| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||||||||||||||||||
RESEARCH ARTICLE |
a Epidemiology, Demography, and Biometry Office, National Institute on Aging, Bethesda, Maryland
David Melzer, Institute of Public Health, Forvie Site, Robinson Way, Cambridge CB2 2SR, United Kingdom E-mail: dm214{at}medschl.cam.ac.uk.
 |
Abstract |
|---|
 TOP Abstract MethodsResultsDiscussionReferences |
|---|
Methods. Data were from 3 sites of the Established Populations for Epidemiological Study of the Elderly, covering 8,871 people aged 65–84 years who were followed for up to 7 years. Participants were classified on years of education received and as disabled if they needed help or were unable to walk up or down stairs or walk half a mile. A Markov model computed relative risks, adjusting for the effects of repeated observations on the same individuals.
Results. Differences between education groups in person-years lived with disability were large. The relative risk of incident disability in men with 0–7 years of education (vs. those with 12 or more years) was 1.65 (95% CI = 1.37–1.97) and in women was 1.70 (95% CI = 1.15–2.53). Both recovery risks and risks of death in those with disability were not significantly different across education groups in either gender.
Discussion. Higher incidence of disability is the main contributor to the substantially higher prevalence of disability in older people of lower socioeconomic status. Efforts to reduce the disparity in disability rates by socioeconomic status in old age should focus mainly on preventing disability, because differences in the course of mobility disability after onset appear to play a limited role in the observed prevalence disparities.
DIFFERENCES in mortality and the prevalence of functional disability in old age across groups with different levels of education are well established, with higher rates reported in less educated subgroups in virtually every study (Guralnik, Fried, and Salive 1996
). Reducing such socioeconomic health disparities is one of the central goals of the Healthy People 2010 initiative (U.S. Department of Health and Human Services 2000).
Epidemiological analyses of socioeconomic differences in health are broadly based on sociological notions that society is stratified into classes or groups, and these groups have different material circumstances; exposures; behaviors; and psychosocial, political, and other experiences (Lynch and Kaplan 2000). The traditional individual measures of education, income, and occupation are seen as indicators of the social and economic factors that dominate the social structure. There is now a large body of evidence based on these markers showing clear trends of poorer health with each step down the hierarchy of social position (Marmot and Wilkinson 1999).
The higher mortality and prevalence of most diseases in less privileged socioeconomic groups have been attributed to a variety of factors (Adler and Ostrove 1999). Early life experiences (including maternal malnutrition) have been implicated in the risk of developing chronic disease in adulthood (Barker and Martyn 1992). Social position may directly result in greater exposure to injury or toxic compounds. Adverse health behaviors, such as cigarette smoking, alcohol abuse, and sedentary lifestyle, may contribute a proportion of observed socioeconomic differences in mortality (Lantz et al. 1998). Psychosocial factors, including stress, appear important in explaining some of the risk not explained by traditional factors (Baum, Garofalo, and Yali 1999; Landsbergis, Schnall, Warren, Pickering, and Schwartz 1999). In addition, differences in access to health services by socioeconomic status have been well documented (Andrulis 1998). Finally, reverse causation, with poorer health causing a fall in social position, may play a limited role (Bartley and Plewis 1997).
Older people often suffer from more than one disease, and disability, usually defined as the inability to carry out the usual tasks of daily life, is established as a powerful measure of health status in old age (Guralnik et al. 1996). Physical disability is linked to acute illness, chronic disease, and injury. Single diseases making important contributions to disability include cardiovascular disease, diabetes, arthritis, and stroke, but comorbidity is also an important risk for disability. Underlying behavioral risks for disability include low levels of physical activity, few social contacts, and smoking (Stuck et al. 1999)
The numbers of prevalent cases of disability in a population during any chosen time period is the result of the dynamic balance between the numbers of new cases (disability incidence), the numbers recovering from disability, and the numbers of deaths in those with and without disability. Understanding the relative contribution of each of these factors to the observed socioeconomic disability prevalence differences could lead to better targeting of efforts to reduce disparities. If most of the excess disability in less privileged groups is attributable to greater incidence, then greater exposure to the underlying causes of disability, less resistance to disability, or less access to effective preventive services are the most important factors. On the other hand, if lower recovery rates are the main factor, then factors facilitating recovery, including curative and rehabilitation services, might be more important. Differences in death rates offer complex potential explanations. The higher prevalence of disability in older people with less education would arise if death rates in less educated disabled elderly persons were lower than in more educated groups. Conversely, higher prevalence in less educated groups would also occur if there were substantially increased death rates in less educated elders without disability.
