Modeling and forecasting U.S. sex differentials in mortality

Trends of Healthy Life Expectancy of the Elderly in China in 1994-2015: Revisiting From the Perspective of Morbidity Transition

Research on healthy life expectancy (HLE) in China has been fueled by a spate of new data sources and studies, yet no consensus is reached on the pattern of HLE changes and the underlying mechanism. This study examined the change of HLE in China over 20 years with long term national data. Health status, measured by activities of daily living, is combined with mortality to calculate the disability-free life expectancy by the Sullivan method. The results show that the HLE rose slower than life expectancy (LE) in 1994-2004, indicating morbidity expansion. However, in 2010-2015, the proportion of HLE to LE increased, manifesting morbidity compression. A counterfactual analysis further shows that health improvement has been increasingly important in increasing HLE in 2010-2015, despite the dominance of mortality decline. The findings suggest that morbidity can transition between compression, expansion and dynamic equilibrium over a long period due to different combinations of mortality and health improvements. Given the limited data in this study, whether and how morbidity transitions unfold in the future remains open and requires further research.

Trends and Projections in Breast Cancer Mortality among four Asian countries (1990-2017): Evidence from five Stochastic Mortality Models

The current study aimed to explore some important insights into the breast cancer mortality (BCM) trends and projections among four Asian countries by using five advanced stochastic mortality models. BCM data over 28 years from 1990-2017 with ages 20-84 were retrieved from the Global Burden of Disease (GBD) Study 2017 for four Asian countries, namely, China, India, Pakistan, and Thailand. Five stochastic mortality models with the family of generalized age-period-cohort were implemented to find the present and future BCM trends in these four Asian countries. Based on Cairns-Blake-Dowd (CBD) model and Lee-Carter model (LCM), overall, results revealed that BCM increased with the passage of time. Aging factor was the most influential factor of elevated BCM in each Asian country under consideration. Projection of BCM showed that mortality rates might continue to grow with time, especially in older ages in each Asian country under study. The highest forecasted BCM rates were observed in Pakistan as compared to other countries. The obvious increase in BCM suggested that earlier tactics should be implemented to reduce the subsequent morbidity and mortality due to breast cancer. The last but not least, some additional tactics to mitigate the BCM in older ages must be adopted.

Mortality trends of liver diseases from 1981 to 2016 and the projection to 2035 in Taiwan: An age-period-cohort analysis

BACKGROUND

Projections on liver diseases mortality can provide a basis for evaluating the World Health Organization's "25 by 25" goal and can help healthcare systems design appropriate control strategies. This study evaluated whether the 25 by 25 goal can be achieved in Taiwan and discussed possible future control strategies.

METHODS

Age-period-cohort models are used to estimate the mortality trends of liver diseases from 1981 to 2016 and project these trends to 2035.

RESULTS

For chronic liver disease and cirrhosis among men, the age-adjusted mortality rate decreased from 1981 to 1991, increased until 1999, and subsequently decreased again until 2016. For women, the age-adjusted mortality rate exhibited an increasing but up-and-down trend from 1981 to 1998 and a decreasing trend to 2016. For liver cancer among men, the age-adjusted mortality rate exhibited an increasing trend from 1981 to 2002 and a decreasing trend to 2016. For women, the age-adjusted mortality rate exhibited an increasing trend from 1981 to 2003 and a decreasing trend to 2016. The age-adjusted mortality rates of chronic liver disease, cirrhosis and liver cancer for both sexes are projected to decrease by more than 30% from 2016 to 2025 and by more than 55% from 2016 to 2035.

CONCLUSION

These results indicated that the 25 by 25 goal can be achieved for liver diseases mortality in Taiwan. In addition to viral hepatitis, the following risk factors may become the major causes of liver diseases in Taiwan in the future: alcohol and tobacco use, diabetes and obesity.

