Skip to main content
Browse Subject Areas

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

The contribution of social participation to differences in life expectancy and healthy years among the older population: A comparison between Chile, Costa Rica and Spain

  • Sarahí Rueda-Salazar ,

    Roles Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Resources, Software, Supervision, Validation, Visualization, Writing – original draft, Writing – review & editing

    Affiliation Centre d’Estudis Demogràfics (CED), Autonomous University of Barcelona (UAB), Barcelona, Spain

  • Jeroen Spijker ,

    Contributed equally to this work with: Jeroen Spijker, Daniel Devolder, Cecilia Albala

    Roles Funding acquisition, Resources, Supervision, Writing – review & editing

    Affiliation Centre d’Estudis Demogràfics (CED), Autonomous University of Barcelona (UAB), Barcelona, Spain

  • Daniel Devolder ,

    Contributed equally to this work with: Jeroen Spijker, Daniel Devolder, Cecilia Albala

    Roles Funding acquisition, Resources, Supervision, Writing – review & editing

    Affiliation Centre d’Estudis Demogràfics (CED), Autonomous University of Barcelona (UAB), Barcelona, Spain

  • Cecilia Albala

    Contributed equally to this work with: Jeroen Spijker, Daniel Devolder, Cecilia Albala

    Roles Conceptualization, Formal analysis, Methodology, Supervision, Writing – review & editing

    Affiliation Unidad de Nutrición Pública, Instituto de la Nutrición y Tecnología de los Alimentos (INTA), Universidad de Chile, Santiago, Chile


We study the health trajectories of the population aged over 60, comparing between one European and two Latin American countries (Spain, Chile and Costa Rica) which have similar longevity patterns. Our focus is on functional limitation and mortality risks, considering differences by gender, education and social participation. Data come from national panel surveys (EPS, CRELES, SHARE). Multistate modelling is used to estimate transition probabilities between two health states: healthy to unhealthy, unhealthy to healthy as well as the transition to death from healthy or unhealthy states, to estimate the duration of stay in a specific state (computing healthy and unhealthy life expectancies) and the effect of the selected covariates. Results show that older Costa Ricans have the smallest gender gap in life expectancy but women have a lower healthy life expectancy compared to those in Chile and Spain. Participation in social activities leads to higher healthy life expectancy among the elderly in Costa Rica and Spain, whilst there were no relevant educational differences observed in longevity in the analysed countries. To conclude: despite the different patterns observed in health transitions and survival across the three countries, social participation is associated with greater health and longevity among people of old age, with little effect coming from educational attainment. Public policies should therefore be aimed at reducing unhealthy life years and dependency at advanced ages by promoting more engagement in social activities, especially among vulnerable groups who are more likely to experience impairment from a younger age.


Ageing in Latin America and the Caribbean (LAC) is a relatively recent phenomenon in contrast to Europe. However, in recent decades this region has witnessed the most rapid increase in the proportion of older population compared to other regions and other periods in history [13]; in the period 2000–2020 it increased by 57% (from 8.3% to 13.0%), with a growth similar to Asia, while European countries only recorded an average increase of about 25% [1]. The speed of the ageing process in LAC is largely due to the combined effect of significant declines in mortality at advanced ages and even sharper decreases in fertility levels since the 1960s, both of which have intensified over the last 2–3 decades [1,4,5].

The countries with the largest proportion of older people (60+) in LAC in 2020 are Cuba (21.3%), Uruguay (20.2%), Chile (17.4%), Argentina (15.5%) and Costa Rica (15%) [1]. The population of these countries also have the longest mean life span in the region [6], with life expectancies (LEs) ranging from 80–82 years for women and 73–79 years for men. However, as we will argue later, the level of LE does not exactly provide sufficient information on the health status and quality of life of the older adults.

Nowadays, the health of older people is represented by the preservation of their physical and mental capacities needed to function in their immediate environment ─which allows for well-being in old age─rather than by the absence of disease or experimenting bad health conditions [7]. Consequently, the ability of the older people to live independently, understood as the possibility for a person to carry out daily activities without receiving supervision or help from others [8], is the main objective for an active and healthy ageing [9].

Various studies indicate that the population of LAC countries that reached sixty years of age in the 2000s has benefited from the deployment of innovations in medicine in childhood and infancy which took place after 1945 [10]. Hence, and in comparison with previous birth cohorts, their better survival is mainly due to improvements in medical treatment and less exposure to contagious diseases, rather than to improvements in living conditions. However, at the same time, fragile economies, rising poverty levels, expanding economic social and economic inequalities and contracting access to collective financed services and resources in LAC [11] are likely to play a part in lower healthy LEs at older ages than those observed in European countries that have a strong welfare state. For instance, at age 60, the average HLE in LAC was 16.1 years in 2019. This is 2.5 years lower than the average for Western Europe. The highest HLE in LAC is recorded in Costa Rica (18.7 years), which is just above the Western European average, but below that of Spain (18.8 years), Sweden (18.9 years), Switzerland (19.5 years) and France (19.7). Chile also has one of the highest HLE among LAC (18.0 years at age 60), but is lower than most Western European countries [6]. Latin American health studies relate their worse health conditions among elderly to poor conditions experienced early in life [12]. Moreover, it is expected that disability at older ages will actually increase in future years as chances of surviving to old age have also improved [13]. Some comparative studies between European countries and Latin America have shown that differences in levels of functional limitations at older ages are also related to differences in general economic welfare, as well as accessibility to and quality of public services, in particular health care and primary attention [14] which are better developed and more effective in Western Europe. This implies a growing demand for health care and attention needs at older ages in the case of LAC [15,16]. In the case of Southern European countries, including Spain, this is also an issue [1719], as public institutions do not have enough capacity to manage the provision of care for the elderly, reason why the family is still the fundamental basis for elderly support and care there. In contrast, in Nordic and Western European countries the State is the main provider of such care [2023].

In LAC, family ties have created a compensatory system of intergenerational support networks within extended family households ─an expansion of the nuclear family─ [4,11,16,19,2429], as is the case in Chile and Costa Rica. While nuclear families are more common in Spain, its average number of household members is higher than in Northern and Western European countries [30], where the partner (women in most cases) represents the largest and principal caregiver [17,3134].

The timing and duration of the demographic transition in each of the Latin American countries [35] have direct consequences on the availability of relatives and on the support network of the older population [19]. Consequently, the countries with the earliest decline in their fertility rates will be those with the highest proportions of elderly people living alone and the lowest proportions of living with children, as is the case of Uruguay and Argentina [15].

In this context, social participation, which includes volunteering, participating in civil or religious organisations, doing sports, etc., emerges as part of successful strategies for preserving independence when older workers transition from labour market exit to retirement [36] and consequently contributes in reducing the costs of long-term care for those in the fourth age [37]. Social participation is considered an important tool to promote the comprehensive well-being of this population group in the guidelines for active aging by the WHO [38]. Studies indicate that social participation not only reduces social exclusion [39] but also have various benefits for the wellbeing of older people in different countries [37,40,41]. In general, they broadly favour two health domains: cognitive processes (increasing cognitive functioning) and prevention of risk factors. The former is associated with psychological objectives focused on emotions such as reducing anxiety, increasing life satisfaction and wellbeing. The latter is aimed at stimulating healthier behaviours in terms of diet, alcohol consumption and smoking [9].

However, studies on elderly health trends in LAC are still scarce [12,4250], and even more so between-country comparisons [5153]. Most studies are descriptive [26], based on the Survey on Health and Well-Being of Elders (SABE) [11]), conducted in seven cities in LAC. This survey is one of the few conducted in the region for studying elderly health conditions since the turn of the century. It was designed to be cross-sectional, although in Santiago de Chile, Sao Paulo and Mexico City there were longitudinal follow-ups.

Most of the analyses on health conditions in LAC are based on the use of prevalence rates for the estimation of Healthy Life Expectancies (HLEs) by means of the Sullivan method [47,54,55] with the exception of the study by Moreno, Albala and Lera et al. [56] who use incidence rates to estimate multiple state models. As we also use longitudinal panel survey data, the same method is applied in this study. Incidence rates are more robust than prevalence rates because they capture changes in health status over the life cycle (i.e. across age) and therefore, HLE estimations are closer to the health dynamics experienced by the population [5760].

The present study analyses the functional health trajectories of the elderly in order to compare recent health trends of the population over 60 years of age between three countries: Chile and Costa Rica ─known in Latin America for their high LE─ and one European country, namely Spain. There are two main reasons for their selection. Firstly, a practical one, as excellent national-level longitudinal data exist for all three countries. Secondly, the three countries are similar in terms of life expectancy, elderly educational levels and income. For instance, in 2019, LE at age 60 in Chile was 22.4 years for men and 25.9 years for women, in Costa Rica 23.6 and 26.4 years, and in Spain 22.6 and 26.7 years respectively [6], with gaps between TLE (Total Life Expectancy) and HLE of 5.9 years in Chile and 6.3 years in Costa Rica and 6.2 years in Spain (both sexes combined). However, at the same time, their health care systems are quite different, which, in the context of the obtained results, is an issue we briefly reflect upon in the discussion.

