Here we go again: Putting COVID numbers in context

A disclaimer: COVID-19 is serious. If you don’t take it seriously by following the rather simple alleviation tactics you have available to you (vaccination, mask-wearing, washing hands, etc.), it can kill you.

There is a difference, however, between it can kill you and it will kill you, and this is a distinction that is seldom clear from the media’s coverage of the pandemic. As I’ve discussed before, numbers are offered out of context and percentages – especially increases in cases, hospitalizations, or deaths – are presented in a hyperbolic manner that make things seem much worse than they really are:

“Alaska and Arkansas more than doubled cases in the last week.” Okay, is that from 50 cases to 100 cases, or is it from 1000 to 2000 cases? Context matters.

“In Missouri, hospitalizations jumped by nearly 30% over the weekend.” Considering the covidestim.org map I included yesterday, this isn’t surprising, and a surge in hospitalizations is what we want to avoid so we don’t overwhelm our facilities and, more importantly, our healthcare workers. But is this from 100 to 130 people hospitalized, or from 1000 to 1300? Again, context matters.

The data aren’t wrong, but the context isn’t there, so they’re less useful than they could be at best, and counter-productive to an informed discussion at worst.

I went to an in-person conference last week in Bellaire, Michigan, which is near Traverse City. Definitely “northern Michigan.” Before I left, I saw a story online that said that COVID-19 cases in “northern Michigan” had risen 50% in the previous week. Which sounds pretty alarming. But there was no definition of what “northern Michigan” meant. (Believe me, there’s no agreement between the various parts of this Great Lakes State on what it means!) Was that “north of Lansing,” or “north of Mount Pleasant,” or “north of Grayling,” or even just the Upper Peninsula? Didn’t say. And, like the examples above, was the 50% rise from 100 to 150 cases (drilling down, that’s what it was) or possibly from 2 cases to 3 (also a 50% increase). So maybe it was concerning… or maybe it wasn’t.

There’s been a lot of concern about children and COVID-19, which is understandable. Nobody likes to see a child suffer or possibly die from an illness, and until very recently vaccines weren’t available for anyone under the age of 18. We continue to recommend that everyone wear a mask indoors, as a precautionary measure. The precautions are more obvious if someone isn’t vaccinated; the mask may help prevent transmission to that person (as well as help prevent them from spreading the virus if they already have it). Unfortunately, the unvaccinated are the least likely to follow those guidelines.

Mask-wearing may help fully-vaccinated people from contracting a “breakthrough” case of COVID, but more often the reasoning is to keep them from spreading the virus unknowingly to unvaccinated people, including children and teenagers.

We rarely distinguish between the likelihood of a particular individual getting COVID, becoming seriously ill, and dying, though. It’s an across-the-board recommendation that treats every American as having an equal chance of death in the face of COVID-19, which isn’t true.

A second disclaimer: This isn’t breaking news; I’m not presenting some discovery that I came up with on my own. (In fact, I’ve already looked at this back in May.) The data have been available for months, but the conclusions, while not terribly complex, don’t fit neatly in a headline or the lead paragraph of a news story.

COVID-related death rates by age group in the United States

Daily death rate, 2 October 2021, per 100,000 people in the United States, by age group.
Graph: Tom Kephart. Source: Our World in Data.

Fully-vaccinated individuals

On 2 October, the number of deaths from COVID-19 per 100,000 fully-vaccinated people in the U.S. aged 12-17 was 0.03, which was 0.3% of all COVID-related deaths of fully-vaccinated people on that date. A fully-vaccinated American teenager had a 1 in approximately 3.3 million chance of dying of COVID on that date.

Also on 2 October, the number of deaths from COVID-19 per 100,000 fully-vaccinated people in the U.S. aged 65-79 was 1.57, which was 18.2% of all COVID-related deaths of fully-vaccinated people on that date. A fully-vaccinated American senior citizen in that age range had a 1 in approximately 63,000 chance of dying of COVID on that date.

Finally, on 2 October, the number of deaths from COVID-19 per 100,000 fully-vaccinated people in the U.S. aged 80 and older was 6.51, which was 75.6% of all COVID-related deaths of fully-vaccinated people on that date. Our oldest fully-vaccinated citizens had a 1 in about 15,300 chance of dying of COVID on that date.

Put another way, those 65 and over fully-vaccinated Americans were over 15 times more likely to die of COVID on October 2 than people in the other four age cohorts combined, and 269 times more likely to die than a fully-vaccinated teenager.

