An assessment of methods for estimating adult mortality from two sets of data on maternal orphanhood

Survey and census data about the survival of respondents’ mothers have been used widely for the estimation of adult mortality. Four methods are described that combine two sets of orphanhood data and yield estimates for the intersurvey period. They are applied to enquiries conducted in Peru, Kenya, and Malawi. This provides improved estimates of recent mortality and also clarifies the nature of the errors that affect the basic data. Age misreporting and other errors affect the information about older respondents and orphanhood of children is sometimes underreported. In contrast, data supplied by young adults seem plausible.     

Estimation of mortality from orphanhood in adulthood

This paper extends earlier research into methods for estimating adult mortality from information on the recent incidence of orphanhood. It presents a series of regression coefficients for estimating female and male mortality from synthetic cohort data on the subsequent orphanhood of those who had a living mother or father at exact age 20. Such information can be obtained either where questions about parental survival have been asked in two inquiries or by asking retrospectively about dates of orphanhood in a single survey. Although the method is somewhat sensitive to errors in the reporting of ages and dates, it is a promising source of up-to-date estimates of adult mortality that are free from bias due to the underreporting of the orphanhood of young children ("the adoption effect").     

The effects of high HIV prevalence on orphanhood and living arrangements of children in Malawi, Tanzania, and South Africa

Using longitudinal data from three demographic surveillance systems (DSS) and a retrospective cohort study, we estimate levels and trends in the prevalence and incidence of orphanhood in South Africa, Tanzania, and Malawi in the period 1988-2004. The prevalence of maternal, paternal, and double orphans rose in all three populations. In South Africa - where the HIV epidemic started later, has been very severe, and has not yet stabilized - the incidence of orphanhood among children is double that of the other populations. The living arrangements of children vary considerably between the populations, particularly in relation to fathers. Patterns of marriage, migration, and adult mortality influence the living and care arrangements of orphans and non-orphans. DSS data provide new insights into the impact of adult mortality on children, challenging several widely held assumptions. For example, we find no evidence that the prevalence of child-headed households is significant or has increased in the three study areas.     

Estimating adult mortality levels from information on widowhood

Summary A range of indirect techniques has been developed for mortality estimation in societies lacking adequate vital registration records. Information on orphanhood has been widely used as an estimator of adult mortality, with generally plausible results. Doubts have remained, however, about potential biases, and the method is less satisfactory for the estimation of male mortality. Information on widowhood, or more strictly the survival of first spouse, has several possible advantages over information on orphanhood. Model first marriage functions and model life tables are used to calculate proportions widowed of first spouse, for both females and males, by marital duration and by age. These proportions widowed are then related to life table survivorship probabilities to provide weighting factors for the conversion of observed proportions widowed into estimates of survivorship probabilities. The application of the method is illustrated with data collected by the 1974 post-enumeration survey of Bangladesh, with apparently encouraging results.     

Death by survey: Estimating adult mortality without selection bias from sibling survival data

The widely used methods for estimating adult mortality rates from sample survey responses about the survival of siblings, parents, spouses, and others depend crucially on an assumption that, as we demonstrate, does not hold in real data. We show that when this assumption is violated so that the mortality rate varies with sibship size, mortality estimates can be massively biased. By using insights from work on the statistical analysis of selection bias, survey weighting, and extrapolation problems, we propose a new and relatively simple method of recovering the mortality rate with both greatly reduced potential for bias and increased clarity about the source of necessary assumptions.     

Estimating African-American mortality from inaccurate data

This paper evaluates the quality of vital statistics and census data for estimating African-American mortality over a period of six decades. The authors employ intercensal cohort comparisons and extinct generation estimates to demonstrate that conventionally constructed African-American death rates may be seriously flawed as early as age 50. Using the crude death rate at ages 50+ for 1978-1982 in conjunction with estimated growth rates and two model life table systems, the authors estimate black age-specific death rates in 1978–1982. These results suggest that if a racial crossover in death rates occurs, the age pattern of mortality among African-Americans must be far outside the range observed in populations with more accurate data.     