More specifically, the study explores the differences between countries in the level of transition probabilities between two health states: healthy to unhealthy, unhealthy to healthy, as well as the transition (from the unhealthy and healthy state) to death. This allows computing the expected duration of staying in a specific state, i.e. the healthy and unhealthy life expectancy at age 60. We also consider differences according to gender, educational attainment and social participation. The analysis focuses on the contribution of these variables to the risk of each health trajectory as they are considered to play a role in the conservation of functional capacities at advanced ages and thereby in the reduction of dependency and the needs for long term health care.

2. Data and method

Two waves of three longitudinal panel-type surveys are used for each country (see Table 1): Chile’s Social Protection Survey "EPS" [61] (years 2004 and 2006), Costa Rica’s Longevity and Healthy Aging Study "CRELES" [62] (years 2005 and 2007) and for Spain the "Surveys of Health, Ageing and Retirement in Europe” (SHARE) [63] (years 2004/05 and 2007). Although each survey has additional waves, we only analysed the interval between the first two waves as this was the only way to ensure that countries could be compared and that the issue of attrition was minimal. However, all waves were analysed and the results are provided in the Supplementary material.

Table 1. Chile, Costa Rica and Spain: Population (60+) Final sample distribution by age, educational level and social participation.

Chile: Longitudinal Social Protection Survey (EPS)

The nationally-representative EPS survey was first conducted by the “sub-secretaria de Protección Social” in 2004 aimed to provide information on the socioeconomic characteristics, family history, household conditions, work history, the social protection system (pensions) and health conditions of the population aged 18 and over [64]. 3,562 respondents aged 60+ participated in the baseline sample. Of those, 3,344 also participated in the 2007 wave (see S1 Fig). And, overall, the survey did not use proxy respondents in any of their waves.

Costa Rica Study on Longevity and Healthy Aging (CRELES)

The CRELES survey is the only source of longitudinal information on older people (60+) in Costa Rica. It is conducted by the Central American Population Centre (CCP) and the Health Research Institute (INISA) of the University of Costa Rica with the collaboration of other national institutes. This survey collects information on household conditions, socioeconomic characteristics, lifestyles, and specific health information (i.e. self-perceived health, physical examinations: blood pressure, anthropometry, mini-mental test, blood and urine collection which allows the study of a wide variety of bio-markers). We used data from the 2005 (baseline) and 2007 rounds without including new interviewed individuals after the baseline [75]. In the 2005 baseline sample, there were 2,827 respondents aged 60+, out of whom 2,631 also participated in the first follow-up (see S2 Fig). For the purpose of our analysis, we also excluded 661 respondents from the baseline sample who required a proxy respondent, which left us with 1970 cases. Proxies are usually employed when the main respondent is not able to do the interview because of health reasons. Taking a closer look at these respondents, their age profile showed that 75% were aged 80+, which is not surprising given the high oversampling of older people at baseline. Removing these cases from the sample therefore made the age-distribution more comparable with that of Chile and Spain. Moreover, the life expectancy at age 60 of these respondents was approximately 10 years less than that of respondents who partook the interview without the help of someone else. Including these cases would have biased the results too much (see also the Discussion).

Health, Ageing and Retirement Survey of Europe (SHARE)

SHARE is a longitudinal (panel-type) survey that collects information on health, socio-economic status, employment trajectories, family (including the partner of the selected persons regardless of their age) and social networks of 140,000 individuals aged 50 and over in 27 European countries and Israel. Currently, information is available for 7 rounds, covering the period 2004–2017 [65]. However, although data is available for more recent rounds, only those respondents from the Spanish sample aged 60 and over who participated in the 2004 and 2007 waves were selected to enable the results for Spain to be compared with those of Chile and Costa Rica. From the 1,554 individuals who were interviewed at baseline, 1,372 persons also participated at the first follow-up, 1317 of whom without a proxy respondent and with a known health state (see S3 Fig). Respondents with a proxy respondent were removed to be consistent with the sample methodology applied to Costa Rica.


The health status (dependent variable) of the population under study is estimated through the self-reporting of functional capacity to perform four Basic Activities of Daily Living (ADLs), namely: bathing, walking across a room, getting in or out of bed, and eating (see S2 Table). Therefore, the individual who did not report any of the four ADLs described above is considered to be healthy. These criteria are also used in the estimation of healthy and unhealthy years of life. We choose this indicator as it can be constructed from the data for all three countries under study and, as we explain in the discussion, the obtained results are compatibles with alternative indicators.

For the identification of the dead state, required for the calculation of transition probabilities between each health state and death (see method section), we were able to obtain the exact date of the occurrence. In the case of Chile, the information on deaths came from civil registries that were added to the survey, while in the Costa Rican and Spanish surveys, the retrospective information on the dates of deaths in the household was reported by a household member [66,67].

The main covariables in our analysis are: education and social participation. The questions on social participation vary according to survey (see S3 Table) but we were able to dichotomise the responses to whether or not the respondent was participating in any activity during the last six months). The education variable was also dichotomised into primary or less and secondary or more. Finally, sex and age are control variables.


The methodology used is based on the estimation of multi-state transition models. These models are used in survival analysis and allow the calculation of transition probabilities between health states by means of a stochastic process. These models consider changes in health status on the life cycle of individuals exposed to current morbidity and mortality conditions. This means that they capture health dynamics of the population between surveys waves through incidence rates. Hence, these are calculated transition probabilities between health states at a specific time period delineated by two ages pij(x,y) (probability that a person in state i at age x will be in state j at age y). The use of these models also allows us to estimate reversible transitions [68], something which is not possible when using prevalence rates (e.g. Sullivan method).

Multi-state models also allow to simultaneously explore the effect of variables on transitions between states and, finally, to estimate the duration of staying in a specific state. In the context of these models, all the LEs that we report here correspond to the expected duration of time regardless of the initial state occupied (Marginal Life Expectancy).

In this study we used an "Illness-death model with recovery" with three states: two transitory states: “healthy” and “unhealthy”, and a terminal (absorbing) state that corresponds to “death”. This is a mixed model [69] where the first two states are censored in the interval of two waves, that is, the exact moment when the transition between the health states occurred is not known, but the time interval between transitions is known, which corresponds to the date of the surveys.

The "healthy" and "unhealthy" states are recurrent as individuals may enter and exit them as many times as they are observed, and the “death” state only allows, by definition, entry into this state, once.

We used the R package "msm", version 1.6.7 [70] for the estimation of the “multistate survival models” to derive the hazard rate or instantaneous probability [71] of the transitions between states and for computing the hazard ratios of the explanatory variables (see S3 Table). These models include age as a continuous dependent variable, which means that transition probabilities between states increases or decreases log-linearly with age, following a Gompertz function. In this case, the probability function is estimated using a constant piecewise approximation conditioned by age [72], which means that the instantaneous probability of transition between states is constant within each age interval.

However, the “msm” package only allows estimating life expectancies estimates based on an exponential function, using age-constant rates [72]. We therefore used the R-package "elect", version 0.2 [73] which can use the model based on the Gompertz function to derive so-called “marginal” life expectancies [74]. These are the expected times spent in healthy or unhealthy states, regardless of the initial state of occupancy, as well as the state-specific life expectancy that refers to the expected occupancy time in each state.

Specifically, life expectancies are estimated based on numerical integration, which also provides an extrapolation of the previously calculated multistate survival model beyond the age range in the data. Additionally, the multinomial logistic regression model is applied using the state distributions at baseline (first observation) to estimate the effects of the time-independent covariates (age, education and social participation) on life expectancy and healthy life expectancy (see Van den Hout [74], section 7.3). Confidence intervals for the estimated values of the life expectancies were also calculated from 500 repeated simulations based on the asymptotic properties of the maximum likelihood estimator applied for the multi-state model [72,74,75], thereby sampling the stochastic process.

In this study we limited ourselves to estimate the total marginal life expectancy (TLE), the marginal life expectancy in healthy state (HLE) and the marginal life expectancy in unhealthy state (ULE), according to the selected variables.

3. Results

Estimates of total life expectancy at age 60 during the mid-2000s (see Table 2) show, as expected, that older women have longer TLE than men as gender differences range from 4.9 years in Spain to 3.4 years in Costa Rica. While women in the three countries have similar life expectancies, the TLE is close to 2 years higher among male Costa Ricans compared to Spanish and Chilean men.

Table 2. Total life expectancy, healthy life expectancy and unhealthy life expectancy at age 60 in years and the percentage of healthy years by gender for Chile, Costa Rica and Spain during the mid-2000s.

