Studies in the demography of supercentenarians in the United States

The proliferation of oldest-old persons in recent decades and emergence of “true” supercentenarians (persons aged 110 years and over) in recent years presents us with new policy debates and new questions regarding maximum lifespan and the relative importance of environment and genes. This dissertation first identifies and verifies the ages of the most longevous population in the United States, the “supercentenarians”, and then investigates some of the nongenetic survival attributes of this extremely longevous population, including the childhood environment and the familial component of longevity. Supercentenarian ages have been well documented in European countries with a history of birth registration, but have not been systematically studied in the United States, which lacks similar documentation, and where inaccuracy of age reporting has been an issue. To verify age we attempt to link Social Security Administration records for about 700 individuals who died in 1980 through 1999 (purportedly at ages 110+) to records of the U.S. censuses of 1880 and 1900, conducted when they were children. Using individual, household, and community-level variables gathered from the childhood census record, we consider whether childhood conditions can predict survival to age 110, and whether those variables retain their predictive importance beyond age 85. When comparing supercentenarians to their overall birth cohort, we find that residential and community characteristics, parents\u27 characteristics and subject\u27s characteristics, all independently influence survival to age 110. When comparing white supercentenarians to those in their birth cohort who survived to at least age 85, we find that living on a farm and/or with a farmer father gave a child significantly higher odds of survival beyond age 85 to age 110. Lastly, we examine the familial component of longevity by comparing the mortality above age 50 of brothers of supercentenarians (persons aged 110+) to their general male birth cohort. Using nonparametric (Kaplan Meier), semiparametric (Relative Risk), and parametric (Kannisto, Gompertz) methods of analysis, we find that, regardless of the method used, the mortality above age 50 of the brothers of supercentenarians (born 1868–1898) is much lower than that of the average male born in 1900, until at least age 95. Relative risk of survival from age 50 to age 100 for brothers versus the general cohort is 5–7:1, depending on the method used