This Article Extract Reprint (PDF) Submit a response to this article Alert me when this article is cited Alert me when Comments are posted Alert me if a correction is posted Services E-mail this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Add to My Personal Archive Download to Citation Manager Reprints & Permissions Citing Articles Citing Articles via HighWire Citing Articles via Web of Science (4) Citing Articles via Google Scholar Google Scholar Articles by Monheit, A. C. Articles by Eisenberg, J. M. Search for Related Content PubMed PubMed Citation Articles by Monheit, A. C. Articles by Eisenberg, J. M. Related Collections Access To Care Medicare

Health Tracking

TRENDS

Moving To Medicare: Trends In The Health Insurance Status Of Near-Elderly Workers, 1987–1996

Research interest has frequently focused on the workforce incentives of employment-based health insurance. For example, the availability of work-related coverage and retiree health benefits to older workers has been found to affect their decisions regarding job mobility and early retirement and to reduce their new employment opportunities.³ In addition, researchers have examined the near-elderly’s health insurance status using single cross-sections of data or relatively short panel data sets.⁴ However, few researchers have investigated older workers’ health insurance status over time or relative to other age-specific worker cohorts or to nonworkers.

In this paper we consider how the health insurance status of specific worker age cohorts has changed over the past decade. Next, we consider whether the health insurance status of older workers differs from that of younger workers in sources of coverage and policyholder and dependent status and identify those near-elderly workers especially at risk of being uninsured. We also consider the insurance status of the nonworking near elderly, including how it contrasts with that of near-elderly workers and how it has changed over the past decade. Finally, we investigate whether the stability of coverage over the year varies across worker age groups.

This examination is important for stimulating research on the impact of legislative initiatives —such as the 1985 Consolidated Omnibus Reconciliation Act (COBRA) and the 1996 Health Insurance Portability and Accountability Act (HIPAA)—designed to ensure access to employment-based or privately purchased coverage. In addition, under former President Bill Clinton’s Medicare buy-in proposal, certain near-elderly workers ages fifty-five to sixty-one who lose coverage as a result of a job loss would be eligible, as well as persons ages sixty-two to sixty-four who lack employment-based and other sources of coverage.⁵ Our analysis can help to inform such policy proposals by identifying those groups of near-elderly workers who would benefit from such initiatives and the potential costs associated with their participation.

Data And Methods

 
Data sources. We derived our data from two nationally representative household surveys sponsored by the Agency for Healthcare Research and Quality (AHRQ): the 1987 National Medical Expenditure Survey (NMES) and the 1996 Medical Expenditure Panel Survey (MEPS). The 1987 NMES is a year-long panel survey of approximately 15,000 households consisting of nearly 36,000 individuals in the civilian, noninstitutionalized population, while the 1996MEPS is a two-year panel survey of approximately 10,000 households consisting of nearly 23,000 individuals. Each of these surveys provides detailed information on the population’s use of and spending for health services, health insurance, health status, demographic characteristics, and employment and economic status. The surveys also have been designed to facilitate temporal comparisons and have been used to assess changes in the population’s health insurance status over the past decade.⁶

Definitions. For purposes of this analysis, we focus on adult’s ages twenty-one to sixty-four. We identify persons as workers if they were employed at the Round 1 interview date in each survey and use information on health insurance status obtained during each survey’s first interview round. Health insurance status is based on whether coverage was held at any time during the first round of each survey. Thus, covered persons had insurance at any time during the Round 1 reference period; uninsured persons lacked coverage throughout the Round 1 reference period. For NMES this yields estimates of the population’s coverage status, on average, for the first quarter of 1987. For MEPS such estimates correspond, on average, to the population’s coverage during the first half of 1996. Although shorter reference periods can yield lower estimates of private and public insurance coverage and higher estimates of being uninsured compared with longer reference periods, disparities in coverage rates observed over a ten-year period are far more likely to reflect changes in population characteristics, economic conditions, and the costs of coverage than reference period length.⁷

