Objectives: To project health outcomes for the period 2011-2031, in the Portuguese population aged 65 years and over, considering the expected changes in its demographic and educational structure. Methods: Demographic projections were produced using the multistate cohort component method. Logistic regression models were fitted to estimate current differences in health outcomes, using data from the Portuguese Health Survey. Such differences were applied to the projection results, in order to project the prevalence of “poor” self-reported health status, functional limitations and disability. Results: Our results point to a future improvement in the considered health outcomes. The prevalence of “poor” self-reported health status, functional limitations and disability will decrease in both genders, but will continue to be higher among women than men. Discussion: Beyond age and gender, other determinants of health, like education, should be considered when studying the possible evolution of health outcomes in an ageing population.
The recent demographic development of the Portuguese population is depicted by a marked change in its age structure, with a clear demographic aging resulting both from decreased levels of fertility and progressive decrease in the mortality levels. The contribution of migration for population growth and renewal has substantially decreased, being now insufficient to ensure an increase in the number of residents in Portugal and to counteract the aging trend, given the diminishing natural increase (Statistics Portugal, Demographic Indicators). As a consequence, the population is increasingly aging: between 1992 and 2011, the number of people aged 65 years and more increased approximately 42% and, at the end of this period, accounted for 19.1% of the total population, a value that is clearly above the EU-27 average (17.5%) and which is only exceeded by Germany (20.6%), Italy (20.3%) and Greece (19.3%) [
Demographic ageing is a major challenge with regard to governance decisions, given its inevitable impact on the public policies sector, namely in relation to social support, health care and the labor market [
Among different structural, behavioural and psychosocial factors, age has often been associated to health status. Results from a study about the evolution of self-rated health status in the elderly, by Orfila, F., Ferrer, M., Lamarca, R. and Alonso, J. [
Considering the combined influence of gender, age and education on population health, the prospective study of the demographic structure defined by these dimensions may represent an important contribution in order to adjust the health policies and practices to the promotion of a better health status. If, on the one hand, older individuals tend to present a worse health status, compared to the younger, and, on the other hand, those with higher levels of education tend to have a better health status, relatively to the less educated [
The objective of this study is to estimate four indicators on health outcomes for the Portuguese population aged 65 years and over, considering the expected changes in population structure by gender, age group and educational attainment. We produced multistate demographic projections for the period 2011-2031 and studied the adjusted impact of education and age on the health outcomes, focusing on self-reported health status and the prevalence of disabilities, using data from the most recent National Health Survey (2005/2006). The following questions were addressed: 1) how will the demographic and educational structure of the Portuguese men and women above 65 years old evolve until 2031? 2) to what extent are there educational differences in self-re- ported health status, functional limitations and disability within this age group? and 3) what will be the combined effect of the expected changes in the demographic and educational structure of this group, on the prevalence of “poor” self-reported health status, functional limitations and disability?
For producing the demographic projections by gender, age group and educational attainment, data from 2001 and 2011 Census, national projections produced by Statistics Portugal (Instituto Nacional de Estatística (INE), I.P.), the Labour Force Survey (INE, I.P.) and the 2010 Revision of the World Population Prospects (United Nations) were all used, in order to estimate the necessary parameters.
The study of current differences in health status and health services utilization by age and educational level was based on the data from the fourth edition of the National Health Survey (4th NHS), a representative survey of the Portuguese population, planned and conducted by the Statistics Portugal and the National Institute of Health Dr. Ricardo Jorge (INSA), in collaboration with the Directorate-General of Health. The 4th NHS collected information on health status, health determinants, use of health services and sociodemographic characteristics of individuals, being the first edition to cover the entire national territory, including the Autonomous Regions of Azores and Madeira. The questionnaire was administered by direct interview to a representative probability sample of the Portuguese population, between February 2005 and February 2006. The study population included individuals living in family housing units (therefore excluding people living in collective accommodation). Housing unit is a unit of observation commonly used in Census. It is defined as a residence such as a house, apartment, mobile home, or room(s) within a larger structure that provides a space for occupants making up a single household to live and eat. By collective accommodation we mean hotels and similar accommodations and people living together in a unit that is not a family (people institutionalized for example). As a whole, 41,193 individuals living in 15,239 family housing units were interviewed, and the interview completion rate was 76% nationwide.
Educational attainment level—The data concerning each individual’s highest level of education were categorized into four groups: 1) No education/Primary Education (second cycle of basic education, corresponding to six years of schooling, or less); 2) Lower Secondary Education (nine years of schooling); 3) Upper Secondary Education (twelve years of schooling); and 4) Higher Education (those who achieved, at least, a degree of Bachelor).
Self-reported health status—The 4th NHS contained the following question about self-perceived health: “In general, how do you consider your health status?”. Based on self-reported health status we construct two new categories: good health (very good or good self-reported health status) and poor health (fair, bad or very bad self-reported health status).
