Nebbia Giacomo, Nussbaum Lisa, Helmkamp Annie, Andersen Stacy, Perls Thomas, Sebastiani Paola
Intelligent Systems Program, University of Pittsburgh, Pittsburgh, PA, USA.
Veteran Healthcare System, Boston, MA, USA.
N Am Actuar J. 2018;22(4):591-599. doi: 10.1080/10920277.2018.1462716. Epub 2018 Sep 27.
A large portion of the baby boomer population will live beyond the age of 90 years and entitlement programs and various insurance products have thusly become interested in longevity risk. Beyond cohort life table predictions, actuaries have little to go on in determining which individuals or portions of populations are at increased risk of living to 90 or 100 or even older. We and others have noted strong familial risk for living beyond the oldest one percentile of survival and we developed an algorithm that uses information about relatives' longevity to compute the chance of an individual surviving to extreme old age. An important step of this work is to compile large samples of pedigrees with and without long lived family members. Here, we describe our process of hand-curation of centenarian pedigrees and software that we have developed for the automated construction of such pedigrees using internet-based resources that can support the manual process.
婴儿潮一代中的很大一部分人将活到90岁以上，因此，福利项目和各种保险产品开始对长寿风险产生兴趣。除了队列生命表预测之外，精算师在确定哪些个体或人群部分活到90岁、100岁甚至更高年龄的风险增加方面几乎没有什么依据。我们和其他人已经注意到活到生存年龄最老百分之一以上存在很强的家族风险，并且我们开发了一种算法，该算法利用有关亲属长寿的信息来计算个体活到极高年龄的几率。这项工作的一个重要步骤是汇编有和没有长寿家庭成员的大型系谱样本。在这里，我们描述了我们手工整理百岁老人系谱的过程，以及我们开发的软件，该软件用于使用基于互联网的资源自动构建此类系谱，这些资源可以支持手工过程。