For each survey, health insurance status is defined hierarchically into the following mutually exclusive classes: having any private coverage, having public coverage only, or being uninsured. Private coverage is further broken down into the following hierarchy: first, whether an individual has any employment-related coverage (including as a policyholder or a dependent) and next, whether other private coverage was obtained outside the workplace. Public health insurance consists of coverage obtained from Medicaid, Medicare, other state or local programs that provide comprehensive hospital and medical coverage, or programs covering military retirees and dependents (for example, TRICARE). We supplement these data with information on whether a worker was offered employer coverage (for wage earners only) and, if so, whether he or she enrolled in the offered coverage. Finally, we draw upon data on monthly health insurance status from the 1996 MEPS to examine the stability of health insurance status by age cohort for workers who were employed throughout 1996 and were present in the survey for the entire year. Unless otherwise indicated, all differences reported in the text are significant at the 0.05 level of significance or better.

Health Insurance Status As Workers Age: 1987 And 1996

Top Data And Methods Health Insurance Status As... Near-Elderly Workers At Risk... Discussion And Policy... NOTES

 
Our data reveal that in 1987 and 1996 the health insurance status of near-elderly workers compared favorably with that of prime-age workers ages 30–54 and that both groups of workers were less likely to be uninsured than were young working adults ages 21–29 (Exhibit 1). Thus as a group, near-elderly workers do not stand out as being especially at risk of being uninsured.

Between 1987 and 1996 uninsurance rates increased for workers of all ages. What is especially striking about this change is that the source of this increase differed between near-elderly and younger workers. In particular, our data reveal that the increase in uninsurance rates for young and prime-age workers was primarily associated with a loss of job-based coverage. By contrast, the decline in private coverage purchased outside the workplace explains most of the increase for near-elderly workers (Exhibit 1).

Age and sex. While private insurance purchased outside the workplace is an important source of coverage for all near-elderly workers, women are especially reliant on such coverage (Exhibit 2). For example, in 1987 nearly a fifth of working women ages 60–64 relied on privately purchased nonwork coverage, more than three times the rate for prime-age working women. Moreover, such coverage compensated for the lower rates of employer coverage for older working women (60–64) relative to their male counterparts. By 1996 the rate of privately purchased nonwork coverage by such older women had declined sharply, while rates of employment-based coverage remained fairly stable.⁹ As a result, a higher percentage of these women became uninsured (from 7.7 percent in 1987 to 13.1 percent in 1996, p< .10).

The nonworking near-elderly. In both 1987 and 1996 the nonworking near-elderly were less likely to be covered by private insurance than were workers in the same age group (Exhibit 3). However, both near-elderly male and female nonworkers were more likely to rely on public sources of coverage, which may have been associated with disabling health conditions that precluded employment.

Similar to older working women, the decline in privately purchased coverage among nonworking near-elderly women (60–64) was a key factor in the increase in their uninsured rates over the past decade. Changes in coverage rates for both workers and nonworkers resulted in statistically equivalent uninsurance rates for both groups in 1996. In contrast, in 1987 nonworking near-elderly males (55–64) and females (60–64) had uninsurance rates that exceeded those of their working counterparts.

As with near-elderly male workers, nonworking near-elderly men were more likely to be covered by a spouse’s work-based health plan in 1996 than they were in 1987. Indeed, this group nearly doubled the rate at which they became dependents on their spouse’s plan (p < .10), which may in part reflect the shift to policyholder status among working near-elderly women documented earlier. In contrast, nonworking near-elderly women (55–59) were more likely to hold job-based coverage in 1996,while those ages 60–64 were less likely to hold such coverage over the decade (p < .10).

Trends in offer and take-up rates. Our data reveal that in both 1987 and 1996 near-elderly workers ages 60–64 were less likely than prime-age workers were to obtain offers of employment-based coverage. Considering age and sex, we find that the lower offer rates for older workers in each year have been borne exclusively by female workers ages 60–64. While offer rates for males across age groups were not statistically different in either year, near-elderly female workers (60–64) had offer rates considerably below those of their prime-age counterparts. Moreover, near-elderly female workers’ access to employer coverage was no different than that of their young female counterparts, the age group that has been consistently shown to have the worst rate of employer coverage.