In each country, gender differences are also observed in unhealthy LE, with older women living more years in poor health than their male counterparts, especially in Spain and Costa Rica (about an additional 3 years). This also translates to a 5%-9% lower proportion of years that women live in good health. In terms of absolute values, country-differences in expected unhealthy years (ULE) at age 60 range from 7.6 years in Costa Rica to 4.2 years in Chile. Women in Costa Rica have a 3.3-years lower healthy life expectancy (HLE) and women in Spain a 0.6-year lower HLE than Chilean women.

The pattern is slightly different in the case of men. During the period studied, TLE at age 60 of Costa Ricans (22.8 years) was 1.8 years higher than what was estimated for both Spanish and Chilean men. On the other hand, levels of HLE were quite similar among the three countries (18.9, 18.7 and 18.7 years in, respectively, Spain, Chile and Costa Rica). This implies that the expected proportion that 60-year-olds can expect to live in good health in Costa Rica (80.0%) is below that of Spain (90.0%) and Chile (88.7%) and that the ULE is also highest in Costa Rica (4.6 years, against 2.4 years in Chile and 2.1 years in Spain).

If we analyse the TLE across ages, a monotonically downward trend for both sexes is observed from age 60 to 90 in the three countries (see Fig 1). This is consistent with a Gompertz distribution [68]. Regarding the trends for healthy LE, the country-specific trajectories converge from about age 80, whilst unhealthy LE remains remarkably constant across age. Both trends apply to each sex, although gender difference in unhealthy LE in Costa Rica and Spain are wider than in Chile. However, it should be noted that the sample of respondents aged 80+ is particularly small for Chile, making it difficult to estimate a robust pattern of life expectancy for this age group.

Fig 1. Total life expectancy, healthy life expectancy and unhealthy life expectancy from age 60 to 90, by sex.

Chile, Costa Rica and Spain. Note: LEs calculated with “msm” and “elect” R Packages, Confidence Intervals computed from 500 replications.

Life expectancy by social participation

Our next step was to calculate marginal life expectancies according to different personal characteristics. First of all, estimates were made by educational level. Results showed no significant effect in either LE or HLE of older men and women in the three countries. However, in terms of the percentage of healthy years, the higher educated lived an estimated 1–2 percentage points more years in good health than the lower educated in the two LAC countries, 5 in the case of Spanish men and 8 percentage points in the case of Spanish women (see supplementary materials S4 Table and S4 Fig).

Larger differences in LE according to health status are observed for social participation, although none are statistically significant (Table 3). Social participation contributes to a higher TLE at age 60 in all three countries, with similar but non-significant differences observed among both sexes between those who participate socially and those who don’t, ranging between 1.0 years among Spanish men to 2.8 years among Costa Rican men.

Table 3. Total life expectancy, healthy life expectancy, unhealthy life expectancy and % healthy years at 60 years old by social participation and gender.

Chile, Costa Rica and Spain.

In the case of Costa Rica and Spain, the overall 1–3 years difference in TLE according to social participation is more than accounted for by extra healthy years as the socially active have roughly 4 additional healthy life years in Costa Rica and 2 extra years in Spain compared to those who are not active (levels are similar for men and women), while at the same time, ULE is between 0.6 (Spanish men) and 2.1 (Costa Rican women) years higher among those who do not partake in social participation. These HLE differences translate to a higher proportion of LE that is spend in good health among the socially active compared to those who don’t, especially in the case of Costa Rica (10% higher among women and 8% higher among men). Differences in Spain are, respectively, 5% and 3%. However, in absolute terms, the Spanish live more years in good health than Costa Ricans (up to 92% in the case of socially active men). Finally, in the case of Chile, virtually all of the additional 1.8/1.4 years of TLE among socially active women/men compared to the non-active is attributed to additional healthy years. As a result, there are no notable differences in the %HLE between the two social participation categories, although men have about a 5% advantage above women, with levels being most similar to those observed in Spain.

Lastly, even when we found no statistical differences between the dichotomous categories of education and social participation in the three LE indicators (see S5 Table), results showed that in terms of the proportion of life expectancy in good health (%HLE), there is a consistent educational gradient, especially in Spain (Table 4).

Table 4. Percentage of healthy years by social participation and educational level.

Chile, Costa Rica and Spain.

For instance, for secondary school or higher educated 60-year old Spanish men, 95% of their remaining life expectancy will be in good health if they also partake in social participation compared to 88% for those who were less educated. For the non-social participants, the percentages are, respectively, 1 and 3 percentage points lower. Although health of Spanish women is slightly worse than that of men, educational differences are about the same.

4. Discussion

This comparative study of old age people in Chile, Costa Rica and Spain, using three different panel data sources and multistate modelling for the estimation of marginal life expectancy and healthy years from age 60 onwards, provides new evidence on the relation between health and survival, taking into account gender, education and social participation.

The observation of women’s greater chance of survival but, at the same time, of having a higher number of unhealthy years than men is well known [6,48,51,76,77]. This “gender paradox” has been explained by a number of factors. While the female advantage in longevity is partly due to biological factors, women are also more likely to have suffered cumulative disadvantages during the life course because of socioeconomic and gender inequalities in educational attainment and employment conditions [78,79]. Additionally, there are specific behaviours and lifestyles differences, in particular obesity and medication use, where women perform worse than men [56,80,81]. Lastly, methodological issues like survey non-response and under-reporting of health problems by men may also be contributing factors [82].

This "gender paradox" is confirmed by our results, although with some exceptions. While all elderly women live longer but suffer a higher proportion of unhealthy years than men, for Costa Ricans there are no gender differences in HLE, although women, like in the other two countries, also live a greater proportion of years in an unhealthy state than men. In the particular case of Costa Rica, where the gender gap in LE was the smallest (3.4 years; see Fig 1), previous research that analysed causes of death has shown that the declining gender gap in LE is due to a more rapid reduction in male mortality, especially from circulatory system diseases [83].

Additionally, the fact that Costa Rica has an even higher male and female LE at age 60 than Spain, known to be one of the countries with the highest survival rates in the world [84], is not surprising given the existence of the so-called blue zones of exceptional longevity in Costa Rica [85,86]. In this sense, our study shows that Costa Rican men have a slight survival advantage over their peers in Chile and have longevity levels that are in the top ranking of longevity in the world, similar to Spain, Sweden, Canada, France, Japan and Switzerland but with a smaller gender gap [84,86,87]. Important known drivers of decreasing gender gaps in survival are changing patterns of risk factors, in particular the reduction in smoking among men [88]. On the other hand, the health condition of Costa Rican women is less favourable as they have the lowest percentages of healthy years compared to the other countries studied here and also compared to Mexico, Puerto Rico and the United States [89].

The advantage observed in Spain and Chile compared to Costa Rica in terms of healthy life years, especially regarding women, could be due to a cohort effect. Some studies suggest that the reduction in mortality and morbidity is largely attributable to changes in the early life cycle [9092]. This is the case of the cohorts that benefited from the health improvements, economic progress and advances in health observed during the mid-1950s and especially after the post-war period in Spain. In addition, studies have shown that during the period 1985–2000, there was a disability compression in Spain, especially for women [93].

However, there are also other elements likely to play a role, including the type of diet and public expenditure on the national health system. With respect to the former, numerous studies suggest that the Mediterranean diet is an important factor in reducing morbidity and mortality for both functional capacities [94] and mental health [95]. In contrast, in Chile, even though, there are high rates of obesity, which is more frequent in women, affecting the difference in healthy years between men and women [96], a recent study conducted showed no evidence to support either a negative or a positive effect of overweight or obesity on LE and DFLE in Chile [56].

The continual improvements in life expectancy at all ages in each of the three countries over the last decades have also been attributed to the benefits of their health care systems, albeit that certain deficiencies do exist in the two LAC countries. While in Spain, health care has been universal and free of charge since 1986 [97], in Chile, it has had a mixed health system since the 1980s, composed of a public part called the National Health Fund (FONASA) and a private part, the Social Security Institutions (ISAPRES) [98]. Although most of the health coverage is assumed by the public part, the State invests relatively little in the health sector and the investment of public expenditure in the national health system is not sufficiently broad to cover the needs of the population in terms of primary and preferential health care, with the poor being the most disadvantaged [99].

In Costa Rica, although the health system has been universal and free of charge since the 1970s [100] and it is perceived by their population as efficient [101,102], there are substantial territorial inequalities in terms of morbidity and access to primary care [103]. The lower percentage of healthy years for women in Costa Rica, compared to the other two countries, may therefore be associated with geographical inequalities and differential access to health care.

In addition, the Costa Rican sample concentrates a higher proportion of cohorts born before 1930 (40%), against 27% and 36% of the sample in Chile and Spain. Especially younger cohorts were benefited by improved living conditions in the countries after the post-war period (1950-) (i.e. higher standards of hygiene, greater access to drinking water and health care, advances in medicine, economic development, educational expansion, healthier lifestyles, among others [26].

The minor variation in life expectancy by educational level in especially the Chilean and Costa Rican elders was also observed in Japan [104] and in older men in Mexico and supported by [105] and [106] who observed a greater contribution of material resources (such as income) than education to changes in health at older ages. The 5 and 8 percentage points more years in good health among higher compared to lower educated Spanish men and women is consistent with recent Spanish research that provided evidence of slowly increasing educational differences in life expectancy at age 50 since the 1960s [105].