Handedness and Age of Death: New Evidence on a Puzzling Relationship

Based on 5743 deaths in the United States and Canada, the present study was undertaken to verify and extend results suggesting that left handers have a shorter life expectancy than right handers. Unlike previous studies, which have merely dichotomized handedness, this study used a 5-point scale to measure handedness (extremely right handed, generally right handed, ambidextrous, generally left handed and extremely left handed). We found a significant tendency for one of the five handedness categories—those classified as generally left handers—to die at a significantly younger age than was true for the other four handedness groups. Our findings add to the controversy over links between handedness and age of death by suggesting that only a portion of left handers are at risk of premature death.

Modeling Future Cardiovascular Disease Mortality in the United States: National Trends and Racial and Ethnic Disparities

BACKGROUND

Accurate forecasting of cardiovascular disease mortality is crucial to guide policy and programming efforts. Prior forecasts often have not incorporated past trends in rates of reduction in cardiovascular disease mortality. This creates uncertainties about future trends in cardiovascular disease mortality and disparities.

METHODS AND RESULTS

To forecast US cardiovascular disease mortality and disparities to 2030, we developed a hierarchical bayesian model to determine and incorporate prior age, period, and cohort effects from 1979 to 2012, stratified by age, sex, and race, which we combined with expected demographic shifts to 2030. Data sources included the National Vital Statistics System, Surveillance, Epidemiology, and End Results (SEER) single-year population estimates, and US Bureau of Statistics 2012 national population projections. We projected coronary disease and stroke deaths to 2030, first on the basis of constant age, period, and cohort effects at 2012 values, as is most commonly done (conventional), and then with the use of more rigorous projections incorporating expected trends in age, period, and cohort effects (trend based). We primarily evaluated absolute mortality. The conventional model projected total coronary and stroke deaths by 2030 to increase by ≈18% (67 000 additional coronary deaths per year) and 50% (64 000 additional stroke deaths per year). Conversely, the trend-based model projected that coronary mortality would decrease by 2030 by ≈27% (79 000 fewer deaths per year) and stroke mortality would remain unchanged (200 fewer deaths per year). Health disparities will be improved in stroke deaths but not coronary deaths.

CONCLUSIONS

After prior mortality trends and expected demographic shifts are accounted for, total US coronary deaths are expected to decline, whereas stroke mortality will remain relatively constant. Health disparities in stroke but not coronary deaths will be improved but not eliminated. These age, period, and cohort approaches offer more plausible predictions than conventional estimates.

SEPARABLE FACTOR ANALYSIS WITH APPLICATIONS TO MORTALITY DATA

Human mortality data sets can be expressed as multiway data arrays, the dimensions of which correspond to categories by which mortality rates are reported, such as age, sex, country and year. Regression models for such data typically assume an independent error distribution or an error model that allows for dependence along at most one or two dimensions of the data array. However, failing to account for other dependencies can lead to inefficient estimates of regression parameters, inaccurate standard errors and poor predictions. An alternative to assuming independent errors is to allow for dependence along each dimension of the array using a separable covariance model. However, the number of parameters in this model increases rapidly with the dimensions of the array and, for many arrays, maximum likelihood estimates of the covariance parameters do not exist. In this paper, we propose a submodel of the separable covariance model that estimates the covariance matrix for each dimension as having factor analytic structure. This model can be viewed as an extension of factor analysis to array-valued data, as it uses a factor model to estimate the covariance along each dimension of the array. We discuss properties of this model as they relate to ordinary factor analysis, describe maximum likelihood and Bayesian estimation methods, and provide a likelihood ratio testing procedure for selecting the factor model ranks. We apply this methodology to the analysis of data from the Human Mortality Database, and show in a cross-validation experiment how it outperforms simpler methods. Additionally, we use this model to impute mortality rates for countries that have no mortality data for several years. Unlike other approaches, our methodology is able to estimate similarities between the mortality rates of countries, time periods and sexes, and use this information to assist with the imputations.

