed, North Carolina didn’t issue any reporting in December. And reporting seems especially in arrears for the holiday weeks ending 12/26 and 1/2. Do you buy into the argument that your numbers don’t synch because CDC December data is incomplete?

    1. Chris Centeno, MD Post author

      Joe, the data is lagging about a month. Hence, right now date through mid-December is pretty accurate. I’ve been following this data every week, so I can see it develop and change.

  4. Jason

    Thanks for the update, Dr C. Very informative.

    As you have shown, the seasonal behavior of COVID is the same as the flu… and latitude and solar radiation dictates the degree of infection more than any other factor, as per Dr Edgar Hope-Simpson. Based upon this, the COVID deaths will drop markedly in February everywhere north of 30 N latitude in the USA and Europe, and will be miniscule by end of May. Daily death rates will likely drop more slowly in southern California, but also beginning in Feb.

    The detractors commenters mention nothing about the irrefutable seasonal behavior of this virus, like the flu, that is likely not in any of the models.

Is Regenexx Right For You?

Request a free Regenexx Info Packet


Learn about the #1 Stem Cell & Platelet Procedures for treating arthritis, common joint injuries & spine pain.

Join a Webinar


Get fresh updates and insights from Regenexx delivered straight to your inbox.

Subscribe to the Blog


9035 Wadsworth Pkwy #1000
Westminster, CO 80021


Copyright © Regenexx 2021. All rights reserved. | Privacy Policy

*DISCLAIMER: Like all medical procedures, Regenexx® Procedures have a success and failure rate. Patient reviews and testimonials on this site should not be interpreted as a statement on the effectiveness of our treatments for anyone else.

Providers listed on the Regenexx website are for informational purposes only and are not a recommendation from Regenexx for a specific provider or a guarantee of the outcome of any treatment you receive.