Why bother with Covid-19 vaccination?

October 20, 2021 - Dave Angelini

Should you get a vaccine to protect yourself from Covid-19? Should you vaccinate your kids?

The simple answer to these questions is, yes. Without reservations and unequivocally. (Summary figure)

To understand the reasoning behind this conclusion, please read on.

---

You may have heard vaccine advocates touting the numbers for effectiveness. Sometimes this information is split up by age group. You may also have heard vaccine skeptics talking about serious adverse effects from vaccination. These arguments rarely cite the same data, and often use different scales to convey risks.

Here, I wanted to present information on a level playing field, to allow anyone to compare risk.

The basics

1. What is risk?

First, what do doctors, scientists, and epidemiologists mean they talk about risk? While there are precise statistical definitions of risk, the risk of a disease can be measured pretty easily. For most of this post, I will use a simple, real world metric: hospitalization. Most people would like to avoid hospitalization.

An illness that lands you in the hospital is likely to be pretty serious.

Intravenous fluids, supplemental oxygen, respirators, huge medical bills. I won't even consider the risk of death here, but that typically follows behind risks of hospitalization.

Hospitalizations are also well documented, and the CDC collects data on hospitalizations due to infectious diseases.

Units of risk

There are many ways we could talk about risk. I am going to use one unit commonly considered by epidemiologists: the number of people effected in a group of 100,000. So we'll mostly consider the number of hospitalized people per 100,000. This allows us to make a fair comparison across groups that may contain different numbers of people, such as different age categories.

Age groups

Speaking of age categories, how should we divide them up? Unfortunately, different government agencies sometimes report data by binning people into different age groups. I've limited this presentation to some groups relevant to the discussion of covid vaccinations:

• 5−11 year-olds (not currently eligible for vaccination)
• 12−17 year-olds
• 18−49 year-olds
• 50−64 year-olds
• 65−74 year-olds

I've omitted children under 5 and adults over 75, because they often have age-related health issues that might confound our consideration of covid-related issues.

Keeping everything in annual terms

Covid-19 is a rapidly moving pandemic. Sometimes rates are reported for the week, the month, the current "wave". But for diseases that are endemic, like flu, rates of hospitalization are typically considered year-by-year. Therefore, to make everything comparable, I will put all the data we discuss in annual terms. For rates of flu, this means choosing a representative year. I'll use 2019, the last full year before the covid-19 pandemic hit the US. For covid-19 data, I'm going to take the last 6 weeks for which data are available: August 30 through October 10, 2021. September is likely to be as "normal" a month as we could get, in terms of weather, school attendance, holiday travel. Plus, it's the most recent!

Correlation and causation

Statisticians will be careful to remind us that just because two things occur together, we cannot assume one causes the other just based on that association. This is especially true of big data sets containing rare events. There are a lot of people in the United States. 329.5 million according to the 2020 census. It's very likely that someone will be hospitalized in a close space of time to any thing we might want to focus on, say having just caught a particular virus, or having just received a vaccine. That coincidence doesn't, by itself, evidence a cause. But larger numbers are more suggestive of a connection. Ultimately, biologists look for a causal mechanism in order to draw conclusions with confidence.

2. Who do we trust?

I am going to use data from US government agencies: the Centers for Disease Control (CDC) and the Vaccine Adverse Event Reporting System (VAERS), a less well-known office that is co-managed by the CDC and the Food and Drug Administration (FDA). I will link directly to those sources of data. And I'll put fold-outs to provide more details info for you to follow-up, if you want to dive deeper.

Can you trust the US government? Federal agencies like these are staffed by career professionals. These people often spend decades at these offices, spanning multiple presidencies. Typically they only make news in the unusual event that politicians try to influence their data gathering or conclusions.

Epidemiological data is decentralized, in the sense that hospitals and state CDC offices around the country send in local data. Data are quickly made public. Academic and commercial researchers routinely comb through these data. These facts would make it very hard to manipulate the data significantly, and most people would have strong incentives to raise an alarm if they found any manipulation.

Still aren't willing to trust government data? If so, we aren't really left with anything to talk about. Because there is no other comparable, nation-wide source of information. Anyone on cable news or on the internet, telling you they have a better source of information about rates of disease is fooling themselves, lying to you, or both.

By the way, who wrote this? I am a college biology professor. I am trained in genetics, developmental biology and evolutionary biology. I have a professional understanding of statistics. However, I am not an epidemiologist, not a virologist, and not a medical doctor. I have been financially supported by US federal agencies, including the USDA, NASA, NIH and NSF. I have not worked directly with CDC or FDA. Importantly, I think, I have 13 years of experience explaining complicated ideas to people who may be new to science.