Diverse patterns by social participation

Results indicated differences in survival according to social participation, although not equally in each of the three countries analysed. Social participation increases LE and the percentage of healthy years in all three countries, but especially in Costa Rica, followed by Spain. The benefits of social participation in older adults are consistent with previous studies that indicate an increase in functional capacities, quality of life and well-being in terms of subjective health perception [107] and survival at older ages [41]. There is evidence that participation contributes to a lower mortality risk, although with differences by age and gender. Younger older women enjoy greater benefit from participation than their older male counterparts [108]. Activities associated with social participation such as volunteering (e.g. in elementary schools), participation in sports teams and clubs, increase strength, walking speed and general activity at follow-up [9] and lower future long-term care costs [37]. Additionally, such activities, that also include time spent with friends, are also reported to have a positive effect on survival in the older population [40].

In this sense, social participation has been proposed as a multi-purpose strategy to address different targets as psychological wellbeing, physical and cognitive functioning and interventions aimed to reduce health risk factors [107]. Regardless of the strategy implemented, participation provides the mechanisms for social support and social cohesion in the community so that older people can experience greater wellbeing within the framework of active ageing [109], reducing situations of dependency and the demand for care at very old ages. These benefits have been widely observed among older people in Costa Rica, due to initiatives promoted by the State and the National Council for Older Persons (CONAPAM) [110] and implemented by various government and non-profit institutions [111] to ease the economic and care burden on families, especially the most vulnerable groups.

However, considering the relatively low participation rates among older people in two of the country samples, ─Spain (23%) and Chile (32%)─ there appear to be barriers to social participation. The factors that explain participation rates among the elderly, particularly in the case of Chile, and the subsequent lack of significance of this indicator in their functional trajectories and survival, are related to their socio-economic and health conditions, but also the other way around [112]. For instance, people who have a functional limitation or a certain condition will have less capacity to participate actively in society due to that health condition.

Other known determinants of social participation in older people include the availability of individual (human capital, which includes educational level) and contextual resources (household composition, social networks, cultural factors) and the housing environment [107,113,114]. It is therefore important that effective strategies are developed and implemented that remove barriers, promote and facilitate social integration between local communities and their older population in order to preserve their functional capacity, both physically and cognitively, as well as their general wellbeing.

Scope and limitations of the study

Although the analysed period roughly coincides between the three countries, one may consider the results somewhat outdated. However, we did this to ensure comparability, as otherwise a string of methodological issues would have placed a shadow on the comparability of the results. Nevertheless, in the supplementary material section we have included the results of a replication of the life expectancy estimations according to health status using all available longitudinal waves (Costa Rica: 2005, 2007, 2009; Chile: 2004, 2006, 2009, 2015, 2017; Spain 2004/05, 2007, 2011, 2013, 2015, 2017). The results do not significantly change for Spain. On the other hand, TLE at age 60 for Chile drops by almost 7 years between 2004–06 and 2009–15 in the case of men and 9 years in the case of women, with HLE dropping even more. The Costa Rican results for a later period also differs somewhat. Given these inconsistencies, we therefore decided to restrict our sample to only those who participated in the first two waves.

One important reason for this is the effect of attrition, i.e. the loss of individuals during survey follow-ups. This is a problem that is commonly reported in panel-type studies [115118]. Using all wave samples would have left us with a final population that was not representative of the national population (see S1S3 Figs). Attrition in terms of participating in panel survey is known to increase with age and, at least in the Chilean sample, is more likely to occur among people with a higher level of education [64].

On the other hand, as mortality is known to be higher in lower social strata, it changes the composition of the surviving population towards a selection of more “robust” individuals which, in turn, may weaken the observed relationship. This may explain why the lower educated in Chile and Costa Rica had slightly (but insignificantly) higher total life expectancies than the higher educated (see S4 Table). Although we controlled for age, there is also evidence that people with better health conditions are more likely to participate in panel surveys [119] and, indeed, in terms of HLE, educational differences were less (Chile) or reversed (Costa Rica).

Another issue that requires some clarification is that of proxy respondents. When it’s share of all respondents is considerable, as was the case of the first wave of the Costa Rican sample, it can have quite a distorting effect on the overall life expectancy. To illustrate: If we would have included the 661 cases who required a proxy respondent, the TLE at age 60 for women in the period 2005–2007 would have been 20.8 years and 20.0 years for men. By having taken these out of the calculations, the TLE increased to, respectively 26.2 and 22.9 years, i.e. levels that are similar to those obtained elsewhere (e.g. respectively 25.2 and 22.3 years in 2005 according to the WHO [6], see S7 Table). In turn, due to the highly selective initial Costa Rican sample that included a high number of proxies, the TLE for individuals who had a proxy respondent was just 13.5 years in the case of women and 12.5 years in the case of men. To be consistent, we also removed the 55 proxies used in the Spanish sample, while the Chilean sample did not use proxies.

A minor source of bias concerns the country comparison of the marginal life expectancy results according to social participation as they cover slightly different activities and items that are comparable but may have different wording (see S2 Table).

In the same vein, it should be mentioned that a different definition of limitation is likely to render different results. To facilitate comparison with other studies (see also next sub-section), we had measured health states according to a definition that had been developed for LAC countries who participated in the SABE study [120]. The authors identified an unhealthy state if the individual self-reported a limitation in at least one ADLs or two instrumental activities of daily living (IADL) or three types of mobility functional activities (MF) that measure the inability of the older adult to relate to his or her environment in daily life [45] (see also S2 Table). This definition had been used earlier by Moreno, Albala, Lera et al. [56].

However, due to the country differences in the phrasing and number of ADLs, IADLs and FM limitations (for instance, the Chilean survey only has one question on IADL and two on FM) it was more appropriate to adopt a more uniform definition of unhealthy state. We therefore opted to exclude the IADLs and FM limitations and define unhealthy state has having at least one out of four identical ADLs (eating/bathing/getting out the bed and walking across room). Hence, in order to get a sense of the sensitivity of these types of health indicators, we replicated the calculations of HLE for Costa Rica and Spain using the aforementioned definition and applying identical types of ADLs, IADLs and functional limitations (S8 Table). Results showed that particularly the health expectancy of women appear to be sensitive to the health indicator that is used as the overall years expected to live in good health is up to 5.6 years lower in the case of Costa Rican women according to the latter, broader, definition of functional limitation.

Comparison with other studies

Studies report that the use of one or another indicator to measure health status also affects differences in healthy life years [121123]. Even if the same indicator is used in a cross-national survey, variations in how the health question was translated in the respective questionnaires and country-specific perception of functional limitations may affect results [124,125]. Providing precise comparisons of our calculated healthy life years for Chile, Costa Rica and Spain with those from earlier studies are therefore hampered due to differences in methodology, indicators and sources. For instance, in the estimates on healthy years obtained by Moreno, Albala and Lera et al. [56] for Chilean older adults functional limitation was defined as having a limitation in at least one ADL, in two IADL, or in three mobility function activities, i.e. more restrictive than ours. Moreover, they also used a different longitudinal (panel) source (SABE) and one which only contained respondents from the city of Santiago. The initial sample was recruited in 1999–2000 and followed up in 2004–2005 and in 2009–2010. Their estimates of the proportion of remaining years from age 60 spent without functional limitation was about 25 percentage points lower than what we obtained (61.6% for men and 48.0% for women against our results of, respectively, 88.7% and 71.7%). On the other hand, if we would apply their indicator on the 2004–06 EPS dataset, we would obtain more similar results (65.3% and 62.6% for men and women, respectively).

Regarding Costa Rica there is more consistency between our results and those from other studies, including those by Rosero-Bixby [87] and Payne [89] that were based on the same CRELES survey. The first author estimated the total life expectancy for Costa Rica’s population aged 60, using a different methodology but results approximate those presented here as their estimates are one year below ours in the case of men two years in the case of women (men: 21.9 vs. 22.9 years and women 24.3 vs 26.2 years). The second study calculated LE according to health state from the age of 65 using "Multistate" models. Not being in a healthy state was defined as having at least one of 5 ADL limitations. Their results reported lower life expectancy values than those estimated in this study, but more similar than those reported by Rosero-Bixby (approximately one year less for both men and women). In terms of healthy years, his results report a fewer number of years in good health (12 healthy years for women and 14 years for men). This comparison with previous studies allows us to ascertain that in the case of Costa Rica the difference in the estimate of life expectancy is mainly due to the application of a different methodology as they used basically the same health indicator as we did here.