On the age dynamics of learned societies-taking the example of the Austrian Academy of Sciences

In a hierarchical organisation of stable size the annual intake is strictly determined by the number of deaths and a statutory retirement age (if there is one). In this paper we reconstruct the population of the Austrian Academy of Sciences from 1847 to 2005. For the Austrian Academy of Sciences we observe a shift of its age distribution towards older ages, which on the one hand is due to rising life expectancy, i.e., a rising age at death, as well as to an increased age at entry on the other hand. Therefore the number of new entrants has been fluctuating considerably-especially reflecting several statutory changes-and the length of tenure before reaching the age limit has declined during the second half of the last century.Based on alternative scenarios of the age distribution of incoming members-including a young, an old, the 'current' and a mixed-age model-we then project the population of the Austrian Academy and its ageing forward in time. Our results indicate that the 'optimum policy' would be to elect either young or old aged new members.

A comparison of models for dynamic life tables. Application to mortality data from the Valencia Region (Spain)

The graduation of mortality data aims to estimate the probabilities of death at age x, q ( x ), by means of an age-dependent function, whose parameters are adjusted from the crude probabilities that are directly obtainable from the data. However, current life tables have a problem, the need for periodic updates due to changes in mortality over short periods of time. The table containing mortality rates for different ages in different years, q ( xt ), is called a dynamic life table, which captures mortality variation over time. This paper proposes a review of the most commonly used dynamic models and compares the results obtained by each of them when applied to mortality data from the Valencia Region (Spain). The result of the comparison leads us to the conclusion that the Lee-Carter method offers the best results for both sexes, while that based on Heligman and Pollard functions provides the best fit for men alone. Our working method is of additional interest as it may be applied to mortality data for a wide range of ages in any geographical location, allowing the most appropriate dynamic life table to be selected for the case at hand.

Coherent mortality forecasts for a group of populations: an extension of the Lee-Carter method

Mortality patterns and trajectories in closely related populations are likely to be similar in some respects, and differences are unlikely to increase in the long run. It should therefore be possible to improve the mortality forecasts for individual countries by taking into account the patterns in a larger group. Using the Human Mortality Database, we apply the Lee-Carter model to a group of populations, allowing each its own age pattern and level of mortality but imposing shared rates of, change by age. Our forecasts also allow divergent patterns to continue for a while before tapering off. We forecast greater longevity gains for the United States and lesser ones for Japan relative to separate forecasts.

Evaluating the performance of the Lee-Carter method for forecasting mortality

Lee and Carter (LC) published a new statistical method for forecasting mortality in 1992. This paper examines its actual and hypothetical forecast errors, and compares them with Social Security forecast errors. Hypothetical historical projections suggest that LC tended to underproject gains, but by less than did Social Security. True e0 was within the ex ante 95% probability interval 97% of the time overall, but intervals were too broad up to 40 years and too narrow after 50 years. Projections to 1998 made after 1945 always contain errors of less than two years. Hypothetical projections for France, Sweden, Japan, and Canada would have done well. Changing age patterns of mortality decline over the century pose problems for the method.

Epidemiology of major age-related diseases in women compared to men

Many observations indicate that women have a longer life expectancy than men. Population-based studies report that mortality and morbidity are higher in men than in women. The gender difference is constant in cardiovascular disease, cancer and dementia, the more frequent diseases in industrialized countries; these chronic conditions strongly influence longevity and quality of life in old persons. Biological, behavioral and environmental factors emerge as major contributors to the difference in mortality, morbidity and case fatality. However, the causes of gender differences remain poorly understood.