As described by Philip Cooper and Barbara Schone, offers of work-related health insurance increased overall for working Americans between 1987 and 1996.¹¹ However, this increase occurred only for working women age thirty and older, while offer rates actually declined for prime-age men and remained constant for other male workers (Exhibit 4). Despite the increase in offer rates for near-elderly workingwomen (60–64), in 1996 such women had only achieved the offer rate (68.8 percent) obtained by prime-age female workers a decade earlier. These changes in offer rates by sex are consistent with the shift in the policyholder and dependent status of near-elderly family members reported earlier.¹²

Acquisition of coverage by uninsured near-elderly workers. In 1996 near-elderly workers were as likely to be uninsured as prime-age workers were. However, MEPS data on monthly health insurance status indicate that when older workers are uninsured, they are less able to obtain health insurance than younger workers are. When we consider workers who were uninsured in January 1996 (data not shown), we find that 19.4 percent of uninsured workers ages 55–64 acquired private or public coverage over the remainder of the year, compared with 28.3 percent of prime-age workers (p < .10) and 35.6 percent of young adults. Considering access to private coverage, only 16.9 percent of near-elderly workers obtained any private coverage over the year, compared with 26.3 percent of prime-age workers and 32 percent of young adults (rates of enrollment in public insurance were quite low and not different among these groups).¹³

Thus, the observed comparability in static uninsurance rates conceals the somewhat greater difficulty older workers have in obtaining health insurance when they are uninsured. This disadvantage is a key policy concern since it may reflect the difficulty near-elderly workers face, especially those with health problems, in finding jobs that provide coverage or in paying high premiums for individually rated health plans. Our data also indicate the absence of a public safety net to address the gaps in coverage among near-elderly and other workers.

Near-Elderly Workers At Risk Of Being Uninsured

Top Data And Methods Health Insurance Status As... Near-Elderly Workers At Risk... Discussion And Policy... NOTES

 
As we have noted, near-elderly workers as a group are not especially at risk of being uninsured but are unlikely to be able to obtain coverage once they become uninsured. Concern for this group largely reflects the fact that health insurance may be especially valued by near-elderly workers because of the greater prevalence of health problems among this age cohort relative to younger workers. When older workers with health problems are insured all year, and therefore more likely to have access to needed health services, they have average annual expenditures of $5,000, nearly twice the level of their counterparts in excellent or very good health ($2,548).¹⁴

One group at particular risk for high out-of-pocket spending is near-elderly women with health problems (approximately 1.8 million persons).¹⁵ Data from the 1996 MEPS reveal that these women are especially at risk of being uninsured. We find that 22.7 percent of female workers ages 55–64 who report themselves in good, fair, or poor health were uninsured, compared with only 9.5 percent in excellent or very good health. In contrast, near-elderly men with health problems were no more likely to be uninsured than those in better health and were also less likely than their female counterparts to be uninsured (15.8 percent of men versus 22.7 percent of women, p < .10).

The relatively high uninsurance rate for near-elderly female workers with health problems may reflect their poor economic status relative to their counterparts in excellent or very good health and, therefore, their ability to purchase coverage or to have access to the kinds of jobs that offer coverage. For example, 25.8 percent of near-elderly female workers with health problems had incomes less than twice the poverty level, compared with only 8.7 percent of those in excellent or very good health; 66.4 percent of the former also had hourly earnings of less than $10, compared with 40.3 percent of the latter (for wage earners only; data not shown).¹⁶

Differences in health insurance status between prime-age and near-elderly female workers also reflect the latter’s lack of dependent coverage. Using the 1996MEPS (data not shown), we find that the likelihood of holding dependent coverage through a spouse’s work-related health plan declines with age, from 22.2 percent for prime-age female workers in good, fair, or poor health to 14.3 percent for their near-elderly counter parts. What is also striking in our data is that we observe the opposite pattern for men in good, fair, or poor health: Near-elderly working men with health problems have higher rates of dependent coverage than do prime-age males with health problems (13.8 percent versus 8.8 percent, p < .10). As a result of this difference, near-elderly working men with health problems have a lower likelihood of being uninsured than their prime-age counterparts have (15.8 percent versus 22.9 percent).