Unvaccinated individuals

On 2 October, the number of deaths from COVID-19 per 100,000 unvaccinated people in the U.S. aged 12-17 was 0.05, which was 0.1% of all COVID-related deaths of unvaccinated people on that date. An unvaccinated American teenager had a 1 in approximately 2 million chance of dying of COVID on that date.

Also on 2 October, the number of deaths from COVID-19 per 100,000 unvaccinated people in the U.S. aged 65-79 was 31.87, which was 39.6% of all COVID-related deaths of unvaccinated people on that date. An unvaccinated American senior citizen in that age range had a 1 in approximately 3100 chance of dying of COVID on that date.

Finally, on 2 October, the number of deaths from COVID-19 per 100,000 unvaccinated people in the U.S. aged 80 and older was 38.28, which was 47.6% of all COVID-related deaths of unvaccinated people on that date. Our oldest unvaccinated citizens had a 1 in about 2600 chance of dying of COVID on that date.

Put another way, those 65 and over unvaccinated Americans were nearly seven times more likely to die of COVID on October 2 than people in the other four age cohorts combined, and over 1400 times more likely to die than an unvaccinated teenager.

Variations between vaccines

There continues to be a lot of discussion about which vaccine is “better.” Looking at the daily death rate data from 10 April to 2 October 2021, there’s not a lot of difference between the three main vaccines available in the U.S. The death rate for those who received the Johnson & Johnson single shot was 0.91 per 100K in April and remained steady at 0.88 in October; Pfizer’s April rate was 0.28 and rose to 0.60 in October (a 119% relative increase); and Moderna’s April rate was 0.15 and rose to 0.46 in October (a 202% relative increase). Those increases for Pfizer and Moderna seem notable as they could indicate the slow wearing-off of immunity, yet they remain very low at less than one death per 100,000 among those who received that vaccine.

Among all age groups, from 10 April to 2 October 2021, unvaccinated people died from COVID at a rate of 3.89 per 100K in April and 7.29 per 100K in October (an 87% relative increase).

So which vaccine a person received has little significance, but being vaccinated versus unvaccinated certainly does.

Risk of dying from COVID versus driving a car

While the data sets don’t align perfectly, a male drivers aged 15-24 had a death rate of 19.0 per 100,000 in 2019; female drivers of the same age cohort had a death rate of 8.2 per 100,000. Both rates are significantly higher than the death rates from COVID for the 12-17 and 18-29 cohorts, approaching the rate of unvaccinated 50-64 year olds.

Conclusion

Let’s repeat this so there’s no misunderstanding of what I’m trying to do here: COVID-19 is serious. If you don’t take it seriously by following the rather simple alleviation tactics you have available to you (vaccination, mask-wearing, washing hands, etc.), it can kill you.

But we’re doing ourselves a disservice by reducing the statistics to the most alarming, and often misleading, components. The SARS-CoV-2 virus is particularly deadly to older people, including those 65 to 79 but especially those 80 and older. Overall health of older and elderly people is certainly a factor, as are possible underlying health issues (many of which are exacerbated by age as well). Vaccinating children is as important to keep them from spreading COVID to their grandparents and great-grandparents as it is protecting them from dying. We’ve been concerned about the health of our children while the focus really needs to be on our oldest community members.

We’re getting bogged down in the various numbers and seem to have lost track of what they mean or whether the really mean anything. In addition to the physical, economic, and mental realities of pandemic fatigue, we also seem to be losing the ability to muster much of a response to the changing data on numbers of cases, how full ERs and ICU wards are, or how many people are dying from a virus that probably won’t be eradicated but certainly could have been much more controlled at this point. Everyone has their own data point they’re watching – and some of them don’t make any sense or have been chosen to fit a specific narrative.

The endless focus on COVID also distracts us from other concerns, some of which are deadly (including other illnesses such as influenza) and some of which aren’t (our widening political gulf between right and left and its implications for the future of our nation). There are new vaccines on the horizon and antiviral pills even closer to being available. An honest discussion of where we are – as opposed to overheated, sensationalistic media and water cooler talk – is critical if we have any hope of resolving the problems that we’re facing.

The rise and (hopefully) fall of the Delta variant in the U.S.

NOTE: This post uses several maps from covidestim.org. Here are the keys for each type of map:

Using the excellent data visualizations on the covidestim.org site, I noticed an interesting “wave” effect as the Delta variant started to take hold in the United States in late spring. The first reports of significant spread of the highly contagious variant were in Missouri and Arkansas in mid-June.

