| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||||||||||||||||||
a Scripps Gerontology Center, Miami University, Oxford, OH
Correspondence: Shahla A. Mehdizadeh, PhD, Scripps Gerontology Center, Miami University, Oxford, OH 45056. E-mail: mehdizk{at}muohio.edu.
Decision Editor: Laurence G. Branch, PhD
 |
Abstract |
|---|
 TOP Abstract MethodsResultsDiscussionReferences |
|---|
Key Words: Dual-eligible women • Health and long-term care expenditures • Long-term care career
Book Reviews
Practice Concepts
The Forum
The U.S. population aged 65 and older was almost 35 million in 2000. Almost 59% of this population was female (U.S. Bureau of the Census 2001
). The female proportion of the older population will gradually decrease from 59% in 2000 to 56% in 2025 and will then remain relatively stable through 2050 (U.S. Bureau of the Census 2000). This reduction in the percentage of the total population who are older women is due to expected improvement in survivorship of the male population.
Although the female proportion of the older population will decrease gradually over the next 50 years, the total number of older women is expected to grow from 20.6 million in 2000 to 45.7 million in 2050, an increase of 224% (U.S. Bureau of the Census 2000). With advanced age comes poor health and multiple chronic illnesses and disability; it is widely believed that women are more likely to be disabled than men of the same age (Laditka and Wolf 1998). In fact, older women tend to encounter chronic illnesses more than men. For example, Camacho, Strawbridge, Cohen, and Kaplan 1993 found that 70% of women aged 80 or older had two or more of the following chronic health conditions: arthritis, cancer, cataracts, diabetes, heart disease, hip fracture, hypertension, osteoporosis, stroke, and varicose veins. The health disadvantage is presumed to be associated, at least in part, with lower economic status. A considerable proportion of older women, especially members of minority groups, experienced this status: about 56% of African American women, 44% of Hispanic women, and 28% of White women older than 75 were classified as poor or near poor in 1992 (Gonyea 1997). Older women's greater numbers, higher life expectancy, lower economic status, higher incidence of chronic illness, and, often, lack of spousal caregivers (Blackburn and Chilman 1987) have led to their higher than proportional use of health and long-term care dollars. The literature on health and long-term care use that tracks individuals' expenditure patterns is mostly limited to a short period of time or to end of life.
In this study, I examined the health and long-term care use patterns in a sample of chronically disabled older women in different care settings. My purpose was to establish a long-term career for disabled older women in need of long-term care on the basis of their health and long-term care use. The long-term care trajectory begins with receiving home care, then proceeds with a period of home care and short stays in a nursing home, and then ends with a long stay in an institution. The study highlights and adds to the knowledge of the increasing demand on the health and long-term care delivery systems as aging female baby boomers reach age 65 and begin to require additional health and long-term care.
|
Methods |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
). The preadmission review is designed to determine whether an applicant is eligible for Medicaid long-term care, to examine the level of care needed by the applicant, to review information about alternative sources of care, and to offer the applicant service options that adequately meet his or her long-term care needs. The sample members were selected randomly each month from the older Ohioans who applied for long-term care between December 1993 and January 1995, a total of 114,558 applicants. Only individuals identified as needing long-term care and deemed eligible for Medicaid are represented in this sample. To construct the sample for this study, as shown in Fig. 1 stratified state data for all long-term care applicants 60 and older by referral setting and requested service outcomes. I selected 2,700 individuals for participation in the study. Because the major objective of the original research was an evaluation of Ohio's nursing home preadmission review process, the study over-sampled applicants from the community who were applying for long-term care by 2.8 times their representation in the preadmission review population and undersampled (by 0.44 times) those who were already residents in a facility and who were now applying for Medicaid assistance.
|
Because the sampling process extended over 3 calendar years and because acquisition of data for 3 additional years would have placed an extra burden on the agencies that were providing data to me (the Health Care Financing Administration, the Ohio Department of Job and Family Services, the Ohio Department of Health, and the Ohio Department of Aging), I further restricted the sample to those who were selected during calendar year 1994 and excluded those who had not reached age 66 by January 1994. There were 1,347 sample members (men and women) who met this additional criteria (Mehdizadeh, Applebaum, Warshaw, and Straker 2000). This study presents the characteristics and use patterns of the 1,016 women in this sample. Although the original sample members (N = 2,700) were weighted, several modifications to the sample (such as exclusion because of age or Medicaid and/or Medicare ineligibility) left the sample weight inappropriate and underrepresenting certain segments of the population; therefore, I did not use the weights. Still, the sample represents those who needed long-term care, either entering a Medicaid-certified nursing home or requesting Medicaid to pay for their care in the community or in a nursing home.
