Kent Sepkowitz is a CNN medical analyst and a physician and infection control expert at Memorial Sloan Kettering Cancer Center in New York. The views expressed in this commentary are his own. View more opinion at CNN.
(CNN)An interesting moment in Fox News reporter Chris Wallace’s memorable interview of President Donald was a confusing back and forth about Covid-19 related deaths. The President claimed that the US had the “number one low mortality rate” in the world — whereas Wallace insisted the US had little to crow about.
Though it is clear that Trump is completely wrong
, comparing the impact of the Covid-19 pandemic on the mortality of each affected country turns out be a very tricky business. Because no one metric is perfect, especially when used to compare impact across countries, there is a large corner of epidemiology dedicated to how to most accurately tabulate death.
The Centers for Disease Control and Prevention (CDC) devotes a long entry to addressing which is the best metric
for a given situation in its online course
, “Principles of Epidemiology in Public Health Practice.”
As Covid-19 reporting has evolved, the death count has been the most cited number to track the size of the tragedy in the US and other countries. According to John Hopkins University, more than 148,000 human lives
in the US have been lost because of this disease.
Much time could and should be spent on understanding the death count, especially since a growing proportion of the US — now 31% — thinks the count is an overestimate
. In contrast, most public health experts feel strongly
that this is an underestimate.
The simple death count is not used in comparisons to other countries, due to vast differences in country population sizes. To account for this variation, most public health experts prefer to use rates, not counts.
The two most popular rate measures are the case fatality ratio (CFR) and the rate per total population (those with and without the disease) called the mortality rate. The Johns Hopkins Coronavirus Resource Center, which Chris Wallace and Fox News cited during the President’s interview
, offers users an easy toggle between the two measures.
The CFR measures the number of deaths divided by the number of people diagnosed with the disease. The advantage of this metric is that it is intuitive and is exactly what an individual person with Covid-19 wants to know — given the diagnosis, what is the likelihood I will die? The disadvantage is substantial because CFR can change depending on how much of a population is tested for a disease.
As an example, think of a country with 100 people where four people are hospitalized with Covid-19 and six others are diagnosed but are well enough to stay home. If no contact tracing or community tracing is done, only 10 people are known to have the infection. Three of these 10 people die, yielding a CFR of 30%.
Now consider if the same country with that population of the 10 infected people above performed contact tracing and community testing and diagnosed 10 additional persons with no or mild symptoms. As with the first example, three people die. But here the CFR is 15% (three deaths in 20 known cases). In other words, a higher testing rate leads to a lower CFR — a point that might make a testing fan out of the President.
To circumvent the large problem introduced by diagnostic test volume, the mortality rate, in contrast, uses total population (those with the disease — diagnosed or not — AND those without the disease). Everyone. Period.
This metric has little value to a patient who wants to know his prognosis but measures exactly what public health authorities need to know — how bad is this in my country compared to others. Using the above example of the 100-person country with three Covid-19 deaths, the mortality rate would be
3%, no matter how many people were tested.
The Johns Hopkins data used by many news outlets, including CNN and others, presents a table with information on the 20 countries most affected by Covid-19. During the interview, Wallace showed CFR results (though he called them mortality rates) and said, correctly, that US is 7th (it is now 10th) among these 20 countries — nowhere near
“number one” lowest in the world.
Even worse? From the way Kayleigh McEnany quickly produced the tables upon thePresident’s request, it appears that the information used by the White House in the interview might be used in their decision-making as well.
Their statistics are pulled from another (non-American) source, Our World in Data compiled by
the UK’s Oxford University. This data uses CFR and, unfortunately for those hoping the President will see the unfiltered truth, allows anyone to select as many or as few countries as they wish to frame the crisis according to their specific goals.
Though it is difficult to discern in the interview video (at about 2:45 in), President Trump used a graph
that included information on only seven countries. In this group, the US ranked 4th (to my eye) with a CFR worse than Iceland, Brazil and South Korea. Again, even with the White House’s carefully curated data, the US is mediocre at best.
An even more sobering view of how the US is doing is our actual mortality rate
per 100,000 people (not CFR). Using this metric in the Johns Hopkins data, the US has the 17th worst mortality rate of the 20 most affected countries.
The number or rate of deaths from Covid-19 in the United States is a disaster no matter what metric is used. And it is clear from the interview that President Trump willingly presents misinformation he is given, even as the cameras are rolling.
But perhaps the most disturbing aspect of watching the interview is how strongly it suggests that the data being given to the President is deliberately incomplete.
He is informed, apparently, of how the US ranks in CFR against six other countries using the metric that public health experts view as inferior to the mortality rate.
At every step of the way, it seems, those entrusted to present the truth to the President have chosen instead to willingly protect their own well-being over that of the people of the United States.