Infections per 100K people as of 20 Jun 2021 (Source: covidestim.org)

The signs were already there in the data, however. The effective reproduction number (Rt), which we’ve discussed before, is “the average number of people who will become infected by a person infected at time t. If it’s above 1.0, COVID-19 cases will increase in the near future. If it’s below 1.0, COVID-19 cases will decrease in the near future.” By May 23, both Missouri and Arkansas were seeing Rt numbers above 1.2 (orange and red on the map).

Effective reproduction number (Rt) as of 23 May 2021 ource: covidestim.org)

By July 4, ironically at the same time as President Biden was suggesting that the nation was declaring independence from COVID (which this New York Times article by Sheryl Gay Stolberg correctly questioned as being too soon), the Rt numbers throughout the southeastern U.S. were very high and most of the rest of the country was following.

Effective reproduction number (Rt) as of 4 July 2021 ource: covidestim.org)

Per capita infections were still low on Independence Day, but as predicted by the high Rt numbers, within three weeks that had changed for the worse.

Infections per 100K people as of 25 July 2021 (source: covidestim.org)

The spread from the southeastern states, where vaccination rates have been the lowest and several Republican governors have deliberately worked to thwart public health measures including mandating (or even promoting) vaccinations and wearing of protective masks, through the southwestern and midwestern states, the Pacific Northwest, and even Alaska, happened next. Here’s the map on August 29 with Delta at what appears to be its widest extent.

Infections per 100K people as of 29 August 2021 (source: covidestim.org)

Notice that at this point 18 days ago, Michigan was still experiencing lower rates of COVID infection (as was most of the rest of the upper midwest and New England). The state is still lagging in the percentage of residents who are fully vaccinated, unlike New England which has some of the highest vaccination rates.

At this same time (August 29), Rt numbers had begun to decline significantly in states that were still bright yellow and white on the infection per capita map.

Effective reproduction number (Rt) as of 29 August 2021 (source: covidestim.org)

Perhaps some of this was the increase in vaccinations that followed the Delta outbreak in the southeastern states, or perhaps it’s a result of there being fewer people left to infect. In any case, with Rt numbers declining back under 1.0 in those states, a predictable decrease in infections per capita should happen… which it has, according to the most recent map from September 14.

Infections per 100K people as of 14 September 2021 (source: covidestim.org)

Here are two animated GIFs of both sets of data covering the period from May 2 to September 14:

Effective reproduction number (Rt) from 2 May 2021 to 14 September 2021 (source: covidestim.org | animation: Tom Kephart)
Infections per 100K people from 2 May 2021 to 14 September 2021 (source: covidestim.org | animation: Tom Kephart)

There’s always the possibility that another variant could emerge and the whole pattern starts over again. However, the combination of vaccine mandates – which have started to have a positive effect as they expand to more businesses and organizations – and so-called “vaccine passports,” already established in Québec, coming to Ontario, and being considered in several U.S. states, which restrict access to non-essential activities such as bars, restaurants, sporting and music events, and more. Ultimately, it will be the level of vaccination that will move COVID-19 from a pandemic to an endemic virus that will likely require periodic boosters, similar to an annual vaccination for influenza, pneumonia, or other already-available prevention techniques.

Watching the Rt numbers, though, is an important way to cut through the media hype and misinformation (both intentional and well-intentioned) that continues to dominate our discussion about what the future holds for the COVID-19 pandemic.

Our exit strategy is death

Craig Newmark founded Craigslist in 1995. A dozen years later – at which point Craigslist was a senior citizen by venture-capital-fueled Internet standards – Newmark was asked for the umpteenth time, “When are you going to sell your company and cash out?” In other words, what was his “exit strategy?”

His response was (more or less) “My exit strategy is death.” He’s repeated that sentiment many times since. He retired from day-to-day operations at Craigslist in 2016 to focus on his philanthropic efforts (Forbes estimated in 2017 that the company was worth $3 billion, making Newmark – who’s assumed to own at least 40% of the company – worth around $1.3 billion or so.) He has no plans to sell Craigslist to any of the many suitors who’ve come calling over the years, nor is he interested in further monetizing the site, which remains pleasantly archaic in appearance and functionality in 2021, yet likely brings in several hundred million dollars per year.

