Formal demography, innovative methodologies, and their applications

This paper proposes a new method to estimate disability transition rates from national cross-sectional health surveys. The proposed method estimates age-specific transition rates from cross-sectional data according to well-documented longitudinal age-specific health transition rates of other populations, used as standards, and the proportion of health and unhealthy individuals by age, reported in cross-sectional datasets. In order to estimate healthy life expectancy, this paper makes use of most recent Brazilian health survey data The preliminary results indicate that the estimated disability transition rates are consistent with the current literature. Moreover, the estimated parameters for the simple model specification seem to produce very reliable results. In 1998, 2003 and 2008 the estimated life expectancy – with and without any disability – do not show significant statistical differences from other estimates, produced by other methods.
A second exercise will be conducted by estimating the parameters including covariates: sex, race and education and to estimate differentials in healthy life expectancy.

This paper aims to study the evolution of adult mortality in small areas in Brazil from 1980 to 2010. We use 45q15 as a summary measure of adult mortality across small areas in Brazil. In order to produce the estimates, we propose a methodological approach that combines death distribution methods (DDM) to indirect standardization and Bayesian statistics, in order to produce more reliable estimates of adult mortality for small areas. We focus on adult mortality because there are much more studies on infant and child mortality for sub-national population using indirect demographic techniques (Souza, Hill and Dal Poz, 2010; Castro and Simões, 2009). The studies on child and infant mortality are also showing a convergence of levels across regions in Brazil, thus the main changes in life expectancy in recent years should be explained by changes and variation in adult mortality levels.

Traditional health statistics are derived from civil/vital registration. In low- to middle-income countries civil registration varies from partial coverage to nothing. We propose a new statistical framework for gathering health and population data - HYAK - that leverages the benefits of sampling and longitudinal, prospective surveillance to create a cheap, accurate, sustainable monitoring platform. HYAK has three fundamental components:

1. DATA MELDING: a sampling and surveillance component that organizes two data collection systems to work together: (1) data from health and demographic surveillance systems (HDSS) with frequent, intense, linked, prospective follow-up and (2) data from linked sample surveys conducted in large areas surrounding the HDSS sites using informed sampling so as to capture as many events as possible;

2. CAUSE OF DEATH: verbal autopsy to characterize the distribution of deaths by cause at the population level; and

3. SES: measurement of socioeconomic status in order to characterize poverty and wealth.

We conduct a simulation study of the informed sampling component of HYAK. Compared to traditionally cluster sampling, HYAK's informed sampling captures more deaths and produces estimates of both death counts and mortality rates that have lower variance and small bias.

Latin American has come a long a long way in reducing child mortality in during the last 60 years, decreasing, as a region, from a Q(5) [probability of dying before age 5] of 191 per thousand in 1950 to 31 per thousand nowadays, according to the latest estimations by CELADE-ECLAC. However, there is evidence of unequal decrease among countries.
As well as differences observed in the decrease rhythm among countries, there are significant differences within countries by regions or by socioeconomic conditions. In this paper, the researchers use census microdata from two Latin American countries (Mexico and Brazil), from the 1980, 1990 and 2000 census rounds, with the aim of estimating a model describing the gap trend in Q(5) by years of study of the mother. Using the estimated model, projections are made for gaps up to the latest census (2010 round) and achievements are evaluated, in terms of gap reduction, establishing whether the rhythm followed the expected trend, whether the rhythm was faster or whether a stalling is observed. Inferences can be made on the gaps evolution comparing population groups, as well as the validity of the variable used to measure more or less vulnerable groups regarding child mortality.