The results of healthy years and LE for Spain are very similar compared to those published in previous studies [79,105] and official statistics [6,126,127], despite the use of different methodologies and data sources for the calculation of LE. The other studies use health prevalence’s through the implementation of the Sullivan method [128131] and do not consider changes in functionality over the age pattern, as it is possible when using incidence rates. Our study is therefore more robust as the current health and mortality dynamics on the life cycle at older ages are captured by the calculation of HLE and LE. Additionally, some of these other studies use health condition prevalence according to distinct health indicators in surveys [131133] and the mortality experience from period life tables (which uses national death cases from vital registers and the whole population). Consequently, their results show a slightly lower proportion of healthy years than our estimates, while the life expectancies for both men and women are very similar. Lastly, as far as we are aware of, no study has yet looked at differences in HLE and LE in Spain according to social participation.

This observed difference between our results and those from previous research is mainly inherent to the source used [121,122]. In particular, mortality studies based on official mortality records have the entire population covered, including the institutionalized population, whereas our study only includes longitudinal information from a sample of private households. As a result, mortality from surveys can be at the same time underestimated, as the institutionalised population is normally excluded, or overestimated, when people who move from households to institutional care appear as deaths in surveys. Notwithstanding, overall, the mortality recorded in the surveys usually results in lower life expectancies than those estimated from vital records [134] as the population exposed to risk of mortality is underestimated due to effect of attrition in subsequent rounds of observation [135137].

Supporting information

S1 Fig. Diagram showing longitudinal analysis of the EPS Data.

Chile 2004–2017.


S2 Fig. Diagram showing longitudinal analysis of the CRELES Data.

Costa Rica 2005–2009.


S3 Fig. Diagram showing longitudinal analysis of the SHARE Data.

Spain 2004–2017.


S4 Fig. Male and female total life expectancy, healthy life expectancy and unhealthy life expectancy from 60 to 90, by educational level within countries.

Chile, Costa Rica and Spain.


S1 Table. Results using all waves for Chile, Costa Rica and Spain: Total life expectancy, healthy life expectancy and the percentage of healthy years by gender.


S2 Table. Questions of Activities of Daily Life (ADLs), Instrumental Activities of Daily Life (IADLs) and Mobility Function Activities (MF) by countries selected.


S3 Table. Question of social participation by studied countries.


S4 Table. Total life expectancy, healthy life expectancy and unhealthy life expectancy at 60 years old by educational level and gender.

Chile, Costa Rica and Spain.


S5 Table. Total life expectancy, healthy life expectancy and unhealthy life expectancy at 60 years old and the percentage of healthy years by social participation and educational level.

Chile, Costa Rica and Spain.


S6 Table. Percentage of healthy years by social participation and educational level.

Chile, Costa Rica and Spain.


S7 Table. Comparison with official statistics and other studies.


S8 Table. Comparison of results using an alternative definition of functional limitation.