Inferences about abortion from foetal mortality data

Abstract In a longitudinal fertility study in Detroit the evidence indicates that the foetal mortality rate in the prospective periods are better reported than in retrospective parts of the fertility histories. While the data do not specifically differentiate between induced abortions and other foetal deaths, the foetal death rates vary in relation to other social and demographic characteristics of the couples in such a way as make induced abortion a consistently plausible explanation of differentials. Foetal mortality rates tend to be high among sub-groups which have the incentive, information and the financial means for induced abortions. If these inferences are correct, it is likely that there is a considerable practice of induced abortion.     

Fertility estimation from open birth-interval data*

Censuses and surveys frequently collect information on period fertility through questions on the timing of last births. The standard approach to estimating fertility with open-interval data uses the proportion of women giving birth in the year before the interview. I propose a more efficient, maximum likelihood method for estimating fertility from open-interval data. I illustrate a mathematical derivation of the new method, perform sensitivity analyses, and conduct empirical tests with Brazilian census data. The new estimators have small biases and lower variance than standard estimators for open-interval data. Consequently, the new method is more likely to generate accurate results from small or moderately sized samples.     

Adult mortality in sub-Saharan Africa: Evidence from demographic and health surveys

This article reports levels, trends, and age patterns of adult mortality in 23 sub-Saharan Africa countries, based on the sibling histories and orphanhood data collected by the countries' Demographic and Health Surveys. Adult mortality has risen sharply since HIV became prevalent, but the size and speed of the mortality increase varies greatly among countries. Excess mortality is concentrated among women aged 25-39 and among men aged 30-44. These data suggest that the increase in the number of men who die each year has exceeded somewhat the increase for women. It is time for a systematic attempt to reconcile the demographic and epidemiological evidence concerning AIDS in Africa.     

Mixed estimation of old-age mortality

"Section 2 will first extend the method of mixed estimation to maximum likelihood estimation in general. Then, we will review generalized linear models with logistic and Poisson regressions as examples. In Section 3 we discuss different approaches for formulating the auxiliary information in practice. Section 4 first reviews the method of Coale and Kisker, provides empirical estimates for it, and then proceeds with the mixed estimation variant. In Section 5 we apply the methods to the estimation of mortality at ages 80+ in Finland in 1980-1993. We will first consider the evidence for mortality crossover between males and females....Then we will estimate life expectancies at age 100." (EXCERPT)     

Further developments in indirect mortality estimation

Summary A variety of indirect estimators of mortality; survival of children by marriage duration of mother, survival of first spouse by marriage duration and by age, maternal orphanhood, and survival of siblings, are investigated by the use of a wide range of model fertility and mortality situations. Survival probabilities are then related by regression analysis to the proportions with a particular characteristic, to yield an equation which can then be used to estimate the survival probability in a population. Maternal orphanhood and survival of first spouse by age have already shown themselves to be useful, and the new developments are only simplifications of the existing methodology. Survival of first spouse by duration of marriage, and survival of siblings are, however, new methods which have yet to be justified by field experience. In conclusion, the features common to all indirect mortality estimation procedures are outlined, and the direction future developments may take in response to gradually improving data quality is suggested.     

Effect of aggregation on the estimation of trend in mortality

There are three approaches to analyzing and forecasting age‐specific mortality: (1) analyze age‐specific data directly, (2) analyze each cause‐specific mortality series separately and add the results, (3) analyze cause‐specific mortality series jointly and add the results. We show that if linear models are used for cause‐specific mortality, then the three approaches often give close results even when cause‐specific series are correlated. This result holds for cross‐correlations arising from random misclassification of deaths by cause, and also for certain patterns of systematic misclassification. It need not hold, if one or more causes serve as “leading indicators”; for the remaining causes, or if outside information is incorporated into forecasting either through expert judgment or formal statistical modeling. Under highly nonlinear models or in the presence of modeling error the result may also fail. The results are illustrated with U.S. age‐specific mortality data from 1968–1985. In some cases the aggregate forecasts appear to be the more credible ones.     

Effect of aggregation on the estimation of trend in mortality

"There are three approaches to analyzing and forecasting age-specific mortality: (1) analyze age-specific data directly, (2) analyze each cause-specific mortality series separately and add the results, (3) analyze cause-specific mortality series jointly and add the results. We show that if linear models are used for cause-specific mortality, then the three approaches often give close results even when cause-specific series are correlated. This result holds for cross-correlations arising from random misclassification of deaths by cause, and also for certain patterns of systematic misclassification....The results are illustrated with U.S. age-specific mortality: (1) analyse age-specific mortality data from 1968-1985. In some cases the aggregate forecasts appear to be the more credible ones." This is a revised version of a paper originally presented at the 1990 Annual Meeting of the Population Association of America (see Population Index, Vol. 56, No. 3, Fall 1990, p. 407).     

