Cancer is a sign
of a long life.
against 2.9% in Central African Republic
That gap is not a verdict on Danish oncology. Cancer is mostly a disease of age, and in Central African Republic too few people reach the age at which it strikes — they die earlier, of other things. Across 185 countries, the share of deaths from cancer goes with how long people live more closely than with how rich their country is. These are 2021 figures, a year in which COVID-19 added deaths to every country’s total; read the shares with that in mind.
Two worlds
The single largest cause of death in each country, 2021 — a COVID-19 year.
Two groups lead in almost the same number of countries. One caution first: in 2021 the infections, maternal and newborn group includes COVID-19, which WHO counts as a communicable disease. In Peru the group is 53.6% of deaths, against 10.8% for heart disease; even in the USA it is 16.5%. So the split is not simply long lives against short ones: 17 of the 83 infection-led countries had a life expectancy above 70. The group ranges from 2.6% of deaths in Finland to 66.4% in Chad.
The clock, not the wallet
Each dot is a country: life expectancy against the share of deaths from cancer, both 2021. 185 countries, correlation +0.87.
BULLDOZER STATS azenha.ai/bulldozer
Richer countries have a larger cancer share too (correlation with income +0.78), but income and longevity move together, so it helps to separate them. Hold income constant and the cancer share still tracks life expectancy strongly (partial correlation +0.64). Hold life expectancy constant and the association with income is much weaker (+0.21). Most of what income and the cancer share have in common runs alongside longevity. These are correlations across countries in one year, so they show which measures go together, not what causes what.
Heart disease is not the rich world’s disease
Median share of deaths by income decile, poorest (1) to richest (10), GDP per capita PPP 2021.
Cancer climbs with income almost without a break, from 5.2% to 25.1%. Cardiovascular disease does not: it peaks in decile 8 at 34.5% and falls back among the richest. The world’s highest shares are in Bulgaria (54.9%, income decile 8), Uzbekistan (53.5%, income decile 4), Ukraine (52.8%, income decile 6) — none of them in the richest decile. In Japan 28.4%, Norway 25.0%, Switzerland 28.6%, it is well below that. Dementia and other neurological causes follow the rich-world pattern most sharply: a median of 2.9% of deaths in decile 8, 10.3% in decile 10.
Where we know least
WHO grades each country’s death-registration data from 1 (complete) to 4 (none usable). Share of the population of the 185 countries, 2021.
For 74 countries, home to 65% of these people — India and China among them — the shares on this page are WHO’s modelled estimates, not counted deaths. 48 of 54 African countries are in that group. And the countries where registration is weakest are the ones where people die youngest: average life expectancy is 77.2 years in grade-1 countries and 64.7 in grade-4 countries. Part of the pattern above is therefore a property of the model that fills the gaps. It is the best estimate there is, and it is labelled as such on every dataset page.
What the clock does not explain
Diabetes has no income gradient at all (correlation +0.00). Its map is mostly islands: Fiji 18.1%, Mauritius 15.1%, Tonga 13.8%, against 0.3% in Ukraine. Something other than age or wealth draws that map. Diet and genetics are the usual suspects, and nothing in these numbers settles which.
Source: WHO Global Health Estimates 2021; life expectancy from the World Bank (UN World Population Prospects), income and population from IMF WEO. Medians of an even-sized group average the two middle values. Explore the data →