Among prime-age and near-elderly workers, the likelihood of obtaining an offer of employment-based coverage declined with age for women workers and was lowest for near-elderly women ages 60–64 in both years. When we consider 1996 offer rates according to age, sex, and health status (Exhibit 5), we find that the low offer rates for older working women reflect in part the experience of those with health problems.¹⁷ This appears to be an important factor in the observed differences in offer rates for near-elderly female workers relative to their prime-age counterparts.

Top Data And Methods Health Insurance Status As... Near-Elderly Workers At Risk... Discussion And Policy... NOTES

Our findings draw attention to the sharp decline over the past decade in health insurance purchased outside the workplace by near-elderly workers, especially women. This decline has been an important factor in the rise in uninsured rates for near-elderly workers (women ages 60–64; men ages 55–59) as well as nonworkers (women ages 60–64). There has been considerable public policy interest recently in expanding the role of the individual health insurance market for persons without access to employer coverage.¹⁸ Our findings suggest that an understanding of the factors associated with the sharp decline in coverage purchased outside the workplace by near-elderly workers over the past decade is essential to assessing whether this market can address the gaps in coverage they experience. Of particular concern is whether and why access to the nongroup market by such persons changed over the decade we studied. In this regard, a 1998 General Accounting Office report identifies high premiums and the use of preexisting condition exclusions in the individual market as important impediments to coverage for near-elderly persons with low incomes and health problems.¹⁹

State insurance reforms directed at the individual market have sought to expand access to this source of coverage through the guaranteed issue and renewal of individual coverage, by relaxing constraints on preexisting conditions, and through rating reforms such as rate bands and community rating. However, the results of such reform efforts across states are decidedly mixed, with the individual market in several states experiencing some adverse selection, rising premiums, declining enrollment, as well as noncompliance by insurance agents and defections by insurers that may or may not be related to the reforms.²⁰ Finally, while HIPAA allows for the transition from group to individual coverage for some persons, the ability of certain near-elderly workers to meet such criteria and to afford individual coverage may pose a major impediment to acquiring individual coverage.²¹

Perhaps our most compelling findings relate to near-elderly working women with health problems. Such women were especially at risk of not having employer insurance and of being uninsured. They were also more likely than those without health problems were to have incomes less than twice the poverty level, making the purchase of private insurance more difficult. Proposals to expand Medicare to uninsured near-elderly workers would clearly include this group, but the extent to which such workers would respond to such an opportunity remains uncertain. Our tabulations of MEPS data suggest that roughly 60 percent of the 2.1 million near-elderly workers uninsured all year have health problems. To approximate the average health care costs should such workers obtain coverage, we use the average health care expenditures for insured near-elderly workers with health problems. These data suggest that the uninsured would incur $5,000 on average (in 1996 dollars), nearly twice that of insured near-elderly workers in excellent or very good health. Since premiums for a Medicare buy-in would likely reflect such costs, some near-elderly workers most at risk to lack coverage may face premiums that represent a significant share of their incomes and may require monetary assistance. As Dennis Shea and colleagues have observed, targeting adults with limited economic means is consistent with recent incremental efforts to expand access to coverage for children through Medicaid and the State Children’s Health Insurance Program (SCHIP).²² Thus, the issue of ability to pay for coverage, especially for groups such as the near-elderly with health problems, is likely to figure prominently in any effort to implement a Medicare buy-in program targeted at this age group.

Editor's Notes

 
Alan Monheit is director and Jessica Vistnes is a senior economist at the Division of Social and Economic Research, Center for Cost and Financing Studies, Agency for Healthcare Research and Quality (AHRQ). John Eisenberg is director of AHRQ.