Newmark’s approach continues to be refreshingly offbeat in a world where the quick buck is treasured. But his plan was based on real information, which we used to call “facts,” and led him to make the decisions that turned Craigslist from a San Francisco-based mailing list helping locals find Bay Area arts and entertainment events to a global company that was truly one of the Internet’s first “social networks.” (Fortunately, his stewardship of Craigslist meant it didn’t turn into a platform with the power to destroy American democracy itself, *cough* “Facebook” *cough*.)

On the other hand, it’s becoming increasingly apparent that the endless misinformation campaign, led by right-wing media such as Fox News, OANN, and Newsmax, as well as *cough* “Facebook” *cough*, against the COVID-19 vaccines has been wildly successful in convincing nearly half of Americans that the vaccines are part of a widespread government plot to steal their guns, make pedophilia legal, and turn the U.S. into a freedom-hating socialist state. COVID cases are rising nearly everywhere as we’ve seen vaccination rates taper off and everyone, including the non-vaccinated who should still be wearing masks, tossed those masks away and started hanging out in crowded restaurants and bars again. Only this time, nearly everyone who is being hospitalized or dying from COVID-19 is unvaccinated:

Appeals to reason and common sense haven’t worked. Most of those who say they still won’t get vaccinated overlap very neatly with those who think the 2020 election was stolen from Donald Trump. Despite all the evidence that the vaccines are safe and effective (and that there was no “steal” to be stopped when it comes to the presidential election), these folks are sure that’s all “fake news.” Until it isn’t:

Maybe if enough unvaccinated people start becoming seriously, even permanently, ill – or perhaps even dying from COVID-19 – it will start to convince the doubters that the only way we can escape endless waves of COVID-19 infections is by getting shots in the arms of everyone who can be vaccinated. Nothing else seems to be working to change minds.

It’s a morbid thought, but perhaps our best shot at an exit strategy from COVID is death.

Keep following the data

It’s been awhile since I took a deep look at the data for COVID infections and vaccinations in St. Clair County and in Michigan, so here we go:

  • According to Bridge Michigan, case numbers have fallen to a point in Michigan where the state will now only update their statistics twice a week. During most of the pandemic, MDHHS had daily updates; they stopped reporting on Saturdays a few weeks ago, and now will only update on Tuesdays and Fridays. The July 2 update showed 101 new cases, up from 40 the week before, but still very low. To compare, at the peaks in December 2020 and in mid-April 2021, the state was reporting over 7,000 new cases per week.
  • Unvaccinated people account for almost all new hospitalizations from COVID-19, as well as nearly all deaths from the virus. In a study released by the Cleveland Clinic, of the 4,300 COVID patients admitted to their facility between January and April of this year, 99.75% were unvaccinated against the virus. Also notable: “The study also looked at 47,000 Cleveland Clinic employees who had received one shot, two shots, or no shots. Among those, 1,991 tested positive for the coronavirus in recent months. About 99.7% of those who contracted COVID-19 weren’t vaccinated, and .3% were fully vaccinated.”
  • In another study at the Cleveland Clinic, over 52,000 employees, those who had already had COVID and those who hadn’t but had been fully vaccinated had almost no chance of getting COVID. Specifically, “The cumulative incidence of SARS-CoV-2 infection remained almost zero among previously infected unvaccinated subjects, previously infected subjects who were vaccinated, and previously uninfected subjects who were vaccinated, compared with a steady increase in cumulative incidence among previously uninfected subjects who remained unvaccinated. Not one of the 1359 previously infected subjects who remained unvaccinated had a SARS-CoV-2 infection over the duration of the study.

Here’s the U.S. map from covidestim.org for July 3:

Missouri, Arkansas, and northeast Texas are dealing with a flare-up, but the rest of the country, including Michigan, is fairly quiet.

The Rt values for several states are above 1.0 again. Rt is the average number of people who will become infected by a person infected at time t. If it’s above 1.0, COVID-19 cases will increase in the near future. If it’s below 1.0, COVID-19 cases will decrease in the near future. Michigan’s Rt number has remained steady at around 0.85 for several weeks. It will be interesting to see if the number rises after the Fourth of July weekend; if it does, it could be concerning, especially for hospitals and clinics that could see a small surge in COVID cases. If it doesn’t rise significantly, however, it would be an excellent sign moving forward.

Per covidestim.org, St. Clair County’s Rt number is 0.62, lower than the state number despite only about 49 percent of county residents being fully vaccinated (52 percent have had at least one dose of the two-dose vaccines, which still provides good protection). 46 percent of St. Clair County residents have had COVID already, again per covidestim.org. While there’s still much to confirm, if you add the percentage of those who’ve already survived a bout with COVID to those who are fully vaccinated, you start to approach 100 percent of the county having at least some protection against the virus (admittedly, some people have both had COVID and gotten the vaccine, so a simple addition – which would result in a 95 percent number – is too simplistic). But you start to see why numbers remain low, even when vaccination rates are much lower than you would hope and no one seems to be wearing masks in public, vaccinated or otherwise.

