For those who wish to debate the lethality of COVID-19, it’s important to get the terminology right to avoid making apples to oranges comparisons – imprecise language in the infosphere is confusing and misleading. In trying to better understand various perspectives, I undertook to try to better understand the metrics of disease severity. When it comes to lethality, I learned there are three significant terms we need to understand and use correctly if we wish to have such discussions – mortality rate, case fatality rate, and infection fatality rate. Each provide a means of quantifying lethality, but for different purposes using different frames of reference.
- Mortality Rate. This is the metric I see most misused in mainstream media reporting. Mortality rate is correctly defined as “…a measure of the frequency of occurrence of death in a defined population during a specified interval.”[1] It is often expressed as deaths per unit population per time, such as “3.2/thousand/year” or something like that. At the moment, for COVID-19 the Worldometer[2] site is displaying deaths/million since onset; as of this post it’s 132/million[3] for the USA. You can also parse this down to per state or any other subdivision for which you have death counts (ex: New York’s mortality rate is 989/million). Mortality rate should be relatively accurate, although there’s always room to debate how causes of death are determined and reported. It is perhaps the best way to compare one cause of death to another and also seems to be the best way to assess the societal impact of deaths due to a disease.
- Case Fatality Rate. Often mislabeled “mortality rate,” the case fatality rate (CFR) is “…the ratio of the number of deaths divided by the number of confirmed… cases of disease.”[4] CFR tells us how many of the actually diagnosed patients with the disease die. Because the denominator in the ratio requires confirmed cases, CFR can vary greatly from place to place or time to time depending on the availability of testing and testing criteria. For example, if an area has the capacity to do a lot of tests and can detect more people infected with a pathogen, it will have a larger denominator and lower CFR. However, if testing capacity is limited and an area can only test people with significant symptoms, the denominator will be small because many people with minor symptoms or asymptomatic infections will never receive a confirmed diagnosis. Right now, per Worldometer the COVID-19 CFR for the USA is 5.42%.[5] You can also parse the CFR for demographic factors like age, gender, certain pre-existing conditions, etc. Because of the extreme inconsistency in testing for COVID-19, this is in my opinion a pretty unreliable metric for judging lethality or comparing COVID-19 with other diseases.
- Infection Fatality Rate. The infection fatality rate (IFR) is “…the ratio of deaths divided by the number of actual infections…” with a pathogen.[6] IFR attempts to quantify the percentage of people actually infected with a disease who die, and given that it is rarely if ever possible to confirm every infection from a pathogen, it is usually going to be just an estimate perhaps based on extrapolation from random data sampling. Hypothetically, if it were possible to test 100% of the people in an area, the CFR and IFR would be identical because all the infections would become confirmed cases. Like CFR, you can also parse IFR for demographic categories like age, gender, pre-existing conditions, etc.
We can make a very preliminary, very rough comparison of the lethality of influenza versus COVID-19 using mortality rate. The U.S. Centers for Disease Control and Prevention (CDC) estimate the U.S. death toll for seasonal influenza as 12,000 to 61,000 per year.[7] With a U.S. population of 329 million, that yields an annual mortality rate of 3.03 to 15.45 deaths per million per month. We only have four months of death data on COVID-19, but as of this writing Worldometer says it has killed 43,663 people in the U.S.A., yielding a mortality rate of 33.18 deaths per million per month so far this year – what it will be over the whole of 2020 remains to be seen (hopefully lower, as mitigation measures and better treatment reduce deaths). The University of Washington Institute for Heath Metrics and Evaluation model projects COVID-19 will kill between 45,000 (best case) and 125,000 (worst case) Americans through August, with a most-likely death toll of 65,967 (the model assumes mitigation measures like social distancing will remain in place).[8] We can discount the lower (best case) number because the disease has already killed over 43,000 and it’s unlikely deaths will suddenly stop. Those projections level at August, yielding an eight-month worst-case projected 2020 mortality rate of 31.67/million/month and most likely of 16.71/million/month. (Note that several states like Florida, Georgia, and South Carolina are relaxing social distancing requirements, so these projections may go up.)
Huge caveat about the above: I was unable to locate monthly death data for influenza for a January to April time period to enable a more direct comparison to COVID-19. Influenza tends to be seasonal, so we could make some assumptions that most of those deaths occur during the winter and spring but for purposes of an illustrative example (which is all I intend the preceding paragraph to be), averaging those over the course of a year is sufficient. More definitive annual mortality rate comparisons will have to wait until January 2021 when we’ll have a year’s worth of COVID-19 deaths.
Anyway, discuss and debate at will – just be sure to compare apples to apples and oranges to oranges.
[1] https://www.cdc.gov/csels/dsepd/ss1978/lesson3/section3.html
[2] https://www.worldometers.info/coronavirus/country/us/
[3] Since disease onset.
[4] https://www.virology.ws/2020/04/05/infection-fatality-rate-a-critical-missing-piece-for-managing-covid-19/
[5] https://www.worldometers.info/coronavirus/country/us/
[6] https://www.virology.ws/2020/04/05/infection-fatality-rate-a-critical-missing-piece-for-managing-covid-19/
[7] https://www.health.com/condition/cold-flu-sinus/how-many-people-die-of-the-flu-every-year
Place Just Right 