Previous work provides some evidence for higher rates of incidence of disability in less privileged subpopulations (Fried and Guralnik 1997; Stuck et al. 1999). Overall mortality rates have been shown to be elevated in those with less education (Feldman, Malone, Kleinman, and Cornoni-Huntley 1989; Lew and Garfinkel 2000), although an Italian study has suggested that mortality in disabled older people with less education is not raised (Amaducci et al. 1998), and disability may be the mediator between education and mortality, perhaps because of higher severity level of diseases. On the key issue of possible differences in recovery rates in older people by socioeconomic status, little is known.
A further problem in exploring the dynamics of socioeconomic status differences in disability prevalence is the variety of definitions of disability. Definitions based on an inability to perform basic activities of daily living identify a severe form of disability, which is often the end result of a progressive disablement process (Ferrucci et al. 1996). Lower extremity disability is often a precursor to more severe disability (Dunlop, Hughes, and Manheim 1997), and because of its relatively high prevalence and the large differences in prevalence across educational groups, it provides an important focus for exploring socioeconomic differences.
The measurement of socioeconomic differences in older people is also complicated by a number of factors. Most older people have no current occupation, and the former occupations, particularly of many older women, are a poor indicator of their social position. Current incomes may also not reflect older people's long-term material circumstances, even when these data are obtainable. In addition, both income and occupation may have been adversely affected by poor health in adult life, rather than the other way around. Years of education was used in our study as a marker of socioeconomic status, because it is closely related to long-term economic position (Smith and Kington 1997) and is less susceptible to the effects of later health status on employment and income.
In this analysis, we aimed to measure incidence rates of mobility disability together with recovery and death rates in data from a large population-based longitudinal study of older people—the Established Populations for Epidemiological Studies of the Elderly (EPESE). We used Markov model–based analysis to estimate relative risks, so that we could take account of the repeated observations of the same study respondents over the follow-up periods.
|
Methods |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
Details of the study methods have been published previously (Cornoni-Huntley et al. 1993). In-home baseline interviews were conducted between 1981 and 1983 followed by seven annual interviews in New Haven and Iowa and six in Boston. Proxy informants were interviewed when participants were unable to answer interview questions. Mortality was ascertained from obituaries and notification of death by friends and relatives. Death certificates were also obtained for decedents.
Assessment of mobility status was based on responses to the following questions: "Are you able to walk up and down stairs to the second floor without help?" and "Are you able to walk half a mile without help (about 8 blocks)?" Persons who responded "no" to either question were classified as having mobility disability. This classification of mobility has been shown to have predictive validity for mortality (Corti, Salive, and Guralnik 1996). We used responses to questions on the number of completed years of education to classify people into three groups: 0–7 years, 8–11 years, or 12 or more years.
Clear trends in the prevalence of mobility disability by years of education were present at baseline in people aged 65–84 but not in older persons, and hence the older group was excluded from the analyses. A total of 3,690 men and 5,618 women aged 65–84 were eligible for this analysis; data on education and mobility status at baseline and death or mobility status on at least one follow-up interview were available for 3,554 men (96%) and 5,317 women (95%).
Transitions in functional status were analyzed in two ways. First, a descriptive analysis treated each full year of follow-up as a "person-year" of observation. For example, if data on disability were available on a respondent at baseline and 7 follow-up years, then this respondent contributed 7 separate person-years of observations to the analysis. Where disability or vital status data were missing for the beginning or end of a study year, that year was excluded from the descriptive person-years analysis.
