Douglas A. Wolf, PhD
Sarah B. Laditka, PhD
James N. Laditka, DA, MPA
SUMMARY. This study examines the distribution of total, unimpaired, and impaired life for several groups of older women defined by race, education, and marital history. Using data from the 1984-1990 Longitudinal Study of Aging, we model transitions among functional statuses using discrete-time Markov chains, and use microsimulation to produce summary indices of active life. Remaining years of life and the proportion of remaining years with disability vary substantially, both within each group of women studied and between pairs of groups. Of all groups studied, never-married, more-educated white women live the longest, healthiest lives. Ever-married nonwhite women with low education have the shortest life expectancy, and experience the most disability. Our findings
show that life expectancy is an incomplete indicator of the time women, in particular sub-groups, can expect to live with and without impairment. These findings highlight the heterogeneity of disability processes and life expectancy for older women.
[Article copies available for a fee from The Haworth Document Delivery Service: 1-800-HAWORTH. E-mail address: <[email protected]> Website: <httÂp:/Â/wwÂw.HÂawoÂrthÂPreÂss.Âcom> © 2002 by The Haworth Press, Inc. All rights reserved.] KEYWORDS. Active life expectancy, disability, functional status, Markovchain, microsimulation
INTRODUCTION
As the number of older Americans grows, both public and private institutions face the possibility of increasing demands for health services, accompanied by rising public and private health care costs. Population aging might similarly add to the hidden costs borne by those who provide the most help to older persons, namely family and friends. Public policies often address these informal service providers, as well. For example, many states have instituted caregiver support programs (Feinberg & Pilisuk, 1999), and consequent to the recent reauthorization of the Older Americans Act, the federal government has launched a National Family Caregiver Program (U.S. Department of Health and Human Services, 2001). Planning for the challenges of an aging population can be informed with improved information about patterns of active life. A better understanding of active life patterns among older women is especially important, since women comprise a large majority of our older population. Given their longer lives and more years of disability, women also use more health care than men.
The percentage of United Statesâ households with single adults has increased greatly (Ahlburg & DeVita, 1992; Smeeding, 1999). Blacks have been particularly affected by declining marriage rates and the increase in households headed by women (Taylor, Chatters, Tucker, & Lewis, 1990). Blacks also constitute a growing percentage of the U.S. population (U.S. Bureau of the Census, 1999). In light of these demographic trends, it is important to understand patterns of active life for groups of older women defined by race and marital status.
Research on active life has focused almost exclusively on the average number of years an individual can expect to live without, or with, disability, that is on âactiveâ or âinactiveâ life expectancy. Such measures are useful for actuarial calculations. They can be used, for example, to determine the insurance value of long-term care services. But averages have their limitations. At any given age, the likely number of additional years lived with and without disability varies considerably across individuals. This variation occurs within groups defined by characteristics associated with disability processes and longevity; some individuals in the group will experience below-average, and others above-average episodes of disability, and many will never experience disability at all. Between-group differences in averages, and in departures from those averages, occur as well. Thus, when considering broader issues of equity and efficiency in the financing and provision of services, or when targeting programmatic resources, it is useful to recognize the full distribution of active, impaired, and total life, and not only the averages of each. To our knowledge, no past research has explicitly traced the frequency distribution that is implicit in calculations of active life expectancy. Similarly, few previous studies have compared averages of such distributions across groups of older women.
Our study examines variations in three measures commonly reported in the active life expectancy literature: total life, active (unimpaired, or disability free) life, and inactive (impaired, or disabled) life. We examine this variability from two perspectives. First, we show how these measures can vary within a given group of older women, where group membership is defined by race, education, and marital status. Second, we investigate differences in total, active, and inactive life expectancy between such groups. Our analysis focuses on older women, using data from the 1984 to 1990 Longitudinal Study of Aging (LSOA) and microsimulation techniques.
FACTORS ASSOCIATED WITH DISABILITY
AND ACTIVE LIFE EXPECTANCY
Past research has shown that disability prevalence and incidence, and active life expectancy, differ substantially across groups of the older population. Studies have consistently shown that older women with more education live longer and healthier lives than those with less education (Crimmins, Hayward, & Saito, 1996; Crimmins & Saito, 2001; Freedman & Martin, 1999; Land, Guralnik, & Blazer, 1994). Researchers have suggested that education may influence individualsâ ability to understand and reduce risk factors, and to adopt healthier lifestyles. There is also evidence of notable mortality and morbidity differences between blacks and whites, although findings in this research area are inconsistent. A growing number of researchers have found that white women have both total and active life expectancies greater than those of black women (Crimmins et al., 1996; Crimmins & Saito, 2001; Geronimus, Bound, Waidmann, Colen, & Steffick, 2001; Hayward & Heron, 1999). Researchers point to socioeconomic, cultural, and genetic factors, as well as other advantages and disadvantages across the life span, as likely causes of racial disparities in mortality and morbidity (e.g., Hayward, Crimmins, Miles, & Yang, 2000).
Many studies have examined associations between marital status and mortality. Studies have variously investigated differentials by marital status, marital history, and marital events. For example, Lillard and Waite (1995) modeled mortality risks over a 20-year period for a sample of adults of all ages, finding elevated mortality rates among never-married and separated women (compared to currently-married women), and little difference between the mortality rates of widowed and currently-married women. Lillard and Waite (1995) also found that the benefits of being married grow as the duration of marriage grows. Others have investigated the consequences of experiencing spousal death, or bereavement. Schaefer, Quesenberry, and Wi (1995) found that womenâs mortality rates were significantly higher 7-12 months after the death of their spouse, but not before or thereafter. Their study did not, however, include comparisons to a never-married group. Korenman, Goldman and Fu (1997), using data from the LSOA, found a significant adverse mortality effect among women widowed more than one year, but not during the first year o...