Measures of Utilization
I employed three different measures of use in this study: (a) average annual health and long-term care use in terms of number of hospital and nursing home admissions and number of physician visits; (b) average monthly health and long-term care expenditures on hospital inpatient and outpatient care, nursing home care, home health care, physician services and durable medical equipment, hospice care, medication, Medicaid home care, and other Medicaid services such as medical transportation, by payer source; and (c) average annual total health and long-term care expenditures regardless of the payer source. I defined health care expenditures as Medicare expenditures on hospital inpatient and outpatient care, physician visits, home health care, and Medicaid as copayer for these expenditures as well as Medicaid prescription payments. The long-term care expenditures were defined as Medicare and Medicaid nursing home care, hospice care, and durable medical equipment, as well as expenditures for home- and community-based care services (HCBCS) and other Medicaid services.
Because some of the sample members were declared Medicaid eligible at the time of preadmission review, no Medicaid payment of Medicare Part B premiums had been made for these newly qualified Medicaid beneficiaries during the year before preadmission review (retrospective use). No additional information was available to inform me whether these sample members had paid Medicare Part B premiums for the previous 12 months. Therefore, on the basis of the Medicare claims data, I could not differentiate between lack of Part B benefits and sample members' nonuse of services such as physician and durable medical equipment services. For this reason, the retrospective examination of use was limited to Medicare Part A benefits.
Health Status Data
Information regarding disease diagnosis came from different sources, based on sample members' referral source to the preadmission review process. A nurse or social worker collected this information on the basis of the sample member's acknowledgment of the condition, verified by his or her physician for the community referrals; a list of diagnoses, checked if the condition was present for the nursing home referrals, identified by the Mimimum Data Set; and the hospital transfer forms completed by a nurse or physician sometimes listed conditions as the sample members were transferring to nursing home from hospital for the hospital referrals. The information was not consistent across referral sources and was not available for all sample members within a referral source. When the information was present, it was not possible to determine the disease progression stage or the specifics about the disease, such as cancer. For this reason, the use of health status variables in this study is limited to presence or absence of certain diseases or conditions prevalent among the older population, as well as the number of diseases present, acknowledging that for about a fifth of the sample I could not differentiate between those who had no known diagnosis and those for whom assessors failed to complete the health status section of the preadmission review form.
Design and Analysis
As the female applicants completed the preadmission review process, they were either referred to a nursing home or given the choice between receiving HCBCS or entering a nursing home. Some applicants were referred to a nursing home, perhaps after a hospitalization, and then received HCBCS after discharge from the facility. Therefore, the sample members were grouped on the basis of the setting from which they received long-term care services during the first year following their preadmission review. The three subsamples in this study are the community subsample, those who remained in the community the entire first year; the transition subsample, those who made the transition between home and nursing home one time or more during the first year, and the nursing home subsample, those who stayed the entire first year in a nursing home. This classification system not only describes this sample, it may also reflect the long-term care career of older disabled individuals, beginning with HCBCS, followed by a combination of community care and stay in a nursing home, and ending with an extended stay in a nursing home.
The nursing home subsample consisted of newly admitted residents from the community, residents transferred from other facilities, or residents who had spent down their assets and had become eligible for Medicaid reimbursement for their health and long-term care services. A comparison of the expenditures and use measures of the different components of this subsample revealed that although these individuals may have been at different points in the same stage of a trajectory, their expenditures and use patterns were very similar. Using t-test statistics to examine significant differences between the newly admitted and those who spent some time in a nursing home, I found that there were no significant differences between expenditures and use in Year 1. However, Medicare nursing home expenditures were significantly higher for the newly admitted residents in Year 2, perhaps reflecting the acute conditions of the newly admitted residents and the need for hospitalization. The differences in expenditure patterns in the year before were consistent with the setting in which each group received its care (significantly higher Medicare home health care expenditures and lower Medicare nursing home and physician care expenditures for the newly admitted residents). Because there was no major difference in the expenditure patterns of the two groups, except those reflecting the care setting, the use patterns of all nursing home residents were analyzed together as a group.