The views in this paper are those of the authors, and no official endorsement by the Agency for Healthcare Research and Quality or by the Department of Health and Human Services is intended or should be inferred. The authors thank Joel Cohen, Steven Cohen, and John Moeller for helpful comments and Suzanne Worth of Social and Scientific Systems, Bethesda, Maryland, for expert data processing support.

NOTES

Top Data And Methods Health Insurance Status As... Near-Elderly Workers At Risk... Discussion And Policy... NOTES

 

1. See, for example, G. Jensen, "Health Insurance and the Near Elderly: A Growing Concern," Medical Care 36, no. 2 (1998): 107–109.
2. Data from the 1996 MEPS reveal that male workers ages 55–64 spent $3,078 per capita on health, nearly five times the amount spent by male workers ages 21–29 ($625), and more than twice that spent by prime-age male workers ages 30–54 ($1,257). Per capita expenditures for female workers ages 55–64 ($2,707) were more than twice those of female workers ages 21–29 ($1,263) and one and a half times those of female workers ages 30–54 ($1,765). MEPS data also indicate that near-elderly workers are more likely to have diabetes, ischemic heart disease, and high blood pressure/hypertension than are prime-age workers.
3. See J. Rogowski and L. Karoly, "Health Insurance and Retirement Behavior: Evidence from the Health and Retirement Survey," Journal of Health Economics (July 2000): 529–540; J. Rogowski and L. Karoly, "The Effect of Access to Post-Retirement Health Insurance on the Decision to Retire Early," Industrial and Labor Relations Review (October 1994): 103–123; J. Gruber and B. Madrian, "Health Insurance Availability and Retirement Decisions," American Economic Review 85, no. 4 (1995): 938–948; A. Gustman and T. Steinmeier, "Employer-Provided Health Insurance and Retirement Behavior," Industrial and Labor Relations Review (October 1994): 124–140; and F. Scott, M. Berger, and J. Garen, "Do Health Insurance and Pension Costs Reduce the Job Opportunities of Older Worker?" Industrial and Labor Relations Review (July 1995): 775–791.
4. For example, see N. Brennan, Health Insurance Coverage of the Near-Elderly, Urban Institute New Federalism Project, Series B, no. B-21 (Washington: Urban Institute, July 2000); L. Pol, K. Mueller, and P. Adidam, "Health Insurance in the Near-Elderly Population," Population Research and Policy Review 19, no. 2 (2000): 97–112; P. Ostuw, "Health Insurance Coverage and the Near-Elderly," EBRI Research Notes (March 1999): 4–8; F. Sloan and C. Conover "Life Transitions and Health Insurance Coverage of the Near Elderly," Medical Care 36, no. 2 (1998): 110–125; and G. Jensen, "The Dynamics of Health Insurance among the Near Elderly," Medical Care 30, no. 7 (1992): 598.
5. For a discussion of COBRA, see Health Insurance and the Uninsured: Background Data and Analysis (Washington: Congressional Research Service, 1988); and M. Field and H. Shapiro, eds., Employment and Health Benefits: A Connection at Risk (Washington: National Academy Press, 1993). For a discussion of HIPAA legislation, see L.M. Nichols and L.J. Blumberg, "A Different Kind of ‘New Federalism’? The Health Insurance Portability and Accountability Act of 1996," Health Affairs (May/June 1998): 25–42. For details of President Clinton’s proposal, see J. Broder, "Clinton Proposes Opening Medicare to Those 55 to 65," New York Times, 7 January 1998, A1; and "An Analysis of the President’s Budget Proposal for Fiscal Year 2001" (Washington: Congressional Budget Office, April 2000).
6. See W. Edwards and M. Berlin, Questionnaire and Data Collection Methods for the Household Survey, National Medical Expenditure Survey Methods, Pub. no. 89-3450 (Rockville, Md.: Agency for Health Care Policy and Research, 1989); and J. Cohen, Design and Methods of the Medical Expenditure Panel Survey Household Component, MEPS Methodology Report no. 1, AHCPR Pub. no. 97-0026 (Rockville, Md.: Agency for Health Care Policy and Research, 1997).
