Years ago, I worked for a health charity that sent out fundraising letters with a message on the envelope proclaiming “One out of every two people who open this envelope with die from [insert disease]!” The folklore was that it even became a joke on late-night US television, as in “who would want to open the envelope if this is what will happen?”
At the time, I tried to get through to the TPTB at the organization that this was a distortion of the facts. At that time, about 48% of Canadian deaths had the disease in question listed on the death certificate. That means that almost half of Canadians who died in any one year were considered to have died from that disease. That doesn’t mean that half of the entire Canadian population of that time would die from the disease – there’s a difference between all Canadians and those Canadians who die in any one year.
Two decades later and the organization in question is still making this sort of claim (although the numbers have shifted). And I still see it as inaccurate and a personal pet peeve.
The problem with the claim is that it confuses two different populations: all Canadians and that sub-set of the population that dies in any one year. There are not the same, primarily because not all people have the same risk of dying in any one year. Children have a much lower risk of dying than adults and men typically have a higher risk than women. The following chart is taken from an article published by Bandolier (a really great site) and shows annual risk of death by age and gender in the UK. The pattern is similar in most western countries.
So you can see the lovely J-shaped curve and the epidemiologic issue. In any one year, a male between the ages of 25-34 has a risk of 1 in 1,215 of dying, while a female in that age group has a risk of 1 in 2,488. In comparison, a male between the ages of 75 to 84 has a 1 in 15 risk of dying and a female 1 in 21. Big difference. So the people dying in any one year are not representative of the general population: they are a sub-set of the population that is strongly skewed to the older age groups.
The other thing to consider is cause of death. Let’s look here at the top three causes of death for Canadians at different ages: 2
- Under 1 years of age: congenital abnormalities, short gestation, maternal complications of pregnancy
- 1 to 14 years: accidents, cancer, congenital abnormalities
- 15 to 24 years years: accidents, suicide, and a tie for third place between cancer and homicide
- 25 to 34 years: accidents, suicide, cancer
- 35 to 44 years: cancer, accidents, suicide
- 45-54 years: cancer, heart disease, accidents
- 55 to 64 years: cancer, heart disease, accidents
- 65 to 74 ears: cancer, heart disease, chronic lower respiratory disease
- 75 to 84 years: cancer, heart disease, stroke
- 85 and over: heart disease, cancer, stroke
The point of this is that younger people have a much lower risk of dying and, more importantly, when they do die, they typically die from different things than older people. Again, a stat developed by looking at deaths (“1 in 3 deaths are due to x”) is not transferrable to the general population (“1 in 3 Canadians will die from x”).
Of course, what we’re actually talking about here is the gap between statistics and marketing. For marketing and fundraising purposes, health charities and other organizations need simple messages that are supposed to be strong and – let’s face it – scary. A message like “1 in 3 Canadians who die will die from our disease” is considered too complicated and obtuse for mass communication. I get that – I don’t like it, but I get it.
Take-away message/bottom line
In trying to create marketing messages, health statistics can sometimes be mis-interpreted. The risk of dying from any one cause differs widely among the general population by age and gender.
1 Risk of death by age and sex. Bandolier “Evidence based thinking about health care” http://www.medicine.ox.ac.uk/bandolier/booth/Risk/dyingage.html
2 Statistics Canada. Leading Causes of Death in Canada, 2007. Table 3 Ranking and number of deaths for the 10 leading causes by age group, Canada, 2007. http://www.statcan.gc.ca/pub/84-215-x/2010001/table-tableau/tbl003-eng.htm