Background: In many low- and middle-income countries, the mortality of adults over 50 years of age is poorly monitored because death registration systems are deficient. Nationally representative surveys currently focus on the survival of children or adults aged 15 to 49 years. Objective: We propose to measure adult survival beyond age 50 via parental survival histories, in which survey respondents provide data on their parents’ ages at the time of the survey, and if deceased, their age at death and date of death. We evaluate the magnitude of possible selection bias in parental survival histories and quantify the sample sizes needed to estimate mortality at ages 50 to 79 with varying levels of precision. Methods: We created a population with known parental survival histories using the 2013 national census of Senegal augmented with microsimulations. Using a stratified two-stage sampling procedure, we then conducted household surveys of this artificial population. We compared reference mortality levels in the artificial population of adults aged 50 to 79 years with those inferred from parental survival histories. We also analyzed selection biases in simulated populations where mortality above age 50 is correlated with the number of adult children. Results: The inclusion of modules on parental survival in large-scale surveys, such as the Demographic and Health Surveys, could provide accurate and precise estimates of old-age mortality and capture time trends and age patterns. Estimates derived from parental survival histories are affected by an upward bias when mortality is positively correlated with fertility, and vice versa, but the bias is modest and can be partially corrected. Contribution: Parental survival histories are a promising method to fill important data gaps around mortality at older ages, although more research is needed on possible reporting errors.

中文翻译：

---

在调查中利用父母生存史改进老年死亡率估计（作者：Bruno Masquelier、Ashira Menashe-Oren、Benjamin-Samuel Schlüter、Atoumane Fall、Stephane Helleringer）


背景：在许多低收入和中等收入国家，由于死亡登记系统不足，50 岁以上成年人的死亡率监测不力。具有全国代表性的调查目前侧重于 15 至 49 岁儿童或成人的生存情况。目的：我们建议通过父母生存史来衡量 50 岁以上的成人生存率，其中调查对象提供有关其父母在调查时的年龄数据，如果已去世，则提供他们的死亡年龄和死亡日期的数据。我们评估了父母生存史中可能的选择偏倚的大小，并以不同的精度水平量化了估计 50 至 79 岁死亡率所需的样本量。方法：我们使用 2013 年塞内加尔全国人口普查创建了一个具有已知父母生存历史的人群，并进行了微观模拟的增强。然后，我们使用分层的两阶段抽样程序对这个人工人群进行了家庭调查。我们将 50 至 79 岁成年人的人工人群的参考死亡率水平与从父母生存史推断的参考死亡率水平进行了比较。我们还分析了模拟人群中的选择偏倚，其中 50 岁以上的死亡率与成年儿童的数量相关。结果：在大规模调查（例如人口和健康调查）中纳入有关父母生存的模块，可以提供准确和精确的老年死亡率估计，并捕捉时间趋势和年龄模式。当死亡率与生育率呈正相关时，从父母生存史得出的估计值会受到向上偏差的影响，反之亦然，但偏差不大，可以部分纠正。 贡献： 父母生存史是填补老年人死亡率重要数据空白的一种很有前途的方法，尽管需要对可能的报告错误进行更多研究。