English only

The Population Division of the United Nations Secretariat organized a Working Group on Projecting Old-Age Mortality and its Consequences from 3-5 December 1996 in New York to discuss global issues on the topic of ageing, particularly for the oldest-old (ages 80 and above). The Population Division organized the Workshop in cooperation with the United States Bureau of the Census, the United States National Institute for Aging, and a non-governmental organization active in the field, the American Association of Retired Persons (AARP). Attending the Working Group were some of the most prominent international experts in the area of ageing and mortality from around the world. The proceedings of this meeting are summarized in the detailed report Projecting Old-Age Mortality and Its Consequences (United Nations, 1997). 1/

The purpose of the present note is to highlight those areas of research that are currently needed in the areas of old-age mortality and ageing. Although we cover a wide range of issues, we concentrate on research issues related to the population characteristics of the aged, health and disability, and mortality. The report draws on the conclusions of the Working Group itself, on follow-up correspondence with some of the participants, and on the author's own views regarding research needs and possibilities for progress. The report is general, in that it identifies areas of research that would be appropriate for any organization that is interested in ageing research (universities, research laboratories, international organizations, governmental offices, non-governmental organizations, etc.). It is recommended that a primary role for the United Nations Population Division (UNPD) should be to monitor ageing research in various countries and provide broad overviews of ageing research, with emphasis on international comparison. The Working Group also noted that UNPD has a comparative advantage in addressing some particular issues, which are highlighted below.

There is general consensus that much of our knowledge about ageing processes comes from research efforts in developed countries, and even then, our knowledge is often limited by a lack of comprehensive data. From the perspective that ageing can be considered a movement through a series of life events, more longitudinal work that follows cohorts through transitions over time is especially required. General recommendations follow:

• highest priority should be given to providing better, in-depth exploitation of all available census and survey data to better understand the technical demography and the social and economic correlates of ageing wherever such data are available;

• because information on levels, trends, and differentials in mortality, causes of death and morbidity, disability patterns, service usage, family care-giving, retirement schemes, etc. is necessary for the formulation, implementation, and evaluation of health and social development policies, there is a need for new and strengthened systems of data collection and processing. International efforts should seek to promote those national data collection and research activities that are conducive to attainment of internationally comparable information and knowledge. Such an approach should be advocated world-wide, but is especially required for developing countries;

• international efforts should seek to develop and promote standardized definitions (of disability, types of service providers, retirement schemes, etc.) for international comparability;

• while all types of data collection strategies should be encouraged for studying issues related to ageing, preference should be given to nationally-representative, micro-level longitudinal sample survey data. Relevant organizations of the United Nations system and other intergovernmental bodies could play a catalytic role in shaping and implementing plans for work of this type, though national and regional organizations could initiate, promote, and carry out the work.

The following sections highlight research needs in particular topical categories. There is a critical need to explore relationships between these groups of variables, particularly from policy and programme perspectives. While research is needed in a wide variety of contexts, particular emphasis should be placed on understanding the dramatic demographic and socioeconomic changes that are occurring in many developing countries. As noted by the United States National Research Council (Martin and Preston, eds., 1994, p. 410), researchers " ... are provided with a unique opportunity to study how societies, institutions, and individuals adapt to ageing populations. Study of these processes can provide insight into the influences of culture and ethnicity, the changing role of family support, particular effects in low-income environments, and the consequences of new policies and programs."

Population estimates and projections for the oldest age groups

The United Nations biennially publishes estimates and projections of population parameters for all countries of the world in its World Population Prospects series. In recent revisions of the population estimates (1996 and earlier), distributions of population and mortality were provided in five-year age increments up to age 80 and for the open-ended, aggregated interval of age 80 and above. The international community does not currently have consistent estimates of the number of people in the oldest age groups (in five-year increments) for all countries of the world. The Working Group supported the United Nations' desire to provide detailed data for age groups 80-84, 85-89, 90-94, 95-99, and 100 and above beginning in the 1998 Revision, agreeing that there is an important need to prepare estimates of the elderly population up to age group 100 and above for all countries using a consistent methodology which would allow international comparisons.

Technical issues concerning estimates and projections of the old-age population were a major focus of the Working Group meeting. Few national life tables can support calculation of survival ratios up to age 100 due to the deterioration of data quality at the older ages. For the same reason, the age distribution of older persons reported in censuses usually cannot be accepted at face value, and in some cases, available tabulations do not provide detail at the higher ages. Thus, for most countries there is a need to employ indirect and model-based estimation techniques to estimate the age distribution of older persons up to age 100, and to supplement the available information about trends (United Nations, 1996a). The methodological and data-quality issues involved are discussed in detail in the report on the meeting (United Nations, 1997). The report also contains recommendations and conclusions on these matters. There was a consensus that, despite the technical difficulties to be overcome, extension of the estimates and projections was both feasible and desirable. The United Nations Population Division is accordingly planning to extend the 5-year age detail in its estimates and projections up to age 100, beginning with the next Revision, due to be completed in 1998. The United States Bureau of the Census is planning a similar extension of its estimates and projections.