In age and sex-specific analyses, four transitions were studied: incidence of mobility disability, recovery from mobility disability, death in those who had disability at the start of the person-year of observation, and death in those who were not disabled at the start of the person-year. Prevalence of disability was measured as the proportion of observations in which disability was present. Incidence was measured as the proportion of those who were nondisabled at the start of each 1-year interval (Time 1) who reported being disabled at the follow-up interview (Time 2). The recovery rate was the proportion of disabled persons at Time 1 who reported being not disabled at Time 2.
We conducted analyses to measure education group differences in each component of prevalence according to mobility status at baseline. Among the nondisabled, risk of incident disablement, death, or no change in 1 year was estimated for men and women with 0–7 years and 8–11 years of education, each compared with those with 12 or more years, and data are presented within four 5-year age groups. Similarly, in the disabled, likelihood of death, recovery, or remaining disabled was estimated in the sex and education subgroups.
We drew inferences regarding differences in these transition rates using a Markov chain model developed for this purpose. Details of this method and examples of its use in disability research have been published elsewhere (Beckett et al. 1996; Mendes de Leon et al. 1995; Muenz and Rubenstein 1985). Briefly, we assumed that each person's likelihood for his or her given series of functional states could be expressed in terms of annual transition probabilities. Transition probabilities were estimated for each group identified by chosen baseline characteristics, but within these groups the estimated likelihood of change in functional status between two interviews was assumed to depend upon the functional state at the most recent interview, independent of the history of preceding functional states.
We modeled transition probabilities using logistic link functions for each of the four dichotomous outcomes: death versus survival from the nondisabled state; death versus survival from the disabled state; and, conditional upon survival to the next interview, incident disability and recovery. All participants with disability and mortality information from at least two of the eight interviews were included in the model-based analyses. We treated interviews with missing disability status flanked by interviews with nonmissing data by averaging (under the model) the predicted transition probabilities for all possible paths in between. We accounted for within-person correlation by using a robust variance estimator that treated each person's data as a cluster, or primary sampling unit. We made an adjustment for the stratified sampling from the New Haven site by treating the combined sample using a separate stratum for each of the other two sites. All analyses were done separately by sex and included main effects of continuous age and education (0–7 years or 8–11 years vs 12 or more years), as well as separate intercepts for EPESE site. We tested the Age x Education interaction but found that it did not significantly influence incident disability or mortality.
|
Results |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
|
|
|
The well-established pattern of higher overall mortality in those with less education was also evident, although absolute numbers of deaths were relatively small, especially for women. For example, deaths occurred in 6.5% (n = 336) of follow-up years in men with 0–7 years of education, compared with 4.8% (412) of men with 12 or more years of education. For women, these rates were 4.2% (310) in those with 0–7 years of education, compared with 3.3% (452) for those with 12 or more years of education. Person-year death rates by disability status at the start of the year (Fig. 3) varied, but showed little systematic difference across education groups in either those who were disabled at the start of the studied years or those who were nondisabled.
|
|
In those who were disabled, relative risks of death were also not significantly different by education. Death rates in those with less education who were not disabled showed complex differences, with odds in the least educated women being significantly lowered and those in the less educated men being higher than in those with 12 or more years of education.
To understand the influence of age on the transition risks, age-specific relative risks (Table 2 ) are presented for 5-year age bands. The incidence and mortality risks that were significant are evident, although they tended not to be statistically significant at the older ages. Relative risks for recovery were not significant at any age in men, and for women, recovery rates were significantly lower only in the youngest and least educated group. Within age groups, among those disabled, relative risks for death with lower education were not raised in women or men.
|
We explored the relative importance of the elevated incidence and elevated death rates in the nondisabled least educated men in a multistate life table model, using methods of Crimmins, Hayward, and Saito 1994 and Leveille, Penninx, Melzer, Izmirlian, and Guralnik 2000. These life table models used the four studied transition probabilities applied to a theoretical cohort to estimate prevalence of disability at each age. In a model of 1,000 men with 0–7 years of education entering the life table at age 65 with a 15% prevalence of disability, only 338 would survive to the age of 75, for example, of whom 92 (27%) would have disability. During their 75th year, there would be 33 incident cases of disability, 12 would recover from disability, 27 disabled men would die, and 13 nondisabled men would die.
