Tuesday, October 20, 2020

Maintaining a Sense of Perspective and Proportionality During the COVID-19 Pandemic (Post #2- 6 months later)

 

It has now been over 10 months since the onset of the current pandemic and many important empirical insights concerning how infectious and lethal COVID-19 is have been established in that time.  This post is the second installment on my earlier post from late March.

 Despite the transition from being a completely “unknown” virus to a “somewhat known but still largely unknown” virus over the past 8 months, parts of Canada and the UK are returning to the lockdown measures of March so I think it is time to put these issues in a broader public health perspective.

I will focus on recent data from the US, since the US remains one of the countries hardiest hit from the virus and it has easily accessible mortality data.  My usual disclaimer applies- this pandemic is very serious (though not close to the most serious public health predicament of the past century). 

There are many serious public health predicaments (e.g. obesity, cancer, mental illness, etc.) and so the response to any specific health risk (e.g. severe COVID-19) must consider both the costs and benefits of the efforts to mitigate those risks AND ensure they are a proportionate response given the other health challenges facing a population.

Two particular types of recent data are helpful for putting this pandemic in the health context of the United States- (1) the total number of Americans who have died to date vs the number of deaths in non-pandemic years, and (2) the percentage of deaths involved with COVID-19 in different age groups across the lifespan.

Now of course mortality does not tell us everything about the severity of a disease, but it is a vivid illustration of the heavy toll it can have on a population.  Cancer, for example, kills about 600 000 people in the United States every year.  But the number of people who survive cancer is estimated to be about 18 million people a year (
https://www.cdc.gov/cancer/dcpc/research/articles/cancer_2020.htm). 

Cancer survivors can carry the toll of their cancer diagnosis for years- financial, emotional and health-related.  To fixate solely on cancer deaths as exhausting the “burden” of the disease would be an understatement.  The same is true of COVID-19 and influenza.  But the number of deaths a specific disease causes is a very significant factor, it is an essential element in determining the story of the severity of different health challenges.

From January through to mid-October the CDC has reported 2,203,637 deaths in the United States.  Is that number unusually high?  There are still 10 weeks left to go in 2020, but the reported deaths in 2018 was 2,839,205 deaths,  2017 was 2,813,503 deaths, in 2016 it was 2,744,248.

So far 2020 does not stand out as having an unusually higher than normal number of deaths.  If the numbers of death for the remaining 10 weeks approximate the average of the last 40 weeks, then 2020 will have an average death toll.  (Aside- Of course some public health officials warn that the weeks to come will be “the darkest yet” of the pandemic, but that has been the headline rolled out for almost every month of the last 8 months now, so take it with a large dash of salt!)

The fact that 2020 is not an exceptional year for total deaths will strike most of us as surprising, even shocking, given that when you turn on the evening news you could easily form the impression that unprecedented numbers of Americans will die this year, and most from COVID-19.  Both of those assumptions are not based on the facts, they are perceptions skewed by hearing daily news updates on COVID-19 infections and mortality.

Now I know what the critic might will retort-- that it is possible that the reported deaths underestimate COVID-19 deaths- see here.  But when such cases fixate so much on the numbers when COVID-19 deaths were at their highest in April (because the virus spread among the elderly (e.g. in nursing homes)) I think some healthy scepticism is warranted.  Significantly higher rates of COVID-19 infection are reported daily now (and have for the past 4 months!) in the United States compared to April, and yet the deaths have remained much lower than in that peak month of reported deaths.  Assuming there are not massive outbreaks of the virus in nursing homes again, I think it is reasonable to assume we will not see higher death tolls than what occurred in April (which was before the vulnerable were shielded).  

I will admit it is still too early to come to any definitive conclusion about COVID-19 mortality this year.  It may take another 2 years before we really get a sense of what is going on.  But even still, I think it is imperative we assess the data we have to date, and modulate our approach to mitigating the risks from the virus accordingly.

When you look at the mortality data from the CDC you see there are 3 categories of deaths, deaths involving pneumonia (with or without COVID), deaths involving COVID-19 and pneumonia but excluding influenza, and deaths from influenza with or without COVID-19 or pneumonia.  So it is a complicated business and will take some time to sort out!  

But based on the data now in hand, it is clear, at least to me, that this is not an exceptional year for death in the United States.  That message contravenes everything you will likely read in the news headlines.  (Aside- perhaps I will write another post later on why the media has become some fixated on this story).

What is also important to keep in mind, from a public health perspective, is that COVID-19 is not the only cause of death.  The CDC helpfully provides the overall deaths and COVID-19 “involved” deaths for different age categories.  I attach the image here:


Let’s compare the mortality numbers at three stages of the lifespan- childhood (age <15), middle age (age 35-54) and late life (age > 65).

For children ages 0-15, approximately 18, 446 children have died in the United States this year.  Of those deaths, only 74 deaths, or 0.4% of the deaths, involved COVID-19.  This means that the cause of over 99% of the childhood deaths this year in the United States did not involve COVID-19.  What did they involve?  In my last point I noted this graph, (source): 


 COVID-19 will not make the top 10 causes of death for children and adolescents (ages 0-19) this year.  That is really good news!  But at the same time it is also rather disturbing news.  Why don’t the top causes of deaths among the young ever make such an impact on the evening news, if they are ever reported on.  We hear a lot more about COVID-19 deaths, but we ignore 99% of the childhood deaths, like the 3000+ childhood and adolescent deaths from firearms, or 1100+ suicides among  the same age cohort. 

OK, let us move on to consider the mortality risks to the parents of these children and adolescents, adults ages 35-54.  For this age cohort, 195, 211 people have died this year. And 14,956 of those deaths, or 7.6%, have involved COVID-19.  The risk of death to these parents is considerably higher than to their children, but still over 90% of their deaths come from things that are not COVID-19. 

Finally let’s look at the grandparent’s mortality risks, those age > 65.  This year 1, 628, 625 people age > 65 have died in the US, and COVID-19 was involved in 160, 080, or 9.8%, of those deaths.  For even the age group most at risk of COVID-19 mortality 90% of those that died were killed by something other than COVID-19.  And yet how many news stories do we hear each night about the deaths from cancer, heart disease, suicide or accidents?  We do not hear about those other causes of death, many of which could be prevented by changes to behaviour and public health policies and priorities. 

I emphasize the above points, concerning the normal death toll in the US, and the fact that COVID-19 accounts for a relatively small portion of deaths this year, not to minimize the seriousness of the pandemic.  It is a big problem, just not the biggest health problem.  Our response to mitigating the risks from this pandemic, like any other risk, should be proportionate and backed by sound empirical evidence.  I don’t think we have a good track record of that to date.

Cheers, 

Colin