The risks

So, what's the risk, right now, of hospitalization with covid-19?

(Data source: COVID-Net)

How should we read this graph? Well, for example, it's saying that if the current conditions (based on data from August 30 - October 10, 2021) represent this year as a whole, then we can expected 2000 hospitalizations of people between the ages of 50 and 64 during the year. It's often been said that covid is more serious among older people, and we see that trend here too. But there are still hospitalizations of children and teenagers. Now, at year's end, this number might not be perfectly precise. But it will probably be close. These 6 weeks have been pretty typical for 2021 in the US so far. Hospitalizations have not been as high as last winter, and not as low as the early summer.

Are there exceptions? The risks shown here average across the entire United States. You may live in a place where risks are higher or lower. But nowhere has been untouched by covid-19. These data also combine people who are healthy with those who may have previous serious health problems that increase their risk. Certain medical conditions can increase the risk of serious illness from covid-19. Asthma and other chronic lung conditions, heart conditions, compromised immunity, diabetes, obesity, and a history of smoking all increase risks above the averages shown here. However, the variation isn't so great as to make healthy people free from risk of serious covid-related illness, including hospitalization and death.

How to put risk in context? Is 2000 hospitalizations per 100,000 people annually a lot? To answer that question, we need to have a more familiar thing to compare against. Here are some other annual rates of injury in the US:

• The US averages 16 shark attacks per year, in a nation of 329.5 million people (source). That's 0.00486 shark attacks per 100,000 people.
• In 2019, lightning killed or injured 120 people in the US (source). So, 0.0364 lightning casualties per 100,000.
• There were 579 deaths or serious injuries that year due to tornados in the US (source): 0.176 tornado casualties per 100,000.
• Car accidents caused 11.0 deaths per 100,000 nationally in 2019 (source).
• In 2017, guns were used to kill 39,773 people in the US, making that death rate 12.1 per 100,000 (source).

These examples can help contextualize the risks of covid-19. However, these examples are rather specific. Not everyone in the US spends time swimming near sharks. Lots of people live where tornados almost never occur. Not everyone drives or rides in a car.

We tend to think of infectious diseases as different, in the sense that they are more universal in who they can affect. For that reason, I will compare covid-19 with two other infectious diseases and their vaccines: the flu and measles.

Here's how the risk of hospitalization compares for measles, flu and covid-19.

(Data: Fiebelkorn et al. 2015, J Ped Infec Disease Soc, CDC, COVID-Net)

The scales of these 3 graphs are all the same. You can see that the rates of measles hospitalization, in any of these age groups, barely registers here. Measles was considered eradicated in the US in 2000, although occasional out-breaks have occurred, mostly due to international travel. Flu is endemic, and every year 100,000's of people are hospitalized for severe cases, and 10,000's die.

However, the impact of the flu is dwarfed by covid-19. By putting hospitalizations in the same terms (annual cases per 100,000 people) it's obvious that covid-19 is easily 10-fold more serious. Children are at a similar risk of hospitalization from covid-19, as adults 50-64 year-olds are from the flu.

Rates of hospitalizations, for all age groups, are higher for covid-19 than for all of the examples above – shark attacks, lightning, tornados, car crashes, and guns – combined.

Careful readers will notice that hospitalization rates for these diseases are reported for age groups with slightly different divisions. That's just due to the arbitary decisions made by different disease survailence programs. The actual ages are at the bottom of each plot. But to make comparisons easier, I've color-coded roughly similar groups, with "kids" in light green and older people in dark red.

The vaccines

Some people may acknowledge the dangers posed by covid-19, but remain skeptical of the effectiveness or dangers of vaccination.

If you consider yourself a vaccine skeptics, the CDC, the Mayo Clinic, and the World Health Organization have answers to common questions and concerns. I won't go into the many misconceptions about vaccinations, but I will address a few very quickly.

• Vaccines do not cause autism. That idea was introduced by one medical article in 1998. It was based on incorrect interpretation of previous data (source). Decades of follow-up research have found no correlation between vaccination and autism and no causal mechanism has been identified. (Read more: 1 2)
• No covid-19 vaccine can give you covid-19. While there are several vaccines available, none of those in the US is made with live virus, which is the only type of vaccine with any possibility of infection.
• Using vaccines to track people is science fiction. It's simply not biologically possible to do this. The idea is evocative tough, and persists in the public imagination. It may help to remember that this idea comes from a plot line from The X-files in 1996.
• Is the speed with which the covid-19 vaccines were developed suspicious? mRNA vaccines, like those from Moderna and Pfizer, have been developed over the last 10 years. One big advantage they have is that they can be customized quickly. So while covid mRNA vaccines are new, mRNA vaccines in general have been under careful and rigorous development for a long time.

