Death certificates and lifespans

Something that rather negates a lot of research into the inequality of lifespans out therethere:

Let’s take Marmot’s contention that life expectancy in the poorest ward of Kensington and Chelsea is 22 years lower than in the richest. As with the Marmot Review into inequality, this fundamentally misunderstands the way life expectancy is measured.

Imagine a world in which Boris Johnson had succumbed to coronavirus – would that have registered as a change in life expectancy in London or New York? It would be recorded as one in London, the place of residence at time of death, not the place of birth. For the numbers used to calculate lifespans across geography are years lived allied with place of death – not the place of birth.

At the national level this is fair enough, even if not totally accurate. Some 15% of the current countrywide population is foreign born, so measuring lifespan by age and place of death is only ever going to be some 85% accurate. When the attempt is made to narrow this down to council ward level it’s not going to be accurate in the slightest. The assumption being made is that some useful portion die in that same little geographic area containing 5,000 or so people that we were born into. This is simply not supportable.

We don;t know the lifespan of the people born into an area. We only know the lifespan of someone who dies in an area.

The implication of this should be obvious. We have no information at all on the effects of childhood upon lifespan when allied with geography…..

11 thoughts on “Death certificates and lifespans”

  1. It’s a good observation, thanks Tim. Furthermore, there is probably a negative correlation in that people most likely to die younger for other reasons are more likely to end up in poorer areas, and conversely those who were born in poorer areas but act in a way to increase their life expectancy likely die in richer areas. Also, we’re looking at data that is severely delayed – even with the best modelling approach, we can’t really know the impact of today’s individual policies on life expectancy due to place of birth any more than we can know whether today’s UK “lockdown” will cause more suffering over the long term than an alternative approach would have (though we may have strong suspicions on this one!).

  2. The ONS *could* analyse deaths by place of birth – to the extent of place of birth specified on death certificate – but it doesn’t because the effort involved isn’t worth it. This still would not differentiate between different wards in Kensington and Chelsea or any other town.
    Infant mortality could be analysed but little information could be gleaned if the number of dead infants was 0 in most areas.

  3. The NHS has lifetime data going back to 1947 (not long enough, OK) based on GP registrations, hospital admissions, prescriptions, etc. Not easily rounded up in a nice spreadsheet but doable. It’s probably the most comprehensive and unbiased data set in the world thanks to universal monopoly health care. So it’s immensely valuable. And the muppets gave it away to google for free!

  4. Perhaps they could also tell us the death rate or infection rate for various classes of immigrant and whether they brought it (e.g. TB) with them. Perhaps it would be interesting to know the IQ and other parameters for the offspring of cousin marriages.

  5. How does the life expectancy of immigrantscompare to those of their fellows who remained in their place of birth?

  6. Philip: the NHS data actually goes back to before 1939 as the initial NHS records were created from the 1939 sort-of-Census. In many places, the actual physical paper survey *was* the local NHS records. The ones for my town have annotations and additions going into the 1950s.

  7. I enjoyed the whinging Bristol Professor ( I think it was him ) that CV-19 took out twice as many people in the most deprived areas compared to the least deprived. He went on to say the link between deprivation and early mortality causes is one that is long established with typical differences of 2.5 to 1 between the most and least deprived.
    Erm, so is CV-19 a bit of an equaliser then?

  8. “The NHS has lifetime data going back to 1947”
    Maybe, sometimes, but I doubt it is complete, nor capable of being matched up. I have an NHS card that dates back to circa 1957, of course that NHS number is not recognised by todays systems. In theory I was registered with the same GP and his successors to that practice from age 3 until a few months ago. But as I last saw a GP over thirty years ago I doubt my records still exist.

  9. Tim, I sent your piece to an Oxbridge academic working in this area. He agreed with you, adding: “I would hesitate before publishing such arguments myself, eg in the Conversation. The inequality argument is now orthodoxy, and it would be a very secure academic who would challenge it in these terms. Even when couched as carefully as the piece you sent, it would attract a deal of obloquy.”

    Time to cleanse the Augean Stables of academe of leftists…

Leave a Reply

Your email address will not be published. Required fields are marked *