Functional limitations and disability—Respondents of the 4th NHS were asked if they have any difficulty or a complete inability in performing a set of tasks. From these questions, three groups of limitations were considered: 1) functional limitations, 2) disabilities in activities of daily living (ADL) and 3) disabilities in instrumental ADL. In the first group, we included the questions related to mobility and to specific physical functional limitations. These included being bedridden or being unable to: go outside the home, walk at least 200 meters or being wheelchair-bound, walk up and down a twelve-step staircase without resting, get into and out of bed, sitting down and getting up from a chair, bend down and lift an object from the floor or grab a small object, alone and without difficulty. In this group, we also considered being unable to listen to a TV or radio show in a reasonable volume, being unable to see in order to recognize a friend at 4 meters of distance or having difficulties with speaking. In the second group, we included the disabilities that show dependence of the individual on others, with need for assistance in daily life. These included being unable to go to the bathroom and use it, get dress and undressed, have bath, wash hands and face or feed him/herself alone, without difficulty. The third group included the disabilities affecting a broader range of activities: being unable to use public transport, to go shopping, to do the housekeeping or to prepare food alone, without difficulty. Each group was considered in the form of a binary variable (1 if at least one of the limitations is present and 0, otherwise).
Additional variables—We worked with individuals aged 45 years and over in the starting year for the demographic projections (2011), since these are the ones with 65 years and more in 2031. Age was categorized into five-year groups until 75+ years.
A multistate demographic model [
To study the differences in self-reported health status, functional limitations and disability by age and educational level, logistic regression models [
The expected proportions of individuals in each category of the dependent variables, for each gender, age group and educational level, were applied to the estimated number of individuals in each of the strata defined by gender, age group and educational level, obtained from the demographic projections previously described.
Portuguese population (65+ years) by gender, age group and educational attainment, 2011-2031
The results from the demographic projections suggest an increase in the total number of residents in Portugal aged 65 years and over, rising from 1,931,457 inhabitants in 2011, to 2,226,497, by 2031 (
In
Self-reported health status, functional limitations and disability, 2011-2031
The results from the logistic regression models based on data from the 4th NHS (
These results also suggest that the effect of education (adjusted for age) among individuals with 65 years and over is observed in all the considered health outcomes and in both genders, with individuals without Lower Secondary education showing a statistically higher risk of having poor self-perceived health status (OR = 2.33; p = 0.00 for men and OR = 3.08; p = 0.00 for women), at least one functional limitation (OR = 1.78; p = 0.01 for men and OR = 1.67; p = 0.02 for women), one disability in ADL (OR = 2.21; p = 0.03 for men and OR = 2.71; p = 0.01 for women) or one disability in instrumental ADL (OR = 2.09; p = 0.02 for men and OR = 1.98; p = 0.01 for women), compared with the whole population of this age group. Also, individuals with Higher education show a lower risk of reporting their health status as poor (OR = 0.58; p = 0.001 for men and OR = 0.56; p = 0.002 for women).
Considering the projected prevalence for each of the indicators, and for both genders, the results suggest a clear improvement in all health outcomes (
This improvement tends to be slightly larger for women than for men. Nevertheless, the proportion of individuals who negatively classify their health status, who have some functional limitation and who have some disability will continue to be considerably higher among women, compared to the male residents.
. Population structure by gender, age group and level of education, 2011 and 2031, individuals aged 65+ years
Year | |||
---|---|---|---|
2011 | 2031 | ||
Total (N˚) | 1,931,457 | 2,226,497 | |
Gender | M | 41.7% | 42.0% |
F | 58.3% | 58.0% | |
Age Group (years) | 65 - 69 | 27.5% | 31.0% |
70 - 74 | 25.1% | 26.1% | |
75+ | 47.4% | 42.9% | |
Level of Education | N.E./Primary | 86.5% | 59.7% |
Lower Sec. | 5.4% | 15.1% | |
Upper Sec. | 3.6% | 12.2% | |
Higher | 4.6% | 13.04% |
Source: Authors’ calculations.
. Adjusted odds ratios from logistic regressions
Gender | Age group/Level of education | Poor health statusa | Functional limitations (≥1)b | Disabilities in ADL (≥1)b | Disabilities in instrumental ADL (≥1)b | ||||
---|---|---|---|---|---|---|---|---|---|
OR | p | OR | p | OR | p | OR | p | ||
Male | 65 - 69 years | 0.86** | 0.01 | 0.84** | 0.02 | 0.68** | 0.00 | 0.62** | 0.00 |
70 - 74 years | 0.96 | 0.54 | 0.99 | 0.88 | 0.86 | 0.24 | 1.09 | 0.46 | |
75+ years | 1.22** | 0.003 | 1.20** | 0.01 | 1.72** | 0.00 | 1.49** | 0.00 | |
NE/Primary | 2.33** | 0.00 | 1.78** | 0.01 | 2.21** | 0.03 | 2.09** | 0.02 | |
Lower Sec. | 0.83 | 0.21 | 0.97 | 0.89 | 0.62 | 0.32 | 0.67 | 0.35 | |
Upper Sec. | 0.88 | 0.47 | 0.79 | 0.46 | 0.62 | 0.41 | 1.02 | 0.97 | |
Higher | 0.58** | 0.001 | 0.74 | 0.35 | 1.18 | 0.71 | 0.702 | 0.46 | |
Female | 65 - 69 years | 0.86** | 0.01 | 0.84** | 0.002 | 0.81** | 0.00 | 0.79** | 0.00 |
70 - 74 years | 0.94 | 0.39 | 0.98 | 0.79 | 0.85* | 0.06 | 0.995 | 0.95 | |
75+ years | 1.23** | 0.003 | 1.21** | 0.001 | 1.45** | 0.00 | 1.28** | 0.00 | |
NE/Primary | 3.08** | 0.00 | 1.67** | 0.02 | 2.71** | 0.01 | 1.98** | 0.01 | |
Lower Sec. | 0.67** | 0.02 | 1.08 | 0.79 | 0.52 | 0.26 | 0.59 | 0.21 | |
Upper Sec. | 0.87 | 0.51 | 0.93 | 0.82 | 0.76 | 0.65 | 1.64 | 0.17 | |
Higher | 0.56** | 0.002 | 0.601 | 0.14 | 0.93 | 0.89 | 0.52 | 0.17 |
an = 6929; bn = 6930. **p ≤ 0.05; *p ≤ 0.10. Source: Authors’ calculations.