  1. 1. United Nations. World Propulation Prospects 2019. Online edition. Rev.1 [Internet]. Department of Economic and Social Affairs/Population Division. 2019 [cited 2019 Aug 30]. Available from:
  2. 2. Rowland DT. Global Population Aging: History and Prospects. In: Uhlenberg P, editor. International Handbook of Population Aging [Internet]. Dordrecht: Springer Netherlands; 2009. p. 37–65. Available from:
  3. 3. Robine J-M, Jagger C, Mathers CD, Crimmins EM, Suzman RM. Determining Health Expectancies [Internet]. Robine J-M, Jagger C, Mathers CD, Crimmins EM, Suzman RM, editors. Chichester, UK: John Wiley & Sons, Ltd; 2002. 444 p. Available from:
  4. 4. Palloni A. Living Arrangements of Older Persons [Internet]. New York; 2000. (Paper presented at the United Nations Technical Meeting on Population Ageing and Living Arrangements of Older People. New York, February 8–10, 2000). Available from:
  5. 5. Rivero-cantillano R, Spijker J. Del rejuvenecimiento al envejecimiento de la población ¿o viceversa?: Chile en el contexto de América Latina, 1950–2050. Notas Poblacion [Internet]. 2015;(101):127–55. Available from:
  6. 6. World Health Organizacion. Life expectancy and Healthy life expectancy. Data by country [Internet]. Global Health Observatory data repository. 2018 [cited 2020 Dec 18]. Available from:
  7. 7. WHO. International classification of functioning, disability and health [Internet]. Organization WH, editor. Geneva: WHO Library Cataloguing-in-Publication Data; 2001. 315 p. Available from:
  8. 8. Zunzunegui MV. Evolución de la discapacidad y la dependencia. Una mirada internacional. Gac Sanit [Internet]. 2011 Dec;25(SUPPL. 2):12–20. Available from:
  9. 9. Menicheti J, Cipresso P, Bussolin D, Graffigna G. Engaging older people in healthy and active lifestyles: a systematic review. Ageing Soc [Internet]. 2016 Nov 27;36(10):2036–60. Available from:
  10. 10. Palloni A, McEniry M, Wong R, Peláez M. The Tide to Come. J Aging Health [Internet]. 2006 Apr 30;18(2):180–206. Available from: pmid:16614340
  11. 11. Wong R, Peláez M, Palloni A, Markides K. Survey Data for the Study of Aging in Latin America and the Caribbean. J Aging Health [Internet]. 2006 Apr 30;18(2):157–79. Available from: pmid:16614339
  12. 12. Monteverde M, Noronha K, Palloni A. Effect of early conditions on disability among the elderly in Latin America and the Caribbean. Popul Stud (NY) [Internet]. 2009 Mar;63(1):21–35. Available from: pmid:19184719
  13. 13. Monteverde M, Palloni A, Guillén M, Tomas S. Early Poverty and Future Life Expectancy with Disability among the Elderly in Argentina. Rev Latinoam Población. 2019;14(26):5–22.
  14. 14. Yokota RT de C, de Moura L, Andrade SSC de A, de Sá NNB, Nusselder WJ, Van Oyen H. Contribution of chronic conditions to gender disparities in disability in the older population in Brazil, 2013. Int J Public Health. 2016;61(9):1003–12. pmid:27339159
  15. 15. Saad PM. Los Adultos Mayores en América latina y el Caribe: arreglos residenciales y transferencias informales. Notas Poblacion [Internet]. 2005;32(80):127–54. Available from:
  16. 16. Wong R, Palloni A. Aging in Mexico and Latin America. In: Uhlenberg P, editor. International Handbook of Population Aging [Internet]. Dordrecht: Springer Netherlands; 2009. p. 231–52. Available from:
  17. 17. Tomassini C, Glaser K, Wolf DA, Broese Van Groenou MI, Grundy E. Living arrangements among older people: an overview of trends in Europe and the USA Examines trends in the living arrangements of older people in several European countries and the USA and the reasons for variabilities between those countries. Popul Trends Natl Stat [Internet]. 2004;Spring 200(115):24–35. Available from:
  18. 18. Glaser K, Tomassini C, Grundy E. Revisiting convergence and divergence: support for older people in Europe. Eur J Ageing [Internet]. 2004 Dec 10;1(1):64–72. Available from: pmid:28794703
  19. 19. Glaser K, Agree EM, Costenbader E, Camargo A, Trench B, Natividad J, et al. Fertility Decline, Family Structure, and Support for Older Persons in Latin America and Asia. J Aging Health [Internet]. 2006 Apr 30;18(2):259–91. Available from: pmid:16614344
  20. 20. Lagergren M, Johnell K, Schön P, Danielsson M. Healthy life years in old age: Swedish development 1980–2010 according to different health indicators. Scand J Public Health [Internet]. 2016 Feb 21;44(1):55–61. Available from: pmid:26392423
  21. 21. Crystal Stephen, Siegel MJ. Population Aging and Health Care Policy in Cross-National Perspective. In: Uhlenberg P, editor. International Handbook of Population Aging [Internet]. springer. Dordrecht: Springer Netherlands; 2009. p. 607–30. Available from:
  22. 22. Norton EC, Stearns SC. Health Care Expenditures. In: Uhlenberg P, editor. International Handbook of Population Aging [Internet]. Dordrecht: Springer Netherlands; 2009. p. 631–45. Available from:
  23. 23. Spijker J, Devolder D, Zueras P. The impact of demographic change in the balance between formal and informal old-age care in Spain. Results from a mixed microsimulation-agent-based model. Ageing Soc. 2020;In press.
  24. 24. De Vos SM. Household Composition in Latin America [Internet]. Boston, MA: Springer US; 1995. (The Springer Series on Demographic Methods and Population Analysis). Available from:
  25. 25. Vos SD. Extended Family Living Among Older People in Six Latin American Countries. J Gerontol [Internet]. 1990 May 1;45(3):S87–94. Available from: pmid:2335736
  26. 26. Palloni A, McEniry M. Aging and Health Status of Elderly in Latin America and the Caribbean: Preliminary Findings. J Cross Cult Gerontol [Internet]. 2007 Aug 7;22(3):263–85. Available from: pmid:17021957
  27. 27. Palloni A, Souza L. The fragility of the future and the tug of the past: Longevity. Demogr Res [Internet]. 2013 Sep 24;29(September):543–78. Available from: =
  28. 28. Pérez Amador J, Gilbert B. Una transición en edades avanzadas: cambios en los arreglos residenciales de los adultos mayores en siete ciudades latinoamericanas. Estud Demogr Urbanos Col Mex [Internet]. 2006;21(3):625–61. Available from:
  29. 29. Feng Z, Falkingham J, Liu X, Vlachantoni A. Changes in living arrangements and mortality among older people in China. SSM—Popul Heal [Internet]. 2017 Dec;3:9–19. Available from: pmid:29349200
  30. 30. Iacovou M, Skew AJ. Household composition across the new Europe: Where do the new Member States fit in? Demogr Res [Internet]. 2011 Aug 16;25:465–90. Available from:
  31. 31. Read S, Grundy E, Foverskov E. Socio-economic position and subjective health and well-being among older people in Europe: A systematic narrative review. Aging Ment Heal [Internet]. 2016 May 3;20(5):529–42. Available from: pmid:25806655
  32. 32. Grundy EM. Living Arrangements and the Health of Older Persons in Developed Countries. In: United Nation, Population Bulletin [Internet]. United Nations Department of Economic and Social Affairs Population Division; 2001. p. 311–29. Available from:
  33. 33. Dykstra PA, Fokkema T. Ties between Parents and their Adult Children: A Western European Typology of Late-Life Families [Internet]. 2011. Available from:
  34. 34. Agree EM, Glaser K. Demography of Informal Caregiving. In: Uhlenberg P, editor. International Handbook of Population Aging [Internet]. Dordrecht: Springer Netherlands; 2009. p. 647–68. Available from:
  35. 35. Rodríguez Wong L, Carvalho JAM de, Aguirre A. Duración de la transición demográfica en América Latina y su relación con el desarrollo humano. Estud Demogr Urbanos Col Mex [Internet]. 2000 Jan 1;15(1):185. Available from:
  36. 36. Alpass F, Towers A, Stephens C, Fitzgerald E, Stevenson B, Davey J. Independence, Well-being, and Social Participation in an Aging Population. Ann N Y Acad Sci [Internet]. 2007 Oct 1;1114(1):241–50. Available from:
  37. 37. Saito M, Aida J, Kondo N, Saito J, Kato H, Ota Y, et al. Reduced long-term care cost by social participation among older Japanese adults: a prospective follow-up study in JAGES. BMJ Open [Internet]. 2019 Mar 30;9(3):e024439. Available from: pmid:30928931
  38. 38. WHO. Measuring Health and Disability: Manual for WHO Disability Assessment Schedule WHODAS 2.0 [Internet]. Üstün, N Kostanjsek, S Chatterji JR, editor. World Health Organization. Malta: WHO Library Cataloguing-in-Publication Data Measuring; 2010. 152 p. Available from:
  39. 39. Walsh K, Scharf T, Keating N. Social exclusion of older persons: a scoping review and conceptual framework. Eur J Ageing [Internet]. 2017 Mar 11;14(1):81–98. Available from: pmid:28804395
  40. 40. Maier H, Klumb PL. Social participation and survival at older ages: is the effect driven by activity content or context? Eur J Ageing [Internet]. 2005 Mar 4;2(1):31–9. Available from: pmid:28794714
  41. 41. Dahan-Oliel N, Gelinas I, Mazer B. Social Participation in the Elderly: What Does the Literature Tell Us? Crit Rev Phys Rehabil Med [Internet]. 2008;20(2):159–76. Available from:,54883b73768ea435,769389123e0370ff.html.
  42. 42. Ashby-Mitchell K, Jagger C, Fouweather T, Anstey KJ. Life Expectancy with and without Cognitive Impairment in Seven Latin American and Caribbean Countries. Sued O, editor. PLoS One [Internet]. 2015 Mar 23;10(3):1–11. Available from:
  43. 43. Zunzunegui MV, Alvarado BE, Béland F, Vissandjee B. Explaining health differences between men and women in later life: A cross-city comparison in Latin America and the Caribbean. Soc Sci Med [Internet]. 2009;68(2):235–42. Available from: pmid:19036488
  44. 44. Alves LC, Andrade FCD, Corona LP, Santos JLF, Duarte YADO. Inequalities in Life Expectancy With Frailty Among Brazilian Older Adults: A Multistate Approach. Thorpe RJ, editor. Innov Aging [Internet]. 2019 Aug 1;3(4):1–9. Available from:
