Methods for projecting all-cause or cause-specific mortality

The deaths by Road Traffic Injuries (RTI) are a serious health public problem in many regions of the world, mainly in developing countries; nevertheless, is a topic no enough studying by the Demography. Moreover, many approach not consider the phenomenon`s heterogeneity and uncertainty, so this paper studies the cause of the death trough a quite approach very used in the economic field: the multiple time series model. This methodology is not quite used in demographic analysis but it can be really useful. This paper pretends to analyze and forecast the deaths by RTI in Mexico City in a five-year term (2011-2015), recognizing the RTI´s heterogeneous in the geographic area, i.e., the analysis will make for each concentric ring are made up the city. We use the Vector Autoregressive models in the forecasting of the deaths by RTI. The expected findings are the RTI is a quite heterogeneous phenomenon thorough the concentric rings; the forecast will be very close to real future observations. We want to motivate to demographic specialists to use stochastic tools like this, because they have a lot of potential in the study of many demographical variables.

Since 1999 IBGE – the Brazilian Institute of Geography and Statistics – is annually publishing mortality tables for the Brazilian population, according to gender and age, the first table in the series being published in 1980. The latest 2011 mortality table is due at the end of 2012 and will crucially include data from the latest 2010 Brazilian Census. Census data in Brazil is collected once in a decade and therefore the latest data are important to update any past study that involved extrapolation in the future.

In this work we use past data and the latest census data to run an analysis of Brazilian life-expectancy using the Lee-Carter model. Such model allows to forecast a trend in life expectancy as well as to estimate probability intervals for deviations from the trend, using a stochastic factor. The data we obtain can be used for scenario analysis and policy making.

We also develop a free computer based tool to run the Lee-Carter model analysis from a given set of mortality table inputs.

We introduce bi-variate and multivariate force of mortality functions. The pattern of mortality in a population is one of the strong influencing factors in determining the life expectancies at various ages in the population. Considering univariate functions of age only to understand the human mortality data without associating with other variables could lead to incomplete analysis. The reasons behind declining forces of mortality globally could be studied using the proposed functions.

When mortality forecasts for subpopulations are treated as independent, historic relationships among subpopulations may not be reflected in the forecasts. The product-ratio method of coherent forecasting (Hyndman, Booth and Yasmeen, Demography, online 2012) takes account of the relationships among subpopulations based on a single criterion such as sex or state. The coherent forecasts have been shown to improve overall accuracy and to equalise accuracy across subpopulations. The aim of this paper is to compare the accuracy of mortality forecasts when the subpopulations are based on sex and state (or country). The product-ratio method is applied to male and female populations of a group of n states, producing mortality forecasts for 2n subpopulations defined by sex and state. Two forecasts per subpopulation are made: the first set are sex-coherent forecasts for each country, while the second set are country-coherent forecasts for each sex. The accuracy of the two sets of forecasts are compared. Examples include four Nordic countries (Norway, Sweden, Denmark and Finland), three constituents of the UK (England & Wales, Scotland, Northern Ireland) and four states of Australia.

A single summery index of mortality can never replace the set of age-specific death rates, it has been found to be extremely useful for a wide variety of purposes. The choice of index depends upon the purposes for which it is to be used, and is important as different indexes can produce very different results. This study examines the variation in mortality and mortality trends in India and its major states from the 1970-75 to 2002-2006 using data from the Sample Registration System (SRS). Several indicators were developed to study mortality .The most basic indicators are: CDR, IMR, ASDR, and LEB and several other indicators were also existed. Arithmetic mean of the Age Specific Death Rates and Geometric Mean of the Age Specific Death Rates (or the Del Index) are examples. Input data used in deriving the DI and OMI Age Specific Death Rates (ASDRs) of the 16 age groups: 0-1, 1-4, 5-9, 10-14, -----65-69, and 70+ “Age Structure of Mortality in India and Its Bigger States“ Del Index refers to the geometric mean of the age-specific death rates as a summery index of mortality, which was postulated by the versatile Demographer Robert Schoen in the year of 1970.

