How Much Do We Know About Really Old People?
Why most of what we know about the 85+ population could be wrong.
Disclaimer: I have previously written about and personally replicated aspects of Dr. Newman’s work. However, I have not had time to fully replicate the study that I’m writing about here. I have downloaded the code, which appears to faithfully recreate all of the figures in the paper, but haven’t rewritten and rerun it to confirm that the findings are all accurate.
Scientists have been looking into the lives of the very old for a long time. The basic idea is simple - if we can find commonalities between people who get to the age of 90 and beyond, we can figure out how they got there. If we know what causes extreme longevity, then maybe we can understand how to make other people live much longer lives as well.
This is the basic concept behind Blue Zones and about half of all longevity science. Look at the very old, and they will show you how to add decades to your life.
Except, as a new paper shows, this entire field of research is potentially…untrustworthy. The data seems to show that the majority of extremely old people are probably much younger than they claim to be.
The Science
The study in question is a preprint - which means it has not yet been peer-reviewed - by Dr. Saul Newman. I’ve written about Dr. Newman’s work before, and generally have found it to be very thought-provoking and well-done. This particular study earned him an IgNobel prize for science that makes you laugh and then makes you think.
Dr. Newman’s observation is fairly simple. All age verification is done through forms. Globally, most of these forms are paper although in the last decade some have become digitized. Verifying someone’s age is usually done by looking at paper forms, and where those forms don’t exist there are various processes to look at other paper documents. If you don’t have a birth certificate - which is not uncommon for people over the age of 80 - there are a many methods that different countries use to try and verify your age.
The problem with all of these systems is that they have an error rate. Birth certificates sometimes have mistakes. Passports can be misprinted for many reasons. Sometimes, people lie about their age to get social benefits that they wouldn’t otherwise be able to access.
One of the most fascinating cases Dr. Newman cites as evidence of this issue is the Japanese centenarians. In 2010, Japan did a careful audit of every person aged over 100 by looking for the actual individual rather than relying on paper records to verify their existence. When they actually looked for these people, they found that 281 of them had been dead, many of them for some time. More than 80% of the total centenarian population in the country simply didn’t exist.
Dr. Newman’s paper looked at this issue globally. He constructed a centenarian attainment rate, which is basically the proportion of people aged 80 who survived until 100+, and compared countries across the globe. We would expect this rate to be very similar to life expectancy at birth, because life expectancy at every age is highly correlated. In countries where people have a longer life expectancy at birth, they also have a longer life expectancy at age 50, 60, and 70.
However, that’s not the case for centenarian attainment. Dr. Newman’s analysis shows that the countries which have the highest centenarian attainment - remember, people living from 80 to 100+ - are…unexpected. The top 10 countries regularly includes Guam, Guadeloupe, Kenya, Thailand, and the unincorporated territory of Western Sahara.
These are all places with poor life expectancies otherwise. Thailand came 54th in 2021 globally for life expectancy at birth; Kenya was substantially lower. Malawi fell out of the top 10 countries for centenarian achievement in 2011, but prior to this it was one of the best countries in the world for survival from 80-100+ for nearly 30 years. This is despite being in the bottom 10 or 20 countries in the world for total life expectancy for that entire period.
This makes no sense. There is no plausible reason why people in Malawi would die far younger than everywhere else at every age we measure except for the very oldest ones. We have no realistic biological explanation for people aged 70-74 and 75-79 dying at higher rates, but somehow people aged 80-100 surviving dramatically longer.
The most likely explanation? Data errors and fraud.
Living Past Age 100
There are possible explanations for these incredibly weird trends. Perhaps there’s some secret quality of the Malawian diet that kills people at young ages but creates superhuman qualities upon reaching age 80. Or, perhaps, the much simpler explanation is that the data is all terrible because it relies on fundamentally inadequate systems.
Death rates are a fascinating thing. They roughly double every 10 years, meaning a 65yo is more than twice as likely as a 55yo to die over a 12-month period. This means that even a relatively small group of people - say, a few hundred in a country of millions - lying about their age has long-lasting impacts. If we look at, say 100,000 people with a reported age of 65 of whom 10 have lied about their age by a decade or more. At age (approximately):
- 75, 2% of the liars have died and 6% of the total population
- 85, 6% of the liars have died and 17% of the total population
- 95, 17% of the liars have died, and 45% of the total population
- 105, 45% of the liars have died, and above 99.9% of the total population
By age 105, you’ll be left with somewhere between 5-10 truthful people and 5 liars still alive. Every year above 100 years of age, around half of the remaining population dies, but the death rate is more than 50% lower for the liars. It’s possible with even a very low rate of liars - this example is 1 in 10,000, stolen from Dr. Newman’s paper - you end up with a situation where liars make up most or even all people over the age of 100.
It doesn’t even take people lying. If the government makes an error in 0.1% of documents and people only correct the errors when they underestimate age, but don’t bother when they overestimate it, then you’ve instantly got issues with the very old population. People are much more likely to correct the government when it says they are a decade younger than they are because then they have to wait longer for pensions and other social support, but if the government overestimates your age by a decade…well. You’ll get a senior’s card, Medicare, or whatever the local benefits are and don’t even have to feel guilty about claiming it because it was the government that made the mistake.
And we know that this happens, regularly. There are countless media stories of people claiming their dead relatives’ pensions for years - even decades - after they had passed. There is also evidence of systematic errors in age records from sources such as the US census.
All of this is a huge problem for longevity researchers. Dr. Newman’s analysis is incredibly simple, but also rather remarkably powerful. He has shown that we probably cannot trust the data for records of very old people, which throws into doubt everything from the Mediterranean diet and intermittent fasting to Blue Zones and moderate alcohol intake. Vast swathes of the current popular health advice is built on the idea that we should emulate the world’s oldest people, which is a big problem if they’re lying about their age.
If you want to be very, very old it turns out that the best solution has nothing to do with food, or exercise, or any other regular behaviour. You just have to move to a place with bad record-keeping and lie to the government about your age, and you too can ‘live’ to be 110.
I very much enjoyed reading this for a variety of reasons. My Dad died when he was 101 years old, born in 1915. I think it is likely this is accurate because my grandfather was a Southern Baptist minister and the birth of my grandparents' children would have been recorded in the church records as well as by, I am guessing, the county.
As for myself, I have a copy of my original birth certificate and decades ago plasticized it for posterity, and my high school diploma from 1962 when I was just turning 18 definitely show me to currently by 80 years old. I have this feeling I will live to be 120. You can put me in the cohort to begin studying when I reach 90 or 100 with my permission. :-)