Estimating infant mortality trends from child survivorship data

Summary Brass's procedure for estimating mortality from census or survey data on numbers of children born to women by age group and numbers of children surviving is generalized to allow the estimation of mortality trends. The new procedure is applied to data for Costa Rica and peninsular Malaysia. The resulting infant mortality rate estimates are compared with rates calculated from vital registration figures. The comparisons suggest, surprisingly, that the estimates derived from statistics for women aged 30-50 are not noticeably inferior to those derived from those for women aged 20-30. This suggests that the common practice of disregarding statistics for women aged over 30 or 35 years may be a mistake. Figures are presented which suggest that estimates based on women aged less than 20 are likely to be very seriously biased because of differential infant mortality by age of mother at birth.     

Orphanhood and adult mortality in the past: a critique of Latin American data and procedures

The author examines the applicability of indirect methods for estimating mortality from information about orphanhood based on methods developed by Brass and Hill using data from Latin America. The author concludes that the data sources available in the region may not be adequate for the application of such techniques.     

Maternal mortality in South Africa: an update from the 2007 Community Survey

The 2007 Community Survey conducted in South Africa included questions on maternal deaths in the previous 12 months (pregnancy-related deaths). The Maternal Mortality Ratio (MMR) was estimated at 702 per 100,000 live births, some 30% more than at the 2001 census. This high level occurred despite a low proportion of maternal deaths (4.3%) among deaths of women aged 15–49 years, which is even lower than the proportion of time spent in the maternal risk period (7.6%). The high level of MMR was due to the astonishingly high level of adult mortality, which increased by 46% since 2001. The main reasons for these excessive levels were HIV/AIDS and external causes of death (accidents and violence). Differentials in MMR were very marked, and similar to those found in 2001 with respect to urban residence, race, province, education, income, and wealth. Provincial levels of MMR correlated primarily with HIV/AIDS prevalence. Maternal mortality defined as ‘pregnancy-related death’ appears no longer as a proper indicator of ‘safe motherhood’ in this situation.     

Prime-age adult mortality in sub-Saharan Africa: examining trends in Demographic and Health Surveys conducted between 1990 and 2014

Adult mortality rate is a critical indicator used to assess the level of national development in most sub-Saharan African countries. However, estimation of adult mortality rates requires comprehensive and accurate reporting of adult deaths, which is one of the challenges faced by most sub-Saharan African countries. Using data from Demographic and Health Surveys conducted between 1990 and 2014 in 25 countries in sub-Saharan Africa, we examine trends in all-cause prime adult mortality (measured by the probability of dying between exact ages 15 and 50) and sex differences in adult mortality by region. Our paper provides a basis for tracking progress in reducing adult mortality and improving overall health. The median probability of dying was 173 per 1000 for women in the latest surveys, an increase from 166 per 1000 during the initial surveys. The median value for men was 177 per 1000; a decrease from the initial surveys which was 202 per 1000. Across all countries, the average annual increase in the probability of dying was higher for women (1.08%) than men (0.49%). Intensive efforts are needed to improve adult survival and ensure that sub-Saharan Africa achieves the Sustainable Development Goals by 2030. In particular, efforts to mitigate the premature risk of dying among women need to be intensified.     

An analysis of indirect mortality estimation

Abstract In this paper the robustness of Brass's child-survivorship indirect mortality estimation technique is investigated. An analytical method is developed for studying the error or bias caused in indirect mortality estimates by poor data, badly chosen model functions, and specific demographic assumptions that are often violated in practice. The resulting analytical expressions give insight into the rationale of indirect methods, the conditions under which they are robust, and the magnitude of errors that occur when specific assumptions are violated.     

An analysis of indirect mortality estimation

In this paper the robustness of Brass's child-survivorship indirect mortality estimation technique is investigated. An analytical method is developed for studying the error or bias caused in indirect mortality estimates by poor data, badly chosen model functions, and specific demographic assumptions that are often violated in practice. The resulting analytical expressions give insight into the rationale of indirect methods, the conditions under which they are robust, and the magnitude of errors that occur when specific assumptions are violated.