The vaccines work. Even Jim Justice, Republican governor of West Virginia, knows what’s up.

Still, lots of people, including many of our elected representatives, are either stupid or intentionally pandering to ignorant people (why not both?):

None of the above guarantees that the Delta variant – or some future variant – won’t be a problem. But working with existing data, I think it’s important to avoid the cherry-picking of bad stories, some of which are anecdotal in nature, that we keep reading in the news every day. Yes, people will continue to get sick from COVID, and some will get seriously ill and die, but at this point that seems to be almost exclusively limited to those who cannot – and more importantly, have chosen not to – get vaccinated.

COVID-19 and the flu, revisited

COVID-19 isn’t the same as influenza. But the way we end up dealing with it long term may be.

As I’ve noted previously, the Rt number, representing the the average number of people who will become infected by a person infected at time “t”, is falling nearly everywhere in the U.S., due to our nation’s superior access to vaccines. (This is not happening in other parts of the world, though, which will allow the coronavirus to potentially continue to evolve into new variants over time.) Not all Americans are willing to get vaccinated, unfortunately, and it now appears that it’s unlikely we will reach the “herd immunity” numbers (whatever level that might be, since it’s inconsistent from one expert to another) if only vaccinated people are counted. (Since those who’ve already had COVID at least once also have some immunity – though it’s not clear yet how effective that is or how long it lasts compared to a vaccine – some calculations of “herd immunity” include those people, which brings us a lot closer to the typical 70 to 75 percent number.)

But as long as Rt continues to fall and remain low, many of our restrictions should be able to be relaxed or lifted. The risk will remain, especially for those who refuse to get vaccinated, but our social lives could return to something close to normal this summer.

The risk of not reaching “herd immunity,” though, is that there will still be a large number of people who potentially could contract COVID, and particularly an existing or yet-to-emerge variant that is more contagious and possibly more resistant to the existing vaccines. Our therapies for COVID patients have improved, so if there isn’t a huge spike down the road that overwhelms our healthcare system again, COVID could become endemic in a similar way to many viruses that haven’t been eradicated but are largely controlled thanks, in large part, to vaccines. This includes measles, chickenpox, and (in most years) influenza. An annual COVID shot seems likely.

The other similarity between COVID and flu, even now, is the reluctance of people to get a vaccine that promises to protect us from illnesses that, while often mild and annoying, can become serious or even deadly. Historically, only about half of Americans who could get a flu shot each year do so. Nobody likes getting the flu, which can last from hours to days or even weeks. Yet we don’t take the time to get even the partial protection offered by the seasonal flu vaccine. I’m guilty of this myself; for years, I never bothered to get the shot, not because I didn’t believe in the science, but because I didn’t think I really needed it. I was (relatively) young, healthy, and figured I’d just ride out a case of the flu if I got it. There’s a word for that: arrogant. And an even better one: stupid.

Are we absolutely sure that the various vaccines are safe in the long run? That there are no side-effects? Well, no. But I think it’s adorable how many people are suggesting that the vaccines aren’t safe who are still smoking, or overeating, or over-drinking… while also taking other over-the-counter drugs or eating packaged foods without knowing exactly what the ingredients are or how they were manufactured. We have a lot of faith in the production of our food and pharmaceuticals otherwise, and justifiably so. What’s so different about the COVID vaccines? They’re tied up, unfortunately yet inextricably, with our current political civil war.

Update on COVID statistics

A couple of weeks ago I wrote a post that included a lot of maps showing the infections per 100,000 persons throughout the United States. A month ago, on April 1, Michigan was glowing white as the worst state in America for COVID infections:

COVID-19 infections per 100K persons as of April 1, 2021

Here’s where we are a month later, on May 4 (maps from covidestim.org):

COVID-19 infections per 100K persons as of May 4, 2021

The Rt number in Michigan is up slightly from April 26, from 0.68 to 0.71, while the Rt number in St. Clair County has fallen to 0.46 from an adjusted 0.54 on April 26. The estimate of those who’ve already had COVID in the county is up to 45 percent, compared to our neighboring counties of Macomb (50%), Sanilac (49%), and Lapeer (38%).