There is also a need for better knowledge of recent trends in old-age mortality from which to make better assumptions for future population projections. As noted by the United States National Research Council, "There is considerable debate about the future course of old-age mortality and life expectancy. Improved sex-, race-, and cause-specific data on deaths at older ages are essential for refining mortality and population projections, which can help gauge the future demand that will be placed on Social Security and other programs that support the elderly" (Martin and Preston, eds., 1994, p. 407).

Health and longevity

Forecasts. Demographers will continue to make major contributions to the understanding of ageing and related issues, through forecasts of adult mortality and morbidity and their socio-economic consequences. However, such studies should not be limited to simple statistical extrapolation of trends in mortality and morbidity, as in the past, but should incorporate specific biomedical factors such as progress in prevention and treatment of cancers, emergence and re-emergence of infective diseases, and impacts of smoking in developing countries. For developing such forecasts, quantitative studies on relationships among ageing, disease, and mortality are very important.

Despite the considerable improvements in mortality over the past decades, it is difficult to predict the course of mortality change that will evolve in the future in both developing and developed countries. As noted by Horiuchi (United Nations, 1997), a number of issues deserve special attention:

• New and re-emerging diseases. Research to evaluate the potential impact on mortality of a broad category of "new" infectious diseases, such as HIV/AIDS, Ebola, Lyme disease, etc., which could spread before effective preventive and/or treatment interventions are discovered. 2/ Similarly, the re-emergence of "old" infectious diseases that are increasingly drug- (and especially antibiotic-) resistant, including malaria and tuberculosis, deserves special attention.

• Smoking. Two centuries of mortality data from Sweden have shown that the spread of smoking has had a significant impact on the history of mortality, especially for males. It remains to be investigated whether changes in smoking behaviours will impact on mortality patterns in the less developed regions.

• Mortality and rapid socio-economic change. Recent examples demonstrate that mortality patterns can change quickly in a country for better or for worse. For instance, experience from the Newly Industrialized Economies has shown that rapid economic development is associated with rapid improvements in life expectancy at birth. In contrast, the countries of the former USSR have demonstrated that considerable increases in mortality can follow rapid economic and political changes. Further research is required to ascertain the potentials for mortality reduction in situations of social disintegration and apathy, in urban slums, where there is excessive alcohol consumption and in areas affected by environmental degradation.

Longevity. A second broad category of promising, needed research is in the study of the genetic, environmental, and medical determinants of longevity. Fundamental, yet unanswered, questions include: "why do some people die at age 60, others at 80, and a few at 100? Why do some species live a few weeks and others scores of years? How important are genetic factors in determining how long an individual will live? How can longevity be extended by a few genetic or environmental manipulations?" (Christensen and Vaupel, 1996).

Cross-national research on disability, service use, and expenditures

Standard classifications. International research would be greatly facilitated by the ability to compare data on disabilities, service use and expenditures (health, long-term care, etc.) from different countries. At present, this is very difficult. An international effort to standardize definitions of disability, types of services, and service providers could contribute, over time, to the ability to do comparative research.

Mortality reduction and quality of life. More policy-related research is needed in order to better understand the relationship between mortality reduction and quality-of-life for the elderly. With mortality reduction, are individuals exposed to longer periods of vulnerability to disease and to prolongation of periods of disability and poor health? Can extension of life be achieved only through interventions that make the quality of life less enjoyable, such as through reductions in caloric intake or through introduction of particular stresses? Related to this topic is research on active life expectancy. How is it changing over time? How does a year of added life expectancy at, say age 90, compare to a year added at age 75?

Geographic concentration of older persons and of services. Related to the issue of service use is the need for a better understanding of geographic variations in the concentrations of the elderly, particularly for the oldest-old, and of geographic distribution of the demand for and cost of services provided through state and local governments (Martin and Preston, eds., 1994). A better understanding of the determinants of concentration is also required. Does the proportion elderly in particular locations reflect migration patterns to that destination or did the population "age in place"?