  45. 45. Fuentes-García A, Sánchez H, Lera L, Cea X, Albala C. Desigualdades socioeconómicas en el proceso de discapacidad en una cohorte de adultos mayores de Santiago de Chile. Gac Sanit [Internet]. 2013;27(3):226–32. Available from: pmid:23291031
  46. 46. Guerra RO, Alvarado BE, Zunzunegui MV. Life course, gender and ethnic inequalities in functional disability in a Brazilian urban elderly population. Aging Clin Exp Res [Internet]. 2008;20(1):53–61. Available from: pmid:18283229
  47. 47. Barbosa AR, Souza JMP, Lebrão ML, Laurenti R, Marucci M de FN. Functional limitations of Brazilian elderly by age and gender differences: data from SABE Survey. Cad Saude Publica [Internet]. 2005 Aug;21(4):1177–85. Available from: pmid:16021255
  48. 48. Lima ALB de, Lima KC de. Activity Limitation in the Elderly People and Inequalities in Brazil. OALib [Internet]. 2014;01(04):1–9. Available from:
  49. 49. Campos ACV, Albala C, Lera L, Sánchez H, Vargas AMD, e Ferreira EF. Gender differences in predictors of self-rated health among older adults in Brazil and Chile. BMC Public Health [Internet]. 2015 Dec 11;15(1):365. Available from: pmid:25884800
  50. 50. Campos ACV, e Ferreira EF, Vargas AMD, Albala C. Aging, Gender and Quality of Life (AGEQOL) study: factors associated with good quality of life in older Brazilian community-dwelling adults. Health Qual Life Outcomes [Internet]. 2014 Dec 30;12(1):166. Available from:
  51. 51. Minicuci N, Bélanger A. Discapacidad y Esperanza de Vida Saludable: una comparación entre Italia, Bulgaria, Canadá, y siete ciudades de América Latina y el Caribe. Resumo. Congreso de la Asociación Latinoamericana de población. 2008.
  52. 52. Prina AM, Wu Y-T, Kralj C, Acosta D, Acosta I, Guerra M, et al. Dependence- and Disability-Free Life Expectancy Across Eight Low- and Middle-Income Countries: A 10/66 Study. J Aging Health [Internet]. 2020 Jun 30;32(5–6):401–9. Available from: pmid:30698491
  53. 53. Chirinda W, Chen H. Comparative study of disability-free life expectancy across six low- and middle-income countries. Geriatr Gerontol Int [Internet]. 2017 Apr 1;17(4):637–44. Available from: pmid:27197085
  54. 54. Belliard M, Massa C, Redondo N. Análisis comparado de la esperanza de vida con salud en la Ciudad Autónoma de Buenos Aires. Poblac Buenos Aires [Internet]. 2013;10(18):7–29. Available from:
  55. 55. Rodríguez-Abrego G, Peña JED La, Zurita B, Ramírez TDJ. Esperanza de vida saludable en la población mexicana. Perinatol y Reprod humana [Internet]. 2006;20(1–3):4–18. Available from:
  56. 56. Moreno X, Albala C, Lera L, Leyton B, Angel B, Sánchez H. Gender, nutritional status and disability-free life expectancy among older people in Santiago, Chile. Devleesschauwer B, editor. PLoS One [Internet]. 2018 Mar 28;13(3):18. Available from:
  57. 57. Molla MT, Madans JH. Estimating healthy life expectancies using longitudinal survey data: Methods and techniques in population health measures. Natl Cent Heal Stat Vital Heal Stat. 2008;2(146). pmid:25093608
  58. 58. Eulenburg C, Schroeder J, Obi N, Heinz J, Seibold P, Rudolph A, et al. A Comprehensive Multistate Model Analyzing Associations of Various Risk Factors With the Course of Breast Cancer in a Population-Based Cohort of Breast Cancer Cases. Am J Epidemiol [Internet]. 2016 Feb 15;183(4):325–34. Available from: pmid:26823437
  59. 59. Jackson C. flexsurv: A Platform for Parametric Survival Modeling in R. J Stat Softw [Internet]. 2016;70(8). Available from: pmid:29593450
  60. 60. Brookmeyer R, Johnson E, Ziegler-Graham K, Arrighi HM. Forecasting the global burden of Alzheimer’s disease. Alzheimer’s Dement [Internet]. 2007 Jul;3(3):186–91. Available from: pmid:19595937
  61. 61. Subsecretaria de Protección Social. Encuesta de Protección Social (EPS) [Internet]. Available from:
  62. 62. Center BP. Costa Rica Estudio de Longevidad y Envejecimiento Saludable (CRELES) [Internet]. 2012. Available from:
  63. 63. Commission E. Survey of Health, Ageing and Retirement in Europe (SHARE). Available from:
  64. 64. Centro UC Encuestas y Estudios Longitudinales. Factores de Expansión Quinta Ronda Encuesta De Protección Social [Internet]. Santiago de Chile; 2015. Available from:ón-eps-2015.pdf.
  65. 65. Mirá P, Fawaz Y. Survey of Health, Ageing and Retiremment in Europe. SHARE SPAIN [Internet]. [cited 2020 Apr 6]. Available from:
  66. 66. Luis R-B, Fernández X, Dow WH. CRELES—Costa Rica: Estudio de Longevidad y Envejecimiento Saludable [Internet]. Centro Centroamericano de Población CCP. 2010. Available from:
  67. 67. SHARE-ERIC. Survey of health, Ageing and retirement in Europe. Release Guide 7.0.0 [Internet]. Munich; 2019. p. 1–83. Available from:
  68. 68. Saito Y, Robine J, Crimmins EM. The methods and materials of health expectancy. Stat J IAOS. 2014;30:209–23. pmid:30319718
  69. 69. Jackson C. Multi-state modelling with R: the msm package [Internet]. Cambridge, UK. 2019. p. 1–53. Available from:
  70. 70. Jackson C. Package ‘ msm ‘. Multi-State Markov and Hidden Markov Models in Continuous Time [Internet]. Version 1.6.7; 2019. Available from:
  71. 71. Kleinbaum DG, Klein M. Survival Analysis. A Self‐Learning Text [Internet]. Third Edit. Gail, M. Krickeberg, K. Samet, J.M. Tsiatis. A. Wong W, editor. New York, NY: Springer New York; 2012. 711 p. (Statistics for Biology and Health). Available from:
  72. 72. van den Hout A, Sum Chan M, Matthews F. Estimation of life expectancies using continuous-time multi-state models. Comput Methods Programs Biomed [Internet]. 2019 Sep;178(September):11–8. Available from: pmid:31416539
  73. 73. van den Hout A. ELECT: Estimation of life expectancies using continuous-time multi-state models [Internet]. Vignette ELECT version 0.2. 2016. p. 1–17. Available from:
  74. 74. van den Hout A. Multi-State Survival Models for Interval-Censored Data [Internet]. 1st Editio. Vol. 152, Monographs On Statistics and Applied Probability. New York: Chapman and Hall/CRC; 2016. 257 p. Available from:
  75. 75. Mandel M. Simulation-Based Confidence Intervals for Functions With Complicated Derivatives. Am Stat. 2013;67(2):76–81.
  76. 76. Van Oyen H, Cox B, Jagger C, Cambois E, Nusselder W, Gilles C, et al. Gender gaps in life expectancy and expected years with activity limitations at age 50 in the European Union: associations with macro-level structural indicators. Eur J Ageing [Internet]. 2010 Dec 3;7(4):229–37. Available from: pmid:28798631
  77. 77. Nusselder WJ, Looman CWN, Van Oyen H, Robine JM, Jagger C. Gender differences in health of EU10 and EU15 populations: the double burden of EU10 men. Eur J Ageing [Internet]. 2010 Dec 4;7(4):219–27. Available from: pmid:21212821
  78. 78. Domínguez-Rodríguez A, Blanes Llorens A. El efecto de la salud en la actividad en los mayores de 50 años en España: 2006 y 2014. Cuad Relac Laborales [Internet]. 2019 Jun 10;37(1):177–202. Available from:
  79. 79. Martín U, Domínguez-Rodríguez A, Bacigalupe A. Desigualdades sociales en salud en población mayor: una aportación desde la salud pública al debate sobre el retraso de la edad de jubilación en España. Gac Sanit [Internet]. 2019 Jan;33(1):82–4. Available from: pmid:29295758
  80. 80. Levine ME, Crimmins EM. Is 60 the New 50? Examining Changes in Biological Age Over the Past Two Decades. Demography [Internet]. 2018 Apr 6;55(2):387–402. Available from: pmid:29511995
  81. 81. Oksuzyan A, Juel K, Vaupel JW, Christensen K. Men: good health and high mortality. Sex differences in health and aging. Aging Clin Exp Res [Internet]. 2008 Apr 25;20(2):91–102. Available from: pmid:18431075
  82. 82. Oksuzyan A, Petersen I, Stovring H, Bingley P, Vaupel JW, Christensen K. The Male-Female Health-Survival Paradox: A Survey and Register Study of the Impact of Sex-Specific Selection and Information Bias. Ann Epidemiol. 2009;19(7):504–11. pmid:19457685
  83. 83. Sundberg L, Agahi N, Fritzell J, Fors S. Trends in health expectancies among the oldest old in Sweden, 1992–2011. Eur J Public Health [Internet]. 2016 Dec;26(6):1069–74. Available from: pmid:27175003
  84. 84. Tapia Granados JA. Health at advanced age: Social inequality and other factors potentially impacting longevity in nine high-income countries. Maturitas [Internet]. 2013 Feb;74(2):137–47. Available from: pmid:23276601
  85. 85. Poulain M, Herm A, Pes G. The Blue Zones: areas of exceptional longevity around the world. Vienna Yearb Popul Res [Internet]. 2013;Volume 11(1):87–108. Available from:
  86. 86. Rosero-Bixby L, Dow WH, Rehkopf DH. The Nicoya region of Costa Rica: a high longevity island for elderly males. Vienna Yearb Popul Res [Internet]. 2014;Volume 11(Pittier 1904):109–36. Available from:
  87. 87. Rosero-Bixby L. High life expectancy and reversed socioeconomic gradients of elderly people in Mexico and Costa Rica. Demogr Res [Internet]. 2018 Jan 5;38(1):95–108. Available from: pmid:29576742
  88. 88. Sundberg L, Agahi N, Fritzell J, Fors S. Why is the gender gap in life expectancy decreasing? The impact of age- and cause-specific mortality in Sweden 1997–2014. Int J Public Health [Internet]. 2018 Jul 13;63(6):673–81. Available from: pmid:29654335
  89. 89. Payne CF. Aging in the Americas: Disability-free Life Expectancy Among Adults Aged 65 and Older in the United States, Costa Rica, Mexico, and Puerto Rico. Journals Gerontol Ser B Psychol Sci Soc Sci [Internet]. 2015 Sep 7;73(2):gbv076. Available from:
  90. 90. Yang Yang. Trends in U.S. Adult Chronic Disease Mortality, 1960–1999: Age, Period, and Cohort Variations. Demography [Internet]. 2008;45(2):387–416. Available from: pmid:18613487
  91. 91. Beltrán-Sánchez H, Crimmins EM, Teruel GM, Thomas D. Links Between Childhood and Adult Social Circumstances and Obesity and Hypertension in the Mexican Population. J Aging Health [Internet]. 2011 Oct 23;23(7):1141–65. Available from: pmid:21948773
