It’s one of my favorite questions to ask whenever I talk about Gapminder in class:
Why did South Africa write the letter “U” over the last thirty years?
gapminder.org/tools
It’s one of my favorite questions to ask for two reasons, and both are related to each other. First, hardly anybody gets the correct answer. Second, the fact that hardly anybody gets it pleases me no end.
That’s an unusual thing for a teacher to say. How can absence of knowledge make the teacher happy?
But in this case, it really and truly is a wonderful thing. And the reason it is a wonderful thing is because AIDS today simply isn’t as big a concern as it used to be at one point of time.
One way to understand this chart is to come to the sobering conclusion that South Africa is only now getting back to life expectancy levels it last saw in the early 1990’s.
But hey, at least it’s getting back to those levels, huh?
And in more excellent news, we learn via Saloni Dattani’s awesome (and very, very welcome) blog post that life expectancy for folks with AIDS who are on AntiRetroViral (ARV) therapy is about as much as it is for folks who don’t have AIDS.
Or, in simpler terms, AIDS seems to be a solved problem.
For people with HIV on ART and with high CD4 cell counts who survived to 2015 or started ART after 2015, life expectancy was only a few years lower than that in the general population, irrespective of when ART was started. However, for people with low CD4 counts at the start of follow-up, life-expectancy estimates were substantially lower, emphasising the continuing importance of early diagnosis and sustained treatment of HIV.
The reason I say “seems to be”, rather than “is”, is simply because I do not know enough to be able to make a definitive statement. But this much I can say with confidence: this is definitely very good news.
And god knows we could do an extra large helping of good news.
Please, do consider subscribing to Saloni’s excellent blog!
Quick post today, because I’m on the road – but an important post nonetheless. So important, in fact, that I hope to do a longer one in the near future.
But are you clear about what life expectancy means?
For example, here’s a chart from the World Bank, showing India’s life expectancy at birth in years:
Does it mean that we should expect a baby born today to live until the age of 67?
But how could we possibly know what a baby born today is going to get in terms of healthcare? Scientific advancements are so rapid that we couldn’t possibly know this. So what does life expectancy at birth mean, then?
It turns out that there are actually two different definitions of life expectancy, and to understand ’em, you really should begin with this picture:
Think of two groups of people. Let’s call the first group “Period”, and the second one “Cohort”.
Who is in the Period group? Everybody alive and in a particular region at a particular point in time. Say, for example, everybody alive in Bangalore in the year 2019.
Who is in the Cohort group? Everybody born in Bangalore in the year 2019, and this group tracked over time.
And, it turns out, you can define life expectancy for both groups.
“Period life expectancy is calculated by assuming people will experience the current year’s mortality rates at each age at the corresponding ages in their lifetime.”
In practice, this is how it is calculated:
Imagine there are 1,000 infants (under one year old)5, and in a particular year, infants had a death rate of 5 per 1,000. This tells us that 995 of them would survive to the age of one.
Now, imagine these 995 have reached the age of one. Based on the death rate among one year olds in the same year, we can now estimate how many might survive to the age of two.
We can then carry on this calculation for the entire hypothetical cohort of 1,000 infants. Along with this, we can keep track of how many years each of them survived.
Then, we can calculate the average number of years lived by the entire group of 1,000 hypothetical infants. This is equal to the period life expectancy at birth.
This means period life expectancy is a summary measure of death rates in one particular year, rather than a prediction of how long people will actually live
This can be confusing, I agree. And what’s more, I remember the confusion I felt the first time I tried to make sense of this myself!
Try this: imagine 1000 babies born this year, in 2024. If the number of babies that make it to 2025 is (say) 995, then we ask how many of those 995 babies will make it to year 2 (2026).
How could we possibly know that, you might say, given all the potential advances that will happen in healthcare in 2024. Agreed! What we do is, we assume that the death rate for 1 year old babies in 2025 is what it was in 2024. Apply that death rate (the death rate for 1 year old babies in 2023) to our 995 babies in 2025. So how many remain?
Whatever your answer, those will be the number of two-year olds from this cohort in 2026. How many two-year olds are expected to pass away in 2026? Again, how could we possibly know? So we assume that the death rate of two year old babies in 2026 is the same as it was in 2024. Apply that death rate, and we now have an answer for how many three year olds (from this cohort) will be around in 2027.
“Period life expectancy is calculated by assuming people will experience the current year’s mortality rates at each age at the corresponding ages in their lifetime.”
Period life expectancy, in other words, is not a forecast of how improvements in health will result in better life expectancy in the future. It is, instead, the exact opposite! It is a projection of current death rates for different age groups, applied to a group of people born today.
Their actual “lived and experienced” life expectancy may well be different, and usually is. See this in the case of France, for example:
The TMKK, of course, is that we should expect to live, on average, for longer than the number indicated by our period life expectancy.
That just leaves us with two questions:
What is cohort life expectancy then? Well, it isn’t a forecast or a projection, it is an actual record of how many people born in a particular year make it to the next year, and the year after that, and the year after that… until eventually none are left.
EconForEverybody 