The prognostic role of gender in survival of adult cancer patients. EUROCARE Working Group

Many observations indicate that women have a much longer expectancy of life than men. Some population-based studies on cancer patients support the idea of the role of gender in predicting survival. However, the data are somewhat contrasting and inconclusive. The purpose of this paper was to evaluate the prognostic role of gender for cancer patients, making use of the large set of survival data made available by the EUROCARE II project for the period 1985-1989. By applying a multivariate approach the major confounders such as age, geographical area and cancer site were considered in analysing survival data on more than 1 million cancer cases collected by 45 population-based cancer registries in 17 European countries. The results were consistent with the general observation that in the industrialised countries women tend to survive longer than men. The multivariate analysis showed better survival from cancer in women than in men, estimated as an overall 2% lower relative risk of dying. The female advantage was particularly evident in young cases, reduced in patients in middle age groups and in the oldest patients completely reversed so that at this age men had the better prognosis. Longer survival for women was not present immediately after diagnosis, but the major advantage was seen after 3 years of follow-up. The risk of death for women was significantly lower for cancer of the head and neck, oesophagus, stomach, liver and pancreas. For bladder cancer, the risk of death was significantly greater for women. These results can be explained by gender differences in sub-site distributions (head and neck and stomach) and by the differences in the stage at diagnosis (presumably bladder). However, the consistency of the data, evident only when a vast set of data is analysed, suggest that women may be intrinsically more robust than men in coping with cancer.

Combining probabilistic and subjective assessments of error to provide realistic appraisals of demographic forecast uncertainty: Alho's approach

"With increasing interest in forecast uncertainty, there is an evolving concern with assessing the degree of certainty we can attach to uncertainty itself. This concern is the subject of recent work by Juha Alho.... I examine [his] approach systematically and draw general conclusions about its efficacy. Suggestions for improvement are made."

What demographers can learn from fruit fly actuarial models and biology

Historically demographers have viewed the results of actuarial studies of nonhuman species, particularly those on invertebrates such as fruit flies, as largely irrelevant to investigations on human populations. In this paper I present life table data from large scale studies on the Mediterranean fruit fly, and show that they provide important insights into fundamental aspects of mortality relevant to human populations: the trajectory of mortality at older ages, sex mortality differentials, the concept of maximal life span, and demographic heterogeneity and selection. An overriding theme of the paper is the need for demographers to acquire a heightened awareness of new developments in biology including areas such as evolutionary ecology, experimental demography, and molecular medicine.

Predictability, complexity, and catastrophe in a collapsible model of population, development, and environmental interactions

"More and more population forecasts are being produced with associated 95 percent confidence intervals. How confident are we of those confidence intervals? In this paper, we produce a simulated dataset in which we know both past and future population sizes, and the true 95 percent confidence intervals at various future dates. We use the past data to produce population forecasts and estimated 95 percent confidence intervals using various functional forms. We, then, compare the true 95 percent confidence intervals with the estimated ones. This comparison shows that we are not at all confident of the estimated 95 percent confidence intervals." (SUMMARY IN FRE)

Disaggregation in population forecasting: do we need it? And how to do it simply

"We have described a method for reducing the dimensionality of the forecasting problem by parsimoniously modeling the evolution over time of the age schedules of vital rates. This method steers a middle course between forecasting aggregates and forecasting individual age specific rates: we reduce the problem to forecasting a single parameter for fertility and another one for mortality. We have described a number of refinements and extensions of those basic methods, which preserve their underlying structure and simplicity. In particular, we show how one can fit the model more simply, incorporate lower bounds to the forecasts of rates, disaggregate by sex or race, and prepare integrated forecasts of rates for a collection of regions. We also discuss alternate approaches to forecasting the estimated indices of fertility and mortality, including state-space methods. These many versions of the basic method have yielded remarkably similar results." (SUMMARY IN FRE)

Stochastic demographic forecasting

"This paper describes a particular approach to stochastic population forecasting, which is implemented for the U.S.A. through 2065. Statistical time series methods are combined with demographic models to produce plausible long run forecasts of vital rates, with probability distributions. The resulting mortality forecasts imply gains in future life expectancy that are roughly twice as large as those forecast by the Office of the Social Security Actuary.... Resulting stochastic forecasts of the elderly population, elderly dependency ratios, and payroll tax rates for health, education and pensions are presented."