Intensity of the ageing process

Implications of rapid population ageing. Demographers have long understood that, in countries with a high fertility rate, the growth rate will not move to zero at once, even if replacement level fertility is achieved. A built-in "momentum" for population increase exists because the large cohorts born at the time of high fertility will still have more children than earlier smaller cohorts. What are the implications of the reverse, i.e. rapid ageing, for populations? Will the proportion young fall so quickly and dramatically in some countries, relative to the rapid increase in the proportion elderly, as to make it difficult for the social and economic structures to adjust? Can the ageing process become so intense, with the proportion elderly reaching 50-60 per cent of the population, implying a large excess of deaths over births, and dramatic reductions in fertility, that entire sub-populations would disappear?

Low fertility and population ageing. Further research is required to ascertain whether it is possible to implement policies to recover fertility in very low fertility countries, and thus to reduce the speed and intensity of the ageing process.

Interactions with society, including care systems and living arrangements

Family care-giving. How do families balance the trade-off between time and money in providing care for their elderly family members?

Care-giving and women's labour force participation. Further investigation is required of the impact of the participation of women in the labour force on care-giving patterns internationally, particularly in those contexts where there has been a rapid change in the percentage of women who have become involved in paid employment outside the home.

Policy impacts on care-giving. Evidence of how governmental and private health care reimbursement policies impact on formal and informal care systems exists for many developed countries, though the full complexities are not yet understood. Further research is also required to evaluate whether, and how, the experience from developed countries can be adopted by countries which are just beginning work in this topic area.

Service needs and social support for the disabled. What are the factors that affect institutionalization and use of services? For continued home living arrangements, which adaptations and programs are needed for the elderly with osteoarthritis and other causes of impaired mobility? What are the physical characteristics of housing units and social support systems for persons with dementia or other cognitive impairments? How do the disabled and their families fare in countries that lack a system of outside-family social support services or a formalized institutional structure?

Work and retirement

Data for understanding retirement choices and their consequences. In countries where formal retirement structures are in place, the greatest need is for more longitudinal, behavioural research in order to understanding the economic and social antecedents and consequences of retirement. The Health and Retirement Survey (HRS) and the Asset and Health Dynamics (AHEAD) Survey provide excellent data on work issues in the United States, and allow multi-period analysis of the behaviour and circumstances of older Americans as they approach retirement. There is a need for HRS-type and AHEAD-type international data sets from different policy environments to better understand cross-national differences in work and retirement decisions among the elderly. Types of issues include:

• Labour force participation. To what extent do policy changes in public (social security systems) and private (retirement plan) benefits structures affect the decision to retire from active labour force participation for employees in pre-retirement ages? To what extent do employees at retirement age continue to work because of a concern that their retirement benefits or health insurance would be insufficient should they retire? What are the determinants of continuing to work part-time after formal retirement? How many people in retirement ages are willing and able to work but are not currently employed?

• Economic security (income and assets of older people). What is the relationship between income and life expectancy at different ages? Do the poor in developed countries collect relatively less from social security systems because they do not live as long? In the case of financial assistance, how are intergenerational transfers related to income and wealth? What are the motivations for transfers, and how are lifetime transfers related to bequests?

* This report was prepared by Ms. Laura Shrestha, Rapporteur, in consultation with members of the Working Group on Projecting Old-Age Mortality and Its Consequences, New York, 3-5 December 1996. The Rapporteur would like to thank the Working Group participants, especially Mr. Eduardo Arriaga, Mr. Shiro Horiuchi and Mr. Miroslav Macura; and Ms. Charlotte Nusbert, Ms. Jane Tilly and Dr. Sara Rix of the American Association of Retired Persons for many helpful suggestions and insights for this report.

2/Recent estimates and projections by the United Nations Population Division (United Nations, 1996b) as well as the United States Bureau of the Census (1996) have included explicit estimates of the demographic impact of HIV/AIDS.

Christensen, Kaare and James W. Vaupel (1996). Determinants of longevity: genetic, environmental and medical factors. Journal of Internal Medicine, vol. 240, pp. 333-241.

Martin, Linda G. and Samuel H. Preston, eds. (1994). Demography of Aging (Washington, D.C., National Academy Press).

United Nations (1996a). Modeling old-age mortality for projection purposes. Paper drafted by Mr. Patrick Heuveline (consultant), University of Pennsylvania, for the Working Group on Projecting Old-Age Mortality and Its Consequences, New York, 3-5 December 1996.

United Nations (1996b). World Population Prospects: The 1996 Revision (United Nations publication, forthcoming).

United Nations (1997). Projecting Old-Age Mortality and Its Consequences, Report on the Working Group, New York, 3-5 December 1996 (New York, Population Division).

United States Bureau of the Census (1996). World Population Profile: 1996 (United States Bureau of the Census, Washington, D. C.).