Substituting the death rates only of men with 12 or more years of education into the life table for men with 0–7 years of education reduced the excess prevalence by a mean of only 10% of the difference between the least and most educated male groups, across the years of age in the studied range. On the other hand, substitution of the lower incidence rates only of the most educated men removed the elevation of disability rates in the least educated (with a mean change of 107% of the difference in prevalence across the age range).
|
Discussion |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
). Markers of socioeconomic status, whether measured by occupation, income, or years of education, have been repeatedly shown to be associated with disability prevalence (Guralnik, Land, Blazer, Fillenbaum, and Branch 1993; House et al. 1994; LaCroix, Guralnik, Berkman, Wallace, and Satterfield 1993; Stuck et al. 1999). In addition, some studies have also reported higher incidence of disability in less privileged subgroups (Stuck et al. 1999). However, in this study we go further in analyzing not only the inflow into the "pool" of prevalent disability in terms of incident cases, but also the outflow in terms of recovery and mortality. The results clearly show that relative risks of disability incidence are significantly higher in groups with fewer years of education compared with those with more education, but relative risks of recovery are not significantly different.
A number of factors need to be considered in evaluating these results. Years of education has been widely used as a marker of socioeconomic status (Smith and Kington 1997), but it does have some drawbacks. Years of education is treated here as a personal variable, although it is possible that effects of education may be mediated at the household level: The effects of spouses' education may result in some misclassification of true socioeconomic status, especially in the study generation of older women (Williams and Collins 1995). In the EPESE data analyzed, data on income were unavailable at baseline for 15.5% of cases in the studied age group, and those with missing income data were significantly more likely to have been disabled at baseline, suggesting potentially important biases in responses. By contrast, education data were missing in only 1.6% of cases. Years of education therefore provides the best available marker of socioeconomic status, not least because of its relative freedom from the effect of poor health on economic and employment status over most of the lifetime of the older people studied, because education is mostly completed early in life. A sex-specific secondary analysis indicated that the number of missing interviews was not correlated with the level of education, adjusting for number of interviews with or without disability and age, which adds some weight to the argument that missing data on education should not have introduced bias.
Mobility disability is measured here by self-report of being unable to walk or climb stairs without help. These measures are good markers of lower limb disability, which in turn are good predictors of progression of disability, nursing home admission, and mortality (Guralnik, Ferrucci, Simonsick, Salive, and Wallace 1995). Mobility disability represents an early stage of disability, and for example, in the Women's Health and Aging study, more than 90% of disabled older people had mobility difficulties (Fried, Bandeen-Roche, Kasper, and Guralnik 1999). Therefore, mobility disability is a key form of disability to study, but it is possible that other measures, for example of severe disability, could produce different results from those we report for mobility disability.
The Markov model presented has several advantages, including simultaneously taking account of all possible outcomes (incidence, recovery, or death); utilizing all the available data, even on those with some missing elements; and dealing with the effect of repeated observations on the same individuals. The model is also stratified for differences between study sites, but numbers are too small to yield significant results within each site. The Markov model does make the technical assumption that the transition probabilities are dependent only on the immediate state at the time of the interview, although this only operates in the model within the "risk factor" groups studied, defined by age, sex, and years of education, and should not introduce any obvious bias. Overall, the Markov model should provide the best possible estimate of the relative risks in which we are interested.
The Markov model results confirm the analysis of the person-year data in showing that the most significant factor influencing the prevalence differences in mobility disability by years of education is incidence of disability. Interestingly, using the same dataset and similar methods, Leveille and colleagues 2000 found that incidence was also the most important factor in sex differences in those aged less than 90. Rates of recovery from disability are not significantly different by years of education, and this may suggest that differences in care received after the onset of disability in different educational group are of limited importance in explaining the excess rates of mobility disability in less educated groups.
