The Changing Meaning of Aging: Historical Trends and Forecasts

Lidia Patricia Tome, CIDEHUS.UE, University of Évora
Filipe Ribeiro, CIDEHUS.UE, University of Évora
Maria Filomena Mendes, CIDEHUS.UE, University of Évora

“How long will we live?” is one of the most important questions that every individual would like to answer. Oeppen and Vaupel (2002) didn’t answer to this question, but demonstrated that life expectancy is breaking old theorized limits and with time every human can expect to live longer and with improved health (Vaupel, 2010). These changes have profound implications, not only for individuals, but also for society and the economy. Thus, the meaning of aging needs to be methodically re-evaluated and conventional measures of aging, as life expectancy at birth, need to be complemented with additional and innovative insights, allowing to present a wide perspective about lifespan, late life autonomy and frailty. The postponement of mortality is, nowadays, intimately connected with the emergence of the extremely old share of population, what was rare until around middle of last century (Vaupel, 2010). In Portugal, e.g., centenarians increased more than three-fold between 2010 and 2015, and age 75 is now seen as the new 65 in what concerns the remaining average lifespan.

Nowadays increasing lifespan is intimately related with improvements in survival after age 65, and thus, a coherent old-age mortality forecast will contribute with less biased forecasts to the female and male old-age mortality in Europe. We use Oeppen’s (2008) and Bergeron-Boucher et al. (2017) Compositional Data Analysis (CoDa) proposals, already proved to present very accurate results, not only to (1) forecast remaining life expectancy at older ages; (2) but mainly to re-examine the changing meaning of aging from the historical perspective and providing insights for the future; (3) analysing why an increasing lifespan may or may not be a synonymous of late life autonomy or frailty; (4) estimate the cause-specific probability of death at later ages; and (5) to evaluate the possible sex convergence in old-age mortality.


Backgroundand aim of the study

“How long will we live?” is one of themost important questions that every individual would like to answer. Oeppen andVaupel (2002) didn’t answer to this question, but demonstrated that lifeexpectancy is breaking old theorized limits and with time every human canexpect to live longer and with improved health (Vaupel, 2010). In Japan, thebest-practice country, a female born in 2014 could expect to live around 87years, resulting from a 6-month average yearly increase since 1947. At the sameyear, in Portugal, e.g., female LE was around 84 years due to a 5-month averageyearly increase since 1940 (HMD, 2017). Thus, as a consequence of low fertilityrates together with increasing lifespan, the elderly population is rapidlyincreasing and according to the UN Population Division, the proportion of individualsover age 60 is projected to more than triple by 2100 while those aged 80+ areprojected to increase more than seven-fold by 2100 (UN, 2015).

Longevity’s increase occurs as a result ofsignificant improvements in health and consequent reductions in mortalityrates. Past high levels of mortality are now experienced later in life,suggesting that senescence is being delayed and not stretched and individualsare reaching older ages in better physiological condition (Vaupel, 2010).

Observed mortality reductions are notfollowing the same rhythm of decrease across different causes-of-death (CODs -circulatory system diseases and neoplasms are leading), countries, ages, or sexes(Meslé, 2006) and in association with the aging population, the incidence ofcoronary artery disease and subsequent myocardial ischemia or infarction isalso rising. As the elderly population is growing, the prevalence ofcardiovascular disease follows as consequence (Shan et al., 2014).

This study main goal is to provide a completeand insightful analysis of mortality, aging and old-age remaining lifespananalyzing heterogeneous patterns between countries, providing simultaneously,an international comparison with best-practice countries and evaluating theimpact of different causes of death in life expectancy itself. To achieve our maingoals, we use Oeppen’s (2008) and Bergeron-Boucher et al. (2017) CoDa proposals,already proved to present very accurate results, not only to (1) forecastremaining life expectancy at older ages; (2) but mainly to re-examine thechanging meaning of aging from the historical perspective and providinginsights for the future; (3) analysing why an increasing lifespan may or maynot be a synonymous of late life autonomy or frailty; (4) estimate thecause-specific probability of death at later ages; and (5) to evaluate thepossible sex convergence in old-age mortality.

Dataand Methods

Data used on this study comes from historicaldata on mortality patterns available on the Human Mortality Database and on theWorld Health Organization Mortality Database. Exposure-to-risk estimates arealso extracted from the HMD. As explained previously, the main focus of thisstudy is to forecast old-age mortality and re-evaluate the changing meaning ofaging for different countries as: Portugal, Spain, France, England, Sweden andJapan.

Concerning the methodological approach, theCoDa equivalent Lee-Carter method for forecasting mortality trends suggested byOeppen (2008) allows not only forecasting mortality trends for overall changesmortality rates (single-decrement forecast), but also disaggregate thoseby cause of death (multiple-decrement forecast). In opposition to mostknown approaches, Oeppen suggested using the life-table distribution of deaths dxand once that its sum equals the life-table radix (l0) we endup with compositional data. Despite maintaining the same baseline, some changeswere made by Bergeron-Boucher et al. (2017) trying to solve the subpopulationcoherence forecasting problems, and this approach is also here explored.

Thechanging meaning of aging and the increase of oldest-old: what can we expect?

Analyzing historical trends of mortalitypostponement and the emergence of the oldest-old (population aged 100+) allowsto understand how and why the elderly population is increasing fast in recent years.Taking the Portuguese population as example, one can realize, by observing Figure1 (where in Panel 1 are presented the historical trends of X5 and X10,i.e., the ages at which remaining life expectancies are, respectively 5 and 10 years;and in Panel 2, the increasing of population aged 100+), that senescence isbeing postponed at a faster pace since around 1980 and centenarians increasedmore than three-fold between 2010 and 2015.

Despite the very decent evolutionary patternsfor Portugal, it is expected that different countries, especially in Europe, presentsimilar insightful or even better prospective results. Nevertheless, anincreasing lifespan may or may not be a synonymous of late life autonomy orfrailty and cause- and sex-specific probabilities of death at later ages may stronglydiffer between the same countries.

 

Figure 1: The changing meaning of aging and the increase of oldest-old in Portugal

Panel 1: Postponing mortality at older ages

Panel 2: Population aged 100+

x5_10.eps

100plus.eps

Source: HMD & Statistics Portugal, authors calculations.

 

Presented in Session 1235: Mortality and Longevity