More importantly, I want to address two questions here:

• Do vaccines against covid-19 work?, and
• How risky are negative reactions to the vaccines?

Covid-19 vaccines work very well

The first vaccine to be developed in the US was produced by the company Moderna, with funding almost entirely from the US National Institutes of Health (NIH). To determine the effectiveness of the vaccine, the NIH conducted a clinical trial. They recruited about 30,000 volunteers. Half of them received the vaccine. The other half received a placebo. Both groups received two doses and were followed over several months, using rigorous methods to test for covid-19.

The graph below shows infections were dramatically lower in the group that received the vaccine.

Data from Baden et al. 2020, New England Journal of Medicine

Compared to the placebo group, which had no protection from the vaccine, the group of vaccinated volunteers had 94% fewer infections. These results demonstrate that the Moderna vaccine is 94% effective at preventing infection. Even better, there were no hospitalizations among people in the vaccinated group during the clinical trial. That's 100% effectiveness at preventing hospitalization! These results are similar to the other vaccines available in the US, and the high effectiveness of the vaccines has been consistent even as they have been distributed to millions of people.

Does the delta variant change things? Delta is more contagious than earlier versions, and it is most common among unvaccinated people. This suggests that vaccines also confer resistance to Delta, although new clinical trials are under way now to rigorously test that possibility. Delta may also be more likely to cause hospitalization, increasing the risks posed by covid-19. Our current vaccines are still the best protection against Delta.

What about break-through infections? Infections in vaccinated people are rare. A recent study found the rate of break-through infection was 20 per 100,000 (source). Importantly, those positive cases were very mild. None required hospitalization, and many people were completely assympotmatic, only knowing they were positive due to a sensitive diagnostic test.

Adverse effects of vaccines are very rare

So far, you may acknowledge that covid-19 is serious problem and that vaccines exist that can protect against the disease. But the danger posed by vaccination may still cause you anxiety.

Doctors in the US have a routine process to document any "adverse events" that are associated with vaccination, the Vaccine Adverse Event Reporting System (VAERS). This process is used with all vaccines. And it includes any "adverse event" that comes to the attention of a doctor, anything from soreness or itch to hospitalization and death. Therefore, it's possible that some events may not be caused by the vaccine, but by some other health problem or environmental exposure that just happened to come near the time of the vaccination in a person's life. That means that the data VAERS collects is likely to over-estimate any potential negative effects of vaccination. In this way, it is an extra-conservative, extra-cautious system.

About 91% of children in the US are vaccinated against measles, with the MMR or MMRV vaccines (source). (These vaccines also protect against other viral pathogens at the same time.) Adults are also encouraged to get a booster at some time between the ages of 19 and 65. Every year hundreds of adverse events are reported to VAERS in association with MMR(V) vaccination. In 2019, among all age groups, there were about 0.46 adverse events per 100,000 people, in association with MMR(V). Hospitalizations in association with an MMR(V) reaction were much more rare: only 0.0024 per 100,000.

The risk of hospitalization associated with the MMR(V) vaccine is less than the risk of shark attack.

How do the rates of hospitalization associated with reactions to vaccination against measles, flu and covid, compare to one another and to the risks posed by those diseases?

Data from VAERS. Rates were calculated using vaccination coverage data from CDC 1 2 3.

These graphs have the same vertical scale we've seen earlier, when we saw the risk of covid-19 hospitalization. The obvious conclusion is that the risk of hospitalization due to any of these vaccines, for any age group, is very, very low.

One potential objection is that, hospitalization is a pretty serious condition. Perhaps people have negative reactions to vaccines that are serious, but do not rise to the level of hospitalization. Below are comparable graphs, including all adverse event reports associated with vaccination.

Data from VAERS. Rates were calculated using vaccination coverage data from CDC 1 2 3.

Here we see higher numbers for covid-19. But remember, we are now including even minor events, such as fatigue and soreness.

Finally, let's make the most important comparison: side-by-side, the risks of hospitalization for covid-19 infection and covid-19 vaccination.

This image should close the case. In every age group, the risk of hospitalization is far higher for contracting covid-19 than for the adverse effects of vaccination.

So, get vaccinated!

While vaccination reduces your own personal risk, I have not even discussed here the biggest advantage of vaccination. It significantly reduces your ability to spread the disease. Vaccination helps you, but it also helps everyone you interact with. And it is essential for people, young and old, who cannot be vaccinated because of health conditions affecting their immune system.

(Summary figure)

(Analysis code used to standardize data and make all the graphs, in R)

---