. Projected prevalence for each health indicator, by gender, 2011-2031, Portugal
Gender | 2011 | 2021 | 2031 | |
---|---|---|---|---|
Poor health status | M | 78.9% | 76.0% | 73.1% |
F | 87.7% | 84.9% | 81.0% | |
Functional limitations (≥1) | M | 13.9% | 12.9% | 11.8% |
F | 16.6% | 15.7% | 14.4% | |
Disabilities in ADL (≥1) | M | 6.8% | 6.1% | 5.4% |
F | 9.3% | 8.5% | 7.4% | |
Disabilities in instrumental ADL (≥1) | M | 7.7% | 6.9% | 6.2% |
F | 11.7% | 10.9% | 9.9% |
Source: Authors’ calculations.
Number of people aged 65+, by level of education, 2011-2031
Projected prevalence for each health indicator, men (left) and women (right), 2011-2031, Portugal
This paper focuses on challenges and opportunities related to projecting health outcomes for Portuguese ageing population. Demographic projections were produced using the multistate cohort component method. Logistic regression models were fitted to estimate current differences in health outcomes, using data from the Portuguese Health Survey. Such differences were applied to the projection results, in order to project the prevalence of “poor” self-reported health status, functional limitations and disability. It has been observed that the prevalence of “poor” self-reported health status, functional limitations and disability will decrease in both genders, but will continue to be higher among women than among men. It has been shown that beyond age and gender, other determinants of health, like education, should be considered when studying the possible evolution of health outcomes in an ageing population.
Albeit its relevance [
To predict health outcome trends in an ageing society, as it is the case for Portugal, it is crucial to know population projections by sex, age and educational level. Projections of health outcomes for the Portuguese population aged 65 years and over, by gender, were produced. Demographic projections by gender, age group and level of education, calculated using the multistate projection model, served as the basis for projecting the expected changes on the population self-perceived health status and the prevalence of functional limitations, disabilities in ADL and disabilities in instrumental ADL. In these demographic projections, the differentials regarding fertility, mortality and migration, by level of education, were taken into account. We found that the self-reported health status is the indicator that has the more obvious expected improvement, with a clear decrease in the proportion of residents who rate their health as poor for both genders. The prevalence of functional limitations and disabilities will also decrease, and will continue to be substantially higher among women, compared to men. This pattern of gender differences in health outcomes has also been found in other studies [
Our results point to an important idea about the consequences of population ageing: the extent to which changes in key lifestyle factors, such as education, influence health during the life course and particularly during old age, may have a counterbalancing effect upon the greater vulnerability to disability and morbidity associated with the increase in age.
We may assume that, if increased longevity is accompanied by an increase in the number of years lived in good health, aging may not necessarily translate into increased health expenditure. A better health status will, in principle, be reflected in a lower use of health services and can consequently lead to lower health expenditure [
Projecting the health status of a population is a complex and challenging task, given the difficulties associated with the estimation of changes in morbidity and the measurement of health [
When dealing with a variable such as education, we have to take into account that the current changes in the schooling rates of the population will only have an influence many years later: the educational structure of people who are expected to reach 65 years and over within two decades is the result of the schooling rates from about 35 to 15 years ago. Considering this, we should try to look into the impact that the present developments in the Education sector might have in a longer term. Until now, we can expect the proportion of people with higher education levels to continue to rise, as the education of younger cohorts seems to evolve positively, but it is still not well understood if this positive development will continue, giving the economic situation that Portugal is currently facing. Moreover, we will expect the better educational levels of the elderly to have an important impact on functional fitness outcomes, enhancing the quality of life of these population, namely within the context of the community physical activity programs that has been recently implemented in the Portuguese society [
This work was co-funded by FEDER funds through the Operational Programme for Competitiveness Factors— COMPETE and by National funds through FCT—Fundação para a Ciência e a Tecnologia as part of the project Aging and Health in Portugal: Policies and Practices, FCOMP-01-0124-FEDER-PTDC/CS-DEM/109967/2009.