| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||||||||||||||||
COMMENTARY |
Laboratory of Epidemiology, Demography, and Biometry, National Institute on Aging, Bethesda, Maryland.
Address correspondence to Jack M. Guralnik, MD, PhD, National Institute on Aging, 7201 Wisconsin Ave., Rm. 3C-309, Bethesda, MD 20815. E-mail: jg48s{at}nih.gov
Robine and Michel have written a provocative piece on the relationships between longevity, morbidity, and disability, how these relationships affect the number of years that older people live with disability, and how this may change over time (1). Furthermore, they propose that there may be a circling back, where, first, sicker people survive into old age and disability rises, then the number of years lived with disability decreases as new cohorts of healthier people enter old age, but, finally, the number of years lived with disability rises again when the average age of death goes so high that many people spend their last years at advanced old age burdened by multiple chronic diseases and frailty. And as if all of this were not complex enough, Robine and Michel propose that it is happening at different times in different countries and perhaps even at different times in the same country within different population subgroups. Particularly provocative and worthy of serious consideration is their proposal that all these changes, both expansion and compression of morbidity, are part of a single unifying process, a "general theory on population aging," and are simply different stages of a single transition. At this point, while they do have a certain amount of supporting evidence for their proposal, it remains within the realm of theory and requires a great deal more empirical evidence. In both recent and future research, some degree of skepticism is warranted in assessing both the validity of the data and its interpretation in regard to changes in population morbidity and disability.
The challenges are considerable in collecting valid evidence to address Robine and Michel's theory, particularly when attempting to assess disability trends across time and among different countries and cultures. There have been questions for many years about the accuracy of older people's self-report of disability and, particularly, whether the kinds of questions we use to assess disability mean the same things to people in different cultures and demographic subgroups. Some quite interesting recent work has used objective physical performance assessment to define the background or underlying level of functional status and, considering this as a latent scale, modeling the cut-points in this scale at which people in different subgroups make the transition from reporting one level of difficulty to another. Analyses from the World Health Organization that used data from the U.S. National Health and Nutrition Examination Survey (NHANES) showed that men, higher income persons, and nonwhites have lower cut-points for reporting transitions (2). This means that they must reach lower levels of performance than women, low-income people, and whites before their self-report transitions from no difficulty to some difficulty, some difficulty to much difficulty, and much difficulty to unable. Remarkably, this same pattern of differences was present for physician-assessed disability. Further work from Melzer and colleagues using this analytic technique shows that the lower levels of self-reported disability in a Dutch cohort compared to the NHANES was due in part to the Dutch reporting disability only when they reach lower levels of performance (3). These analyses provide some of the most compelling evidence to date that we must be quite cautious in making comparisons of self-report across population subgroups. Furthermore, as we follow populations over time, improvements in socioeconomic status may have an impact not only on real change in disability but on how people report their disabilities.
Despite the methodological difficulties in assessing disability, there is a fairly solid body of evidence that disability rates have fallen in recent years in a number of countries. Considering the kind of issue discussed above, it would seem that self-reports that assess more basic functional tasks, what have been termed functional limitations (4), might provide the best evidence, and here the work of Freedman and Martin, using items such as stair-climbing, lifting, and walking, is compelling (5). However, there are potential pitfalls in addressing some of the issues in Robine and Michel's theory on population aging using only disability prevalence data. A compression of morbidity requires that the number of person-years spent in a disabled state declines in the older population as a whole. Even in the face of reduced age-specific disability prevalence [Freedman demonstrated that prevalence fell even in the population aged 80 and older (5)], it is still possible that, in the total older population, more years will be spent disabled. This results from a shift in the age structure of the older population (6) so that many more people are alive at the oldest ages, where disability rates, even if they have come down somewhat, are still much higher than for younger populations.
In the original proposal on compression of morbidity by Fries (7), he predicted that there would soon be a rectangularization of the survival curve and a halt in the rise of life expectancy. Life expectancy is rising now, however, and is expected to continue to rise, but a compression of morbidity is still possible if the postponement of disability outpaces the postponement of death. Recent research has provided evidence that lifestyle improvements are associated with less disability as death approaches, supporting a compression of morbidity (8–10). However, these reports also must be interpreted cautiously due to the age structure of the populations studied. For example, in Vita's analyses of persons who attended the University of Pennsylvania in 1939 and 1940 and were followed through 1994, those with better health habits (low risk) had substantially less cumulative disability and less disability before they died (8). However, the average age at the end of the follow-up period was 75 years and was lower among people who died. The study, therefore, can give us no information on the low-risk participants' cumulative disability when they lived to very old ages.
It is probably unwise to extrapolate findings on disability prior to death at age 75 to disability prior to death at age 95. Deaths in the 80s and 90s, especially deaths in women, are the ones in which very long survival may be accompanied by years of disability from conditions such as arthritis, osteoporosis, and dementia, disability that would have been absent or much reduced had these people not survived so long. This, indeed, is the final stage of the theory of population aging proposed by Robine and Michel. In order to document whether this last stage is in fact occurring, it will be necessary to develop a new epidemiology of extreme longevity. The challenge in doing this is that there is often a large under-representation of sick and disabled people in population-based studies of the very old population. Validly tracking the full impact of disability at the end of life in people of this age requires identification of participants and follow-up in many different living situations and care settings, home visits, and use of proxies to increase response rates and more frequent assessments than are necessary in younger populations.
Ultimately, applying a theory of aging such as that proposed by Robine and Michel should help us to understand the public health impact of having a population that is dying at older and older ages, and it is useful to consider just what our measures of success should be. A century ago, less than 4% of 50-year-olds could expect to live to age 90, but now, over a quarter of women who are age 50 can expect to live to age 90 and beyond (6). In a study using data on an older population with no upper age limit, it was found that higher education blacks and whites had a longer life expectancy and longer active life expectancy than lower education persons (11). However, greater life expectancy was a result of spending longer time in the disabled state as well as in the nondisabled state. Thus, if future populations look more like our current high education population, they would live longer and have higher active life expectancy but their total number of person-years of disability could be higher.
It is clearly a measure of success if many more people survive to advanced old age, if active life expectancy is longer, and if the percentage of life spent disabled is lower. But all of this may be at the expense of extra person-years of disability, years for which a very long-lived population will require support and medical care. This is important from a public health standpoint, but perhaps even more important is a point that is often missed in discussions about the fall in disability rates and compression of morbidity. That point is that, as the population expands and the sheer number of people who are over age 80 drastically increases, the absolute number of people with disability will rise substantially. Even if we encounter lower rates of disability, and if overall there is a compression of morbidity, the total number of disabled years experienced by this increasingly large and aged population could grow dramatically. The true measure of success is whether the care that is available to them can reduce the impact of chronic disease and improve quality of life in the ninth and tenth decades.
References
This article has been cited by other articles:
 |
|
 |
|
  K. Christensen, M. McGue, I. Petersen, B. Jeune, and J. W. Vaupel Exceptional longevity does not result in excessive levels of disability PNAS, September 9, 2008; 105(36): 13274 - 13279. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