Overall mortality differences by years of education are well established (U.S. Department of Health and Human Services 1998; Wilkinson and Marmot 1998) and present in the study data. The results also suggest that the excess mortality rates are partly mediated through the higher rates of disability in less educated groups. Death rates in older people with disability did not vary significantly by years of education, supporting the suggestion that differences in course of illness after the onset of disability are not factors in mortality differences by socioeconomic status in old age. The excess risk of death in less educated men who were not disabled may reflect the effects of higher prevalence of cardiovascular and other conditions that can kill suddenly without prior disability. The multistate life table model for men, however, shows that differences in death rates between education groups make a small contribution to prevalence differences overall, compared with differences in incidence.
The chief importance of these findings is in targeting efforts to reduce disparities in health, of the sort envisaged in the Healthy People 2010 initiative (U.S. Department of Health and Human Services 2000). Although a great deal is already known about the risk factors for disability in old age (Stuck et al. 1999) and about programs to prevent disability (Wagner 1997), there is a popular tendency to assume that differences in care for already disabled elderly persons from less privileged socioeconomic groups must account for a substantial part of the poorer disability experience of these populations. This analysis shows, however, that at a population level differences in influences on recovery or survival with disability are of limited importance. It follows, therefore, that efforts to reduce disability disparities in old age should target prevention, including better management of disabling medical conditions before they progress to disability.
The most significant influence on the higher prevalence rates of mobility disability in older people who had fewer years of education is higher incidence rates of disability. Both recovery rates and death rates in those with disability do not differ significantly by socioeconomic status. This analysis suggests that programs aiming to reduce socioeconomic disparities in the prevalence of disability in old age should focus on prevention of disability onset.
|
Acknowledgments |
|---|
Received for publication April 26, 2000. Accepted for publication March 20, 2001.
|
References |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  M. I. Parahyba, K. Stevens, W. Henley, I. A. Lang, and D. Melzer Reductions in Disability Prevalence Among the Highest Income Groups of Older Brazilians Am J Public Health, January 1, 2009; 99(1): 81 - 86. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. A. Shaw and L. S. Spokane Examining the Association Between Education Level and Physical Activity Changes During Early Old Age J Aging Health, October 1, 2008; 20(7): 767 - 787. [Abstract] [PDF]
|
|
|
|
 |
|
 J. Adamson Commentary: Disability and education--the Madonna factor? Int. J. Epidemiol., May 12, 2007; (2007) dym028v2. [Full Text] [PDF]
|
|
|
|
 |
|
 C. O. Weiss, L. P. Fried, and K. Bandeen-Roche Exploring the Hierarchy of Mobility Performance in High-Functioning Older Women J. Gerontol. A Biol. Sci. Med. Sci., February 1, 2007; 62(2): 167 - 173. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Sulander, T. Martelin, P. Sainio, O. Rahkonen, A. Nissinen, and A. Uutela Trends and educational disparities in functional capacity among people aged 65-84 years Int. J. Epidemiol., October 1, 2006; 35(5): 1255 - 1261. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. K. Coppin, L. Ferrucci, F. Lauretani, C. Phillips, M. Chang, S. Bandinelli, and J. M. Guralnik Low Socioeconomic Status and Disability in Old Age: Evidence From the InChianti Study for the Mediating Role of Physiological Impairments J. Gerontol. A Biol. Sci. Med. Sci., January 1, 2006; 61(1): 86 - 91. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. F. Schoeni, L. G. Martin, P. M. Andreski, and V. A. Freedman Persistent and Growing Socioeconomic Disparities in Disability Among the Elderly: 1982-2002 Am J Public Health, November 1, 2005; 95(11): 2065 - 2070. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. S. House, P. M. Lantz, and P. Herd Continuity and Change in the Social Stratification of Aging and Health Over the Life Course: Evidence From a Nationally Representative Longitudinal Study From 1986 to 2001/2002 (Americans' Changing Lives Study) J. Gerontol. B. Psychol. Sci. Soc. Sci., October 1, 2005; 60(suppl_Special_Issue_2): S15 - S26. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. Breeze, D. A Jones, P. Wilkinson, A. M Latif, C. J Bulpitt, and A. E Fletcher Association of quality of life in old age in Britain with socioeconomic position: baseline data from a randomised controlled trial J Epidemiol Community Health, August 1, 2004; 58(8): 667 - 673. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