Demographic censuses usually contain information about a graded education system, i.e. age and grade declarations. This information can be used to estimate a series of indicators, useful for diagnostics and prognostics of the educational system. One of the principal goals of this paper is to provide a new demographic technique to better understand the population trends in terms of levels of schooling, in a country or region. A detailed scrutiny of the derived indicators proposed can be a powerful tool for policy makers. Our technique follows in the tradition of formal demographic methodologies used in analyzing and projecting population, such as Life Tables. Thus, one could study the probable social consequences of the implementation of any educational policies related to promotion and retention practices, over the medium and long runs. The methodological procedures were applied to the Latin American Countries, although it could be replicated or adapted to other developing countries or regions that have the usual information of grades concluded by age.
Key Words: Demographic analysis, trends and forecast of educational indicators, regional inequalities

China’s excess of males, as a consequence of skewed sex ratio at birth for more than 30 years, has attracted much attention on the scale of marriage squeeze and the socioeconomic impact of the forced unmarried males, but studies on the death of unmarried males are underdeveloped due to the limitation of data. In this paper, we firstly develop life tables of the unmarried males and nuptiality tables of males in China. In addition, the mortality and life expectancy of the unmarried males and married males, unmarried males across education levels and health difference for the elderly by marriage status are compared. Secondly, we simulate a projection model to evaluate the influence of the relatively high mortality of unmarried males on the prediction of excess males in 21th century of China. We find that an overestimation of at least 16% to the number of unmarried males aged more than 50 could be generated if the excess mortality of the unmarried to the mortality of overall males not accounted. Finally, we pay special attention to the current social security system on the unmarried male elderly and the paper’s findings suggest a need to relax the age restriction and expand the coverage of social security programs to the unmarried males, especially in rural areas.

Intercensal methods have been broadly used to estimate mortality in developed and less developed countries with deficient or incomplete data. These methods have several advantages over indirect methods because they do not require the use of model life tables and provide sufficiently accurate results even in the presence of age distortions and death under-registration. The drawback of these methods, however, is that generated life tables do not provide ex post projections of the baseline population that are consistent with the subsequent enumeration, even after adjusting for migration and age misreporting. This article demonstrates these inconsistencies by reviewing and comparing the results of three well established methods. We introduce a simple procedure to solve this inconsistency by providing life tables that are accurate and which generate identical projected and input populations. The empirical illustration demonstrating its efficacy draws on data from Vietnam, but the method can be extended to any context and time period.

This paper contributes to the analysis of the relation between child malnutrition and morbidity by providing a comprehensive assessment of the mutual impacts of the two phenomena. We investigate the synergistic relationship between malnutrition and morbidity among infants in Nairobi's poor urban settlements using data from 3,459 children enrolled in an ongoing Maternal and Child Health Longitudinal study. For our analysis we use a joint dynamic framework to account for persistence in a given nutritional or morbidity status, and also for dynamic cross-effects between the two phenomena. We address the econometric challenges associated to our analytical framework using a bivariate random effects dynamic probit model, a bivariate extension of Wooldridge approach. This paper contributes to the literature in two ways. First, we propose an analytical framework that examines both dynamics of malnutrition and morbidity and analyze the contamination process between them, which permits to investigate the interactions existing across the two problems. Second, we contribute to shed light on a very important empirical question: does malnutrition interact identically with all common forms of morbidity, or is its effects stronger for some types of morbidity than others? Answer to this question may have important policy implications.

China’s life expectancy at birth is in debate, and the 2010 census data may exaggerate the figure and the increase pace in it. In this paper, with an extension of Lee-Carter method for limited data, we use China’s 1982, 1990 and 2000 census to forecast mortality pattern and life expectancy for 2000 to 2030 period. We find that the annual gain in life expectancy from 2000 to 2030 is 0.18 years for males, and 0.23 years for females, and the infant mortality rate will decline to 10.39‰ in 2030 for males, and to 20.32‰ for females.

Traditional demographic methods correct adult mortality by assuming a uniform distribution of the errors in between censuses. We propose the use of quality control charts to detect which years should be corrected at a larger scale than others within inter-census periods. This methodology proves new in the area and highly improves the current available demographic methods, both for all age groups, including the oldest ages usually hard to correct with traditional methods. We used Colombian official death records from 1950 to 2010, as recorded from the Latin American Human Mortality Database (www.lamortalidad.org), and found that this methodology will enhance current available demographic methods.