To establish a long-term care career for the sample members, I ideally needed a longitudinal data set that tracked health and long-term care use and expenditures of these women for an extended period of time. In absence of such a data set, I used subsamples at different stages of this career. To show the differences in use and expenditures at each stage of this trajectory, I used analysis of variance.
However, I discuss the expenditures across time within each group (stage) but did not test them for significant changes over time within each group (stage), for two reasons. First, those with the most serious health conditions died during the 1st and 2nd years; therefore, the remaining sample members' conditions might have been more stable and their health and long-term care expenditures might have been lower. Second, the assignment of sample members to the three groups (community, transition, and nursing home) applied only to the 1st year. In the following year, some sample members' group membership changed as their needs changed.
|
Results |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
Table 1 summarizes the total sample's demographic, health, functional, mortality status, and expenditure patterns. The average age of the sample members in the study was about 81 years, and 12% were 91 years or older. About 13% were married, 28% were cognitively impaired, and about 63% had some form of circulatory disease. More than 90% were impaired in bathing; more than 68% were impaired in dressing and grooming; and 32% of the sample had four or more activity of daily living (ADL) impairments. The average ADL score for the sample was 2.8, and about 16% of the sample members died during the first year. On average, the sample members had 1.4 hospital admissions and 32 physician visits during the first year. Seventy percent also had a nursing home admission during the 1st year. Average monthly total health and long-term care expenditures for the sample members were about $2,750.
|
); this difference reflects the minority population's extent of frailty and poverty and perhaps preference for noninstitutional care.
|
The transition subsample members were the second youngest group. Overall, the transition subsample represented a population of unmarried older women (hence without a spousal caregiver) with the highest proportion of non-Whites (27%). This subsample, although slightly more disabled than the community subsample on the basis of average ADL scores, differed from the other two subsamples in regard to the specific ADLs in which they were disabled. More than 90% were impaired in bathing, about two thirds were impaired in dressing and grooming, and almost half were impaired in transferring and toileting; this combination of ADL impairments made staying in the community a challenge for this group as they resisted a long-term institutional stay. It appears that their frequent nursing home stays were follow ups to hospitalization for certain acute conditions, a situation that was not always observable from the hospital transfer forms. About 22% of the transition subsample members were diabetic, two thirds had circulatory diseases, more than a quarter had musculoskeletal diseases, and about 20% had respiratory diseases. About a third of this subsample had one disease, and about half had at least two of the listed diseases. The mortality rate was highest, 24%, in this group.
The nursing home subsample was older and predominantly unmarried (88%). Forty-four percent of the women who stayed in a nursing home for the entire year were 85 years or older. Fifteen percent of the nursing home sample were non-White, again reflecting the minority population's fraility but also emphasizing that the members of minority populations are as likely as the White population to use nursing home care. The members of this subsample showed the highest proportion of impairment in all ADL items except bathing and the highest overall average ADL score (3.2). Slightly fewer than 50% were cognitively impaired. About one half of the members in this subsample had circulatory diseases, and 20% had musculoskeletal diseases. About one third had no diagnosis or a diagnosis was missing, a third had one disease diagnosed, and another third had at least two diseases diagnosed.
Table 3 shows the sample members' use patterns. The use and expenditure patterns of the three groups are equally interesting; they reflect sample members' health status and the practice patterns and specifications of the care settings. The clients who remained in the community, on the average, were hospitalized at least once per year; the hospitalization rate for this group also increased throughout the study. The rate of nursing home admission for the community subsample was about one admission for every five subsample members in the year before preadmission; the rate was none during the 1st year, by subsample definition, and then it increased to one for every two subsample members during the 2nd year following preadmission.
|
), that perhaps in nursing homes illnesses are recognized at an earlier stage, and residents receive care for certain acute conditions in the nursing home rather than being transferred to a hospital. The rate of hospitalization was lowest in the nursing home group. The annual number of physician visits in this group was comparable to that in the community subsample. Long-term care expenditures were significantly different in the three groups. As might be expected, expenditures were highest among those who spent the entire year in a nursing home, closely followed by those in the transition group. These expenditures were lowest among those who remained in the community. This finding is due in part to the conditions of Ohio's 2176 Medicaid waiver: HCBC clients are subject to a 6-month caseload cost cap, which on average should not exceed 60% of nursing home care.
