Centenarian Clusters: Don’t Trust the Data

 Reaching the age of 100 is a significant achievement.  Making it to 105 labels one a semi-supercentenarian, hitting 110 a full supercentenarian.  Such accomplishments have been rare enough that if a town or region where an unusual number of people reached these ages was discovered, it was a newsworthy event.  It became a common practice to analyze the lifestyles and living conditions in the area in order to suggest reasons why people there were living longer than expected.  Articles would appear suggesting the particular diet of the local population was responsible.  Other suggestions included a physically active life, strong social support systems, and even genetic isolation.  A recent note provided by The Economist indicates research provides a simple explanation: such clusters might not even exist.

The article, Places claiming to be centenarian hotspots may just have bad data, focuses on the research of Dr. Saul Newman of Oxford University.  It was noted that in 2010 Japan had audited its records and discovered that 230,000 supposed centenarians in their data base were actually dead or missing.  If the data on aged citizens in Japan was that incorrect, could bad data explain the anomalies?  Many centenarians today could have been born at a time when birth certification did not exist.  Today, we are constantly required to produce our birthdate and forgetting our age is almost impossible.  But if there was no birth certificate to refer to and no constant reminder provided, would we remember our age accurately after 100 years?  The data from Japan suggests that in some cases dead people might be kept alive for fraudulent purposes such as a continued arrival of pension checks.

Newman studied data available from the United States, Britain, Italy, France, and Japan.  His first observation was that regions with exceptional numbers of centenarians tended to be in the poorer regions of a nation, where record keeping might make one suspicious.

“Okinawa, for example, has a poverty rate nearly twice the Japanese average and 1.6 times as many listed centenarians for each reported nonagenarian.”

There were examples of contradictory data sets.

“In Italy provinces where more people reach the age of 105 tend to have more people die before 55. On the island of Sardinia, renowned for its abundance of very old people, residents have among the lowest chances of reaching midlife of any Italians.”

Places that were late in issuing birth certificates have also tended to have an excess in centenarians.

“The most concrete evidence that mistakes could be causing variations in the numbers of very old people came from America. Between 1841 and 1919, states introduced birth certificates, making age estimates more accurate and fraud more difficult. By aligning data on the numbers of old people in each state with the date that birth registration was introduced, Dr Newman found that it resulted in a 69% drop in the prevalence of supercentenarians.”

As with all things dealing with humans and health, simple correlations can be terribly misleading—perhaps, this one as well.

You can learn a little about a lot of things or you can learn a lot about a very few things. Guess which is the most fun.