7. R. Weinick and A. Monheit , "Children’s Health Insurance and Family Structure, 1977–1996," Medical Care Research and Review (March 1999): 55–73.
8. Among all workers in both years, young adults ages 21–29 were less likely to have employer coverage and were roughly twice as likely as the near-elderly and prime-age workers were to lack health insurance.
9. While near-elderly male workers also experienced a decline in rates of privately purchased coverage, the decline was only significant for those male workers ages 55–59.
10. Significant at the p < .10 level.
11. P. Cooper and B. Schone, "More Access, Fewer Takers for Employer-Provided Health Insurance: 1987 and 1996," Health Affairs (Nov/Dec 1997): 142–149.
12. As Cooper and Schone report, access to coverage through either spouse’s job (that is, family access rates) remained stable for the near-elderly over the past decade. Given this stability, our focus on individual offer rates allows us to examine who among the near-elderly experienced a change in offer rates.
13. We also examined the stability of the health insurance status of insured near-elderly workers and insured prime-age workers (that is, the extent to which each group retained their coverage over the year) and we found no difference between the groups.
14. Tabulations from the 1996 MEPS.
15. Sample-size considerations necessitated grouping individuals reporting good health with those reporting fair or poor health. Our data indicate that the results reported regarding insurance status are dominated by the experience of workers in fair or poor health. Since the categories for perceived health status in the 1987 NMES were expanded to include "very good" health in the 1996 MEPS, estimates for self-reported health status may not be directly comparable across surveys. As a result, we focus on cross-section results for near-elderly men and women from the 1996 MEPS.
16. Data from the 1996 MEPS.
17. Sample-size considerations precluded a separate analysis of offer and take-up rates for women ages 60–64 in good, fair, and poor health. Consequently, we examine estimates for such workers ages 55–64.
18. See, for example, the papers on the individual health insurance market published in the Journal of Health Politics, Policy and Law (February 2000).
19. U.S. General Accounting Office, Private Health Insurance: Declining Employer Coverage May Affect Access for 55 to 64-Year-Olds, GAO/HEH-98-133 (Washington: GAO, June 1998).
20. See papers by K. Swartz and D. Garnick, M. Hall, and A. Kirk describing state experience with individual market health reform in the Journal of Health Politics, Policy and Law (February 2000). Note also that Sloan and Conover, "Life Transitions," conclude that state policies to expand access to private insurance were not effective in preventing the near-elderly from losing coverage or in improving their prospects of acquiring coverage.
21. Nichols and Blumberg, "A Different Kind of ‘New Federalism’?"
22. D.G. Shea, P.F. Short, and M.P. Powell, "Betwixt and Between: Targeting Coverage Reforms to Those Approaching Medicare," Health Affairs Jan/Feb 2001): 219–230.

                      What's this?

This article has been cited by other articles:

				B. Han, R. B. Tiggle, and R. E. Remsburg Characteristics of Patients Receiving Hospice Care at Home Versus in Nursing Homes: Results From the National Home and Hospice Care Survey and the National Nursing Home Survey American Journal of Hospice and Palliative Medicine, January 1, 2008; 24(6): 479 - 486. [Abstract] [PDF]

				D. E. Hoffmann and A. J. Tarzian Dying in America - An Examination of Policies that Deter Adequate End-of-life Care in Nursing Homes J. Law Med. Ethics, June 1, 2005; 33(2): 294 - 309. [PDF]

				J. L. WERTH Jr. and D. BLEVINS Public Policy and End-of-life Care American Behavioral Scientist, November 1, 2002; 46(3): 401 - 417. [Abstract] [PDF]

				M. B. Buntin and H. Huskamp What Is Known About the Economics of End-of-Life Care for Medicare Beneficiaries? Gerontologist, October 1, 2002; 42(90003): 40 - 48. [Abstract] [Full Text] [PDF]