The medication expenditures were significantly different across the three groups. A comparison of Part A Medicare expenditures revealed that the transition group had significantly higher average monthly expenditures in the year before as well. In fact, the three groups maintained a similar relationship to each other in Years 1 and 2 (the year before and the 1st year) when expenditures are compared. This indicates that expenditures for the year before could be used as predictors of current-year expenditures. The mortality rate was significantly different in the three groups, with the transition subsample having the highest rate for both years. By the end of the 2nd year, almost a quarter of the community and nursing home subsamples and more than one half of the transition subsample had died (cumulatively). In general, the transition group had a significantly higher rate of hospital admissions, nursing home admissions, physician visits and mortality rate compared with the other two groups.
Next, I examined health and long-term care expenditures. Regardless of where the subsample members received their long-term care services, they represented a frail population that had embarked on its long-term care career. Fig. 2Fig. 3Fig. 4 present trajectories of 3-year expenditures for the older women in the three subsamples.
|
|
|
The average monthly nursing home expenditures for the transition group started with a relatively small amount (less than $100 per month Medicare payment) and increased to a payment of $570 per month by Medicare in the next year. Medicaid paid more than half of the nursing home bill, a reflection of the longer nursing home stays (see Fig. 3). This pattern changed further in the 2nd year (following preadmission review) to a much larger payment of nursing home expenditures by Medicaid ($1,006) and a much smaller payment, about 20% of total nursing home expenditures ($183), by Medicare. This change may have signaled the transition to a long-stay nursing home.
The group that remained in the community for the entire 1st year became progressively more disabled and needed additional health care services. In fact, at the end of the 2nd year following preadmission, the health care expenditure patterns for this group were similar to the retrospective expenditures of the transition group. As shown in Fig. 4, the sample members' health and functional status grew worse (according to their expenditure patterns), and a shift from health care expenditures to long-term care expenditures and from Medicare as the major payer for care to Medicaid's paying for a more substantial portion of the care occurred. As this figure shows, when clients spent the entire time in a nursing home (the nursing home subsample), Medicare's share of the total care cost declined to 25% from a high of 70% in the transition group and 57% in the community group.
|
Discussion |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
). In addition, the women in this sample were considerably more disabled than the overall national dually eligible population: 21% of the sample had none or only one ADL impairment (vs 34%), and 32% of the sample had four or more ADL impairments (vs 18%; Mollica and Riley 1997). This study began when individuals started to experience certain health conditions, followed by functional impairments that made them unable to care for themselves (or made their family unable to care for them). Taking this stage as the beginning of the sample members' long-term care career, I classified the study participants into three groups on the basis of their needs and the long-term care options they received. The three groups were those who received long-term care services in the community in the 1st year following their preadmission review; those who received such care in a nursing home and in the community; and those who received care in a nursing home. The 2176 Medicaid waiver, pertaining to HCBCS in Ohio, is structured so that this option is offered only to those whose average estimated care plan cost for 6 months does not exceed 60% of the cost of nursing home care. Those who maintain this level can continue to receive HCBCS as long as they meet eligibility criteria for Medicaid and nursing home level of care. We learned that those who were able to take advantage of HCBCS were at an earlier stage of their long-term career.
If each group (community, transition, and nursing home) is viewed as representing a stage in a long-term care career in which one stage follows another, one finds a 6-year trajectory for a population in need of long-term care, where the expenditures in the 2nd year of each stage are similar with the retrospective expenditures (year before) of the preceding group (see Fig. 2 and Fig. 4). This linking of expenditures shows how individuals' patterns of expenditure shift from considerable acute and inpatient costs (paid by Medicare), shown in Fig. 2, to chronic care in a nursing home (paid by Medicaid). I acknowledge that the most vulnerable persons, suffering grave conditions, die in the process, thus displaying a much shorter long-term care career.