  92. 92. Montez JK, Hayward MD. Cumulative Childhood Adversity, Educational Attainment, and Active Life Expectancy Among U.S. Adults. Demography [Internet]. 2014 Apr 27;51(2):413–35. Available from: pmid:24281740
  93. 93. Sagardui-Villamor J, Guallar-Castillon P, Garcia-Ferruelo M, Banegas JR, Rodriguez-Artalejo F. Trends in Disability and Disability-Free Life Expectancy Among Elderly People in Spain: 1986–1999. Journals Gerontol Ser A Biol Sci Med Sci [Internet]. 2005 Aug 1;60(8):1028–34. Available from: pmid:16127108
  94. 94. Sofi F, Macchi C, Abbate R, Gensini GF, Casini A. Mediterranean diet and health status: an updated meta-analysis and a proposal for a literature-based adherence score. Public Health Nutr [Internet]. 2014 Dec 29;17(12):2769–82. Available from: pmid:24476641
  95. 95. Muñoz M-A, Fíto M, Marrugat J, Covas M-I, Schröder H. Adherence to the Mediterranean diet is associated with better mental and physical health. Br J Nutr [Internet]. 2008 Dec 15;101(12):1821–7. Available from: pmid:19079848
  96. 96. Arroyo P, Lera L, Sánchez H, Bunout D, Santos JL, Albala C. Indicadores antropométricos, composición corporal y limitaciones funcionales en ancianos. Rev Med Chil [Internet]. 2007 Jul;135(7):846–54. Available from: pmid:17914541
  97. 97. Sevilla F. La universalización de la atención sanitaria. Sistema Nacional de Salud y Seguridad Social [Internet]. Documentos de trabajo (Laboratorio de alternativas). 2006. Available from:
  98. 98. Ministerio de Salud F. Superintendencia de Salud [Internet]. [cited 2019 May 14]. Available from:
  99. 99. Coker TR, Thomas T, Chung PJ. Does Well-Child Care Have a Future in Pediatrics? Pediatrics [Internet]. 2013 Apr 1;131(Supplement):S149–59. Available from: pmid:23547059
  100. 100. Sáenz M del R, Acosta M, Bermúdez JL, Muiser J. The health system of Costa Rica. Salud Publica Mex. 2011;53(SUPPL. 2).
  101. 101. Fernández A. El crecimiento de la oferta de servicios privados de salud y su posible impacto en el sistema universal de salud costarricense. Semanario Univesidad [Internet]. 2018 Nov 7; Available from:
  102. 102. Costa Rica entre los seis países del mundo con mejor asistencia en salud. [Internet]. 2019 Jan 18; Available from:
  103. 103. Organización Panamericana de la Salud. Las desigualdades de salud en Costa Rica: una aproximación geográfico—poblacional. Ser Análisis Situación Salud. 2003;8:37.
  104. 104. Chiu C-T, Hayward M, Saito Y. A Comparison of Educational Differences on Physical Health, Mortality, and Healthy Life Expectancy in Japan and the United States. J Aging Health [Internet]. 2016 Oct;28(7):1256–78. Available from: pmid:27590801
  105. 105. Permanyer I, Spijker J, Blanes A, Renteria E. Longevity and Lifespan Variation by Educational Attainment in Spain: 1960–2015. Demography [Internet]. 2018 Dec 15;55(6):2045–70. Available from: pmid:30324395
  106. 106. Marmot M. The Influence Of Income On Health: Views Of An Epidemiologist. Health Aff [Internet]. 2002 Mar;21(2):31–46. Available from: pmid:11900185
  107. 107. Cachadinha C, Pedro JB, Fialho JC. Social participation of community living older persons: importance, determinants and opportunities [Internet]. London; 2011. (6th International Conference on Inclusive Design ‘The Role of Inclusive Design in Making Social Innovation Happen’). Available from:
  108. 108. Wu J, Li J. The impact of social participation on older people’s death risk: an analysis from CLHLS. China Popul Dev Stud [Internet]. 2018 Jun 25;2(2):173–85. Available from:
  109. 109. Douglas H, Georgiou A, Westbrook J. Social participation as an indicator of successful aging: an overview of concepts and their associations with health. Aust Heal Rev [Internet]. 2017;41(4):455. Available from:
  110. 110. Medellin N, Jara P, Matus-Lopez M. Envejecimiento y atención a la dependencia en Costa Rica. Nota Técnica IDB-TN-1820 [Internet]. 2019;56. Available from:ón_a_la_dependencia_en_Costa_Rica.pdf.
  111. 111. Universidad de Costa Rica (UCR), Consejo Nacional de la Persona Adulta Mayor (CONAPAM). I Informe estado de situación de la persona adulta mayor en Costa Rica. 2008;237 p. Available from:
  112. 112. Pinto JM, Neri AL. Factors related to low social participation in older adults: findings from the Fibra study, Brazil. Cad Saúde Coletiva [Internet]. 2017 Sep 28;25(3):286–93. Available from:
  113. 113. Galenkamp H, Deeg DJH. Increasing social participation of older people: are there different barriers for those in poor health? Introduction to the special section. Eur J Ageing [Internet]. 2016 Jun 18;13(2):87–90. Available from: pmid:28804373
  114. 114. Katagiri K, Kim J-H. Factors determining the social participation of older adults: A comparison between Japan and Korea using EASS 2012. Mogi M, editor. PLoS One [Internet]. 2018 Apr 6;13(4):e0194703. Available from: pmid:29624588
  115. 115. de Graaf R, Bijl R V., Smit F, Ravelli A, Vollebergh WAM. Psychiatric and sociodemographic predictors of attrition in a longitudinal study. Am J Epidemiol [Internet]. 2000;152(11):1039–47. Available from: pmid:11117613
  116. 116. Rothman S. Estimating Attrition Bias in the Year 9 Cohorts of the Longitudinal Surveys of Australia Youth. Longitudinal Surveys of Australian Youth. Technical Report No 48 [Internet]. Australian Council for Educational Research. Victoria; 2009. Available from:
  117. 117. Sierra H, Cordova M, Chen C-SJ, Rajadhyaksha M. Confocal Imaging–Guided Laser Ablation of Basal Cell Carcinomas: An Ex Vivo Study. J Invest Dermatol [Internet]. 2015 Feb;135(2):612–5. Available from: pmid:25178106
  118. 118. Berg CJ van den, Lindeboom M. Attrition in Panel Data and the Estimation of Dynamic Labor Market Models [Internet]. Serie Research Memoranda. Amsterdam; 1994. Report No.: 1994–22. Available from:
  119. 119. Van Oyen H, Nusselder W, Jagger C, Kolip P, Cambois E, Robine J-M. Gender differences in healthy life years within the EU: an exploration of the “health–survival” paradox. Int J Public Health [Internet]. 2013 Feb 22;58(1):143–55. Available from: pmid:22618297
  120. 120. Albala C, Lera L, García C, Arroyo P, Marín P, Bunout D. Searching a Common Definition for Functional Limitation in Latin America. Gerontologist. 2004;44:550.
  121. 121. Vrabcová J, Daňková S, Faltysová K. Healthy Life Years in the Czech Republic: Different data sources, different figures. Demografie [Internet]. 2017;59(4):315–31. Available from:
  122. 122. Malhotra R, Chan A, Ajay S, Ma S, Saito Y. Variation in the Gender Gap in Inactive and Active Life Expectancy by the Definition of Inactivity Among Older Adults. J Aging Health [Internet]. 2016 Oct;28(7):1279–98. Available from: pmid:27590802
  123. 123. European Health Expectancy Monitoring Unit. Interpreting Health Expectancies [Internet]. Ehemu. 2007. Available from:
  124. 124. Berger N, Van Oyen H, Cambois E, Fouweather T, Jagger C, Nusselder W, et al. Assessing the validity of the Global Activity Limitation Indicator in fourteen European countries. BMC Med Res Methodol [Internet]. 2015 Dec 2;15(1):7. Available from: pmid:25555466
  125. 125. Cambois E, Grobon S, Van Oyen H, Robine J-M. Impact of Question Wording on the Measurement of Activity Limitation. J Aging Health [Internet]. 2016 Oct;28(7):1315–38. Available from: pmid:27590804
  126. 126. EUROSTAT. Healthy life years and life expectancy at age 65 by sex. [Internet]. 2019. Available from: Accessed 20/05/2019.
  127. 127. EHLEis. Healthy Life Years (HLY) [Internet]. European Health & Life Expectancy Information System. [cited 2019 Sep 15]. Available from:
  128. 128. European Commission. European Core Health Indicators [Internet]. [cited 2019 Sep 15]. Available from:
  129. 129. EUROSTAT. Healthy life years (from 2004 onwards) (hlth_hlye) [Internet]. 2012. [cited 2017 Jun 15]. Available from:
  130. 130. Mathers C, Ho J. WHO methods for life expectancy and healthy life expectancy [Internet]. Global Health Estimates Technical Paper. 2014. p. 1–24. Available from:
  131. 131. Robine J-M, Cambois E, Nusselder W, Jeune B, Oyen H Van, Jagger C. The joint action on healthy life years (JA: EHLEIS). Arch Public Heal [Internet]. 2013 Dec 4;71(1):2. Available from: pmid:23379576
  132. 132. van Oyen H, Van der Heyden J, Perenboom R, Jagger C. Monitoring population disability: evaluation of a new Global Activity Limitation Indicator (GALI). Soz Praventivmed [Internet]. 2006 Jun;51(3):153–61. Available from: pmid:17191540
  133. 133. Berger N, Robine J-M, Ojima T, Madans J, Van Oyen H. Harmonising summary measures of population health using global survey instruments. J Epidemiol Community Health [Internet]. 2016 Oct 1;70(10):1039–44. Available from: pmid:27165845
  134. 134. Charafeddine R, Berger N, Demarest S, Van Oyen H. Using mortality follow-up of surveys to estimate social inequalities in healthy life years. Popul Health Metr [Internet]. 2014 Dec 12;12(1):13. Available from: pmid:24855457
  135. 135. Vega S, Benito-León J, Bermejo-Pareja F, Medrano MJ, Vega-Valderrama LM, Rodríguez C, et al. Several factors influenced attrition in a population-based elderly cohort: Neurological disorders in Central Spain Study. J Clin Epidemiol [Internet]. 2010 Feb;63(2):215–22. Available from: pmid:19473811
  136. 136. van Groenou MIB, Deeg DJH, Penninx BWJH. Income differentials in functional disability in old age: Relative risks of onset, recovery, decline, attrition and mortality. Aging Clin Exp Res [Internet]. 2003 Apr 25;15(2):174–83. Available from: pmid:12889850
  137. 137. Feng D, Silverstein M, Giarrusso R, McArdle JJ, Bengtson VL. Attrition of Older Adults in Longitudinal Surveys: Detection and Correction of Sample Selection Bias Using Multigenerational Data. Journals Gerontol Ser B [Internet]. 2006 Nov 1;61(6):S323–8. Available from: pmid:17114312