The purposes of this study are to examine a method to overcome the shortage of historical data on mortality of the elderly and to find the best model to forecast Korean mortality rates overall. To extend the mortality for ages 75 and over, we test two methods of estimating death probabilities: the 2-parameter logistic model and the Brass-Logit model. Based on the Mean Absolute Percent Error (MAPE), the logistic model has better performance than the Brass-Logit model. Four stochastic forecasting models (the Lee-Carter Model, the adjusted Lee-Carter Model, the Lee-Miller Model, and the Coherent Lee-Carter Model) are fitted to the period 1970-2010. The forecasts are compared to actual mortality for that period. The results of this evaluation show that the Coherent Lee-Carter Model is consistently more accurate in forecasting Korean mortality rates than other compared models. The Coherent Lee-Carter model yields a higher life expectancy at birth for both sexes and a larger difference between sexes than other models in which sex differentials diminish rapidly.

Though the cardiovascular diseases are the leading cause of deaths in India, there are limited studies that provide estimates on burden of cardiovascular diseases. The aim of this paper is to project the age and sex pattern of deaths and hospitalisation due to cardiovascular diseases in India and its major regions. The data have been drawn from multiple sources; Special Survey on Cause of Death, 2001-03, Sample Registration System 2004-2010, Expert Committee Projection on Population 2001 and Census of India 2011. The projections are carried out, for 2004-21, under baseline, optimistic and pessimistic scenario. A sensitivity analysis has been carried out on the estimates of cardiovascular deaths. Under the baseline scenario the number of deaths due to cardiovascular diseases was estimated at 1.4 million in 2004, 1.6 million in 2010, 1.8 million in 2016 and 2.1 million in 2021. The total hospitalisation was estimated as 6.7 million in 2004, 7.7 million in 2010, 9.5 million in 2016 and 10.9 million by 2021. While, deaths were relatively higher among older adults, hospitalisation was concentrated among prime working age group. The regional estimates suggest that the majority of the deaths and hospitalisation were in the southern and eastern region of India.

Mortality analysis seen from the cohort perspective is in demography often associated with purer and more adequate way of study of mortality changes and mortality development. Cohort mortality is more stable in time and reflects the mortality conditions during the whole life of the generation. Unfortunately, traditionally the complete cohort life tables could be calculated only for cohorts which are already extinct. In this paper a simple method of mortality estimation is applied for cohorts which are not yet extinct. This method was developed in previous research steps of the author. Through application of this estimation method the intensity of mortality of adults will be calculated and then analyzed in connection to the concept of rectangularization of the survival curve or compression of mortality. According to the assumptions, the verification of this concept could be taken as a proof of the existence of a limit of the human life span. Analysis of this process is applied to the empirical as well as modeled cohort data from several selected European countries. The differences between post-communist and other European countries will be traced and described with the aim of finding some more general patterns and regularities usable for example in population forecasting.

United Nations Millennium Development Goals(MDGs) has set the target of reducing high rates of infant (and child) mortality (IMR) by two thirds to be reached by 2015 using 1990 as the benchmark year. The availability of time series data on Infant Mortality rate from Inter- Agency Group for Child Mortality Estimation(IGME) led by UNICEF and WHO, it has become possible to track the rate of progress towards this goal. Using the IGME 2012 data for all the South Asian Countries, I have considered three specific issues in this article. (1). How does the South Asian Countries fair in reducing the IMR towards this MDG target? Although the time series data exhibit declining trends for all the countries in South Asia, to what extent such trends are attributed by their average annual progress trajectory over the period for which data are available? (2). Whether deterministic or stochastic trend can attribute the IMR decline in South Asian countries and what alternative time series models be used to forecast the decline in Infant Mortality? Can we find a serviceable representative model for the entire region? (3) In case, a satisfactory representative model for the entire region exists, how do we assess the forecast accuracy using the model and quantify the propagation of forecast error?

This paper presents life expectancy forecasts for 159 countries explicitly assuming mortality convergence. We develop a model that takes into account country-specific heterogeneity in life expectancy historical trajectories, between-countries heterogeneity across gains and uncertainty through experts’ based arguments (Lutz et al., 2001). The relevant literature has focused on forecasting mortality for a single population. Exception to this rule is the work by Li and Lee (2005) where the authors develop mortality forecasts that take into account patterns in a larger group using the Lee-Carter model.
Torri and Vaupel (2012) argue that life expectancy in different countries tends to be positively correlated and forecast life expectancies in individual countries by forecasting the best-practice level and the gap between the national performance and the best-practice level. We build upon their work by varying the speed of convergence, taking into account differential rates of linear increase in life expectancy across group of countries.