In a previous study, Mehdizadeh, Applebaum, and Straker 2001 learned that residents who stayed in a nursing home beyond 9 months would remain there for the rest of their lives. Therefore, for those who survived beyond the 6th year of the trajectory, the pattern of expenditures after the 6th year would be similar to the nursing home stage of the trajectory. The average age of the women in this study was a little older than 80; about one quarter were under 74 and almost one third were older than 85. The breakdown of age in each of the three groups supports the construction of a 6-year long-term care trajectory, as shown in Table 2 .
Not all women reaching 65 years of age and older would be as disabled as those presented in this study. Several studies examined various data sources from the mid-1980s to the mid-1990s and found that the proportion of life spent free from impairment (hence active life expectancy) has increased, and the prevalence of disability in the older population has decreased (Crimmins, Saito, and Ingegneri 1997; Freedman and Martin 1998; Manton, Corder, and Stallard 1997).
A recent study attempting to predict determinants of active life expectancy found that the period of inactive life decreases with higher educational attainment (Laditka and Laditka 1998). Women reaching old age in the first half of the 21st century are expected to be better educated (U.S. Bureau of the Census 1994) and more financially secure than their predecessors, such as the sample members in this study. Labor force participation for all women increased from 52% in 1980 to 60% in 1997; this rate is projected to increase to 61% in 2006 (U.S. Bureau of the Census 1998, p. 403). In addition, the types of employment opportunities available to women have improved, as has the level of pay: women's median weekly earnings increased by 55% between 1985 and 1997 (U.S. Bureau of the Census 1998, p. 436). (The gains in education, labor force participation, and retirement income security are much higher for White women than for minority women.) Women's tenure with current employer, indicating an uninterrupted employment history, has also increased for all ages, although this tenure is shorter among younger women (age 25 to 54) while they are raising families (U.S. Bureau of the Census 1998, p. 415). These changes have brought some women higher earnings while they are employed, and sometimes have provided additional retirement income and health benefits beyond Social Security and Medicare. In 1996, for example, 26.5 million women (45% of all working women) had some form of pension benefits (U.S. Bureau of the Census 1998, p. 385).
Even as women's life expectancy continues to increase (although the differences between men's and women's life expectancies are narrowing), and as some women become more educated and have better retirement benefits, most women, particularly members of minority groups, still face a period of inactive life marred by functional impairments and dependence.
Implications of Findings
In this study, I have attempted to document health and long-term care use during the year that one became physically inactive. Ohio will, as will the rest of the nation, experience a steady increase in its older population and its older female population. The nation will face a new challenge in the near future. As age increases, the level of impairment and the need for assistance will increase. A certain proportion of this increased population will rely on state-reimbursed long-term care. As the states plan and reevaluate their long-term care policies, they should remember that those who come to rely on the state for their care may need this care for a long time (6 years or longer).
Summary
By examining the health and long-term care use trajectories of a sample of chronically disabled older women, one gains insight into the use patterns of this population as they reached this life stage, regardless of their age. One also learns that health conditions and chronic disability occur gradually and grow worse with each passing year. Older women facing these conditions seek assistance and try to remain in the community; as their conditions become more serious, they negotiate between living in the community and an institutional stay. Those who survive these two stages will eventually enter a long-stay nursing home.
|
Acknowledgments |
|---|
I am grateful to Brooke Trisel, health services policy specialist at the Ohio Department of Job and Family Services, and to Neil Jordan, a graduate student at ResDAC, University of Minnesota School of Public Health, who diligently assisted me in understanding and linking the Medicare and Medicaid claims data. I am also thankful to my colleagues Lisa Groger and Suzanne Kunkel, who commented on an earlier version of this article. The useful and constructive comments from the anonymous reviewers greatly improved this article.
Received for publication April 19, 2001. Accepted for publication October 30, 2001.
|
References |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  R. L. Kane, P. Homyak, B. Bershadsky, Y.-S. Lum, and M. S. Siadaty Outcomes of Managed Care of Dually Eligible Older Persons Gerontologist, April 1, 2003; 43(2): 165 - 174. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
