Data from Czechia, Israel, Sweden, and U.S. Show an Exponential Growth of the COVID-19 Mortality with Age – the Fraction of People Dying with COVID-19 Increases About Three Times for Every Ten Years of Age

The published data on Covid deaths in age groups (ten years range) in four countries (Czechia, Israel, Sweden, U.S.) have been combined with the age distributions in said countries (see Table 9). It can be seen that the deaths numbers (absolute, not recalculated per million people) seem to form approximately geometric series with a common ratio of between 2 and 4 till the age of about 75. More consistent geometric series are obtained till still higher ages by recalculating the deaths numbers in each

age cohort per million people in said cohort (see the penultimate column in the Table). The last column in the Table shows the ratio of the deaths number per million in the age group divided by the deaths number in the adjacent younger age group (for example, dividing the values 48:13 from the penultimate column provides the first yellow-marked ratio in the last column of 3.7 – a common ratio estimation for a supposed geometric series).

The Czechia data correspond to an earlier epidemic stage (May 2020), with only 279 total deaths in the country, so that only the deaths values for people above the age of 54 are sufficiently high to be statistically significant. In the data for Israel, Sweden, and the U.S., with 960, 5880, and 194 000 total deaths (on the shown date of September 2020), respectively, the values are sufficiently high for ages above 44, 39, and 24, respectively. The ratio values derived for the ten years ranges with relatively reliable deaths values are marked yellow in the Table.

The three small countries (all having a population of about 10 million) show nearly the same average ratio value of about 3.5 + 0.1, despite different sample sizes. The U.S. show a lower value of 2.5. Thus, the fraction of people dying with COVID-19 in Czechia, Israel and Sweden increases 3.5 times for every ten years of age, and in the U.S. 2.5 times for every ten years of age. Roughly, the mortality (death per million) increases 3 times for every ten years in all countries.

The people 85 years old or older form 2% US population (=6.6 million), so that about 60 000 deaths in this age group constitutes less than 1% in the cohort. This is the probability that a person 85+ will die with Covid in the U.S.; the people around 75 years old have have about three times lower probability of dying with corona – i.e. only 0.3%, the people around 65 old 0.1%, around 55 old 0.03%, around 45 old 0.01%, and around 35 old 0.004%. So, any panic is indeed unfounded. Let us remember that about 1% of the population die every year regardless COVID-19 (see also here).

Goldstein & Lee (PNAS 117, 2020, pp 22035-22041) estimate that COVID-19 mortality rate in the U.S. increases by about 9.5% per one year of age; this provides (1.095¹⁰= ) 2.5 fold increase per 10 years of age – exactly the value as observed here in the Table, despite the fact that my data are from September and their from April. The publication further estimates the increase of 12% per year of age for South Korea, Italy, France, Germany, England and Wales, and Spain, which provides (1.12¹⁰ = ) 3.1 fold increase per 10 years. The increase of 3.5 for Czechia, Israel, and Sweden as observed here is higher (corresponding to 3.5^0.1 = 13% per year of age) than estimated for the above countries, but there may be factors specific for different populations, as noted also in the cited publication.

The above authors hesitated to relate to younger ages, since they did not have so many cases in April when preparing the publication, but they guessed that the mortality for the young would be less than predicted from the exponential relation found for the older ages. When reviewing the U.S. data in the Table (the other countries have too few cases for younger ages), it indeed seems that the ratio is higher than 3 for all groups below 35 years old. Shortly, the danger for young people and children, in contrast with influenza, is extremely low. It would be interesting to have case histories for the children who supposedly died with corona.