This Article Abstract Submit a response to this article Services E-mail this article to a friend Alert me to new issues of the journal

H E A L T H  T R A C K I N G : T R E N D S
W E B  E X C L U S I V E


27 February 2002

Trends In Medicare Supplemental Insurance
And Prescription Drug Coverage, 1996-1999

Beneficiaries in the highest-earning group saw their supplemental coverage rates
erode more than was true for their lower-earning peers.

by Mary A. Laschober, Michelle Kitchman, Patricia Neuman, and Allison A. Strabic

ABSTRACT

Medicare Current Beneficiary Survey (MCBS) Access to Care data indicate a five-percentage point decline in the share of Medicare beneficiaries having Medigap coverage between 1996 and 1999; this was matched by a commensurate rise in the share enrolled in Medicare HMOs, contributing to an increase in the percentage with drug coverage. During this period, high-income beneficiaries, and to a lesser extent healthier and rural beneficiaries, experienced greater net declines in supplemental coverage and smaller relative gains in drug coverage, compared with others. By fall 1999, 38 percent of beneficiaries lacked drug coverage, based on point-in-time estimates. This is much higher than previous estimates that measured beneficiaries' drug coverage at any time during the calendar year. Many of Medicare's most vulnerable beneficiaries–rural (50 percent), near-poor (44 percent), and oldest old (45 percent)–were most likely to lack drug coverage in the fall of 1999.

Medicare has been, and continues to be, a vital and reliable source of health insurance for older Americans and for many younger adults with permanent disabilities. However, Medicare's high cost-sharing requirements and gaps in coverage-particularly the lack of an outpatient prescription drug benefit-have led a majority of beneficiaries to seek supplemental coverage through private and public sources. There is some concern that rising Medigap premiums, eroding employer-sponsored retiree health coverage, and unstable Medicare+Choice (M+C) markets are limiting beneficiaries' access to supplemental coverage, intensifying the national debate for Medicare benefit improvements.

This paper examines trends in Medicare beneficiaries' supplemental insurance and prescription drug coverage between 1996 and 1999, the most recent years for which data are available. Building on the research of Nadereh Pourat and colleagues, who used the 1996 Medicare Current Beneficiary Survey (MCBS) Access to Care files to examine socioeconomic differences in supplemental coverage, this paper also focuses on coverage differences among traditionally vulnerable subgroups of Medicare beneficiaries in 1999.¹ Recognizing that health insurance markets have continued to change rapidly since 1999, the paper concludes with a discussion of the future viability of supplemental insurance markets to meet beneficiaries' health care financing needs. The findings from this study provide important context for discussions of current policy proposals to reform Medicare and extend prescription drug coverage to the entire Medicare population.

Data And Methods

This study uses data from the 1996, 1997, 1998, and 1999 MCBS Access to Care files. The data provide point-in-time estimates of beneficiaries' self-reported supplemental insurance and prescription drug coverage. These estimates differ from the more commonly reported calendar-year estimates of coverage derived from the MCBS Cost and Use files and provide an alternative measure of coverage in a given year.² The Access to Care files assign a source of coverage to individuals based on their insurance status at the time the MCBS survey is conducted in the fall of each year. By contrast, studies based on the Cost and Use files, which are constructed from the MCBS survey administered to respondents three times during the year, generally assign coverage to an individual if he or she had that coverage at any time during the year.³ As a result, the Access to Care files generally produce lower estimates of supplemental insurance and drug coverage rates than the Cost and Use files do, primarily because of beneficiaries' gaps in coverage during the course of the year. Bruce Stuart and colleagues found, for example, that nearly 17 percent of beneficiaries had gaps in their prescription drug coverage at some point during 1995, and nearly one-third had gaps at some point during 1995-1996.⁴

Some beneficiaries have more than one source of supplemental coverage. Thus, for analytic purposes, beneficiaries were counted in only one of the following categories (in hierarchical order for beneficiaries with more than one type of supplemental insurance): current employer-sponsored insurance, Medicaid, Medicare health maintenance organization (HMO), employer-sponsored retiree insurance, individually purchased coverage (referred to as Medigap), "other" public insurance, or traditional fee-for-service (FFS) Medicare only.⁵ Medicare HMOs are considered supplemental coverage in this analysis because they generally provide benefits in addition to those covered by FFS Medicare.

To help ensure that all sources of drug coverage are accounted for, we assigned drug coverage to beneficiaries who reported having drug benefits from any insurance source at the time of the interview, but we assigned that coverage to their primary source of supplemental insurance.⁶

This study presents the results of bivariate statistical tests to convey important variations among beneficiary subgroups in 1999 (results from difference-in-proportions tests) and to indicate key trends from 1996 to 1999 as well as deviations from these overall trends among beneficiary subgroups (results from difference-of-differences-in-proportions tests). Standard errors for significance tests were generated from software specifically designed to account for the complex survey design of the MCBS. Logistic regression analyses also were conducted to identify beneficiary characteristics most closely associated with alternative sources of supplemental coverage and drug benefits in 1999. All trend and comparative statistics highlighted in the findings are statistically significant at the 95 percent confidence level or higher (although not all results shown in the exhibits are statistically significant).

Study Findings

Trends in supplemental insurance coverage, 1996-1999. From 1996 to 1999 almost nine of ten Medicare beneficiaries had supplemental coverage in the fall of the year, and this ratio remained virtually constant over the four-year period. However, notable changes took place in the distribution of sources of supplemental coverage during that time. The most dramatic shift occurred between beneficiaries enrolled in a Medicare HMO and those with a Medigap policy: Growth in the former more than offset the decline in the latter (Exhibit 1). Medicare HMO enrollment rose by approximately two million beneficiaries (+5.1 percentage points), from 4.1 million in 1996 to 6.0 million in 1999, while Medigap policyholders declined by 1.5 million (-5.1 percentage points), from 9.9 million in 1996 to 8.4 million in 1999. The share of beneficiaries with supplemental benefits from other sources remained generally constant over the period, as did the share who lacked any form of supplemental coverage.

The shift between Medicare HMO and Medigap coverage rates observed between 1996 and 1999 did not occur to the same extent for beneficiaries in the highest income group (more than $30,000). This group experienced both declines in Medigap (-4.2 percentage points) and employer coverage (-1.7 percentage points) that were not offset by matching gains in Medicare HMO enrollment (+3.1 percentage points). Consequently, the group with the highest income had a larger net decline in supplemental coverage than was true for groups earning less (Exhibit 2).

Rural beneficiaries and those who reported better health also experienced modest net declines in supplemental coverage over the four-year period–possibly explained by their Medicare HMO enrollments' failing to make up for declines in employer-sponsored or Medigap coverage or both. The poorest beneficiaries made gains in supplemental coverage during this period, in part because they had only relatively small declines in Medigap and had gains across all other sources.

Trends in outpatient prescription drug coverage, 1996-1999. The share of beneficiaries who had drug coverage increased by 5.5 percentage points between 1996 and 1999, from 56.8 percent to 62.3 percent.⁷ This increase appears to be a direct reflection of the shift in coverage from Medigap to Medicare HMOs because Medicare HMO enrollees were far more likely to have drug coverage than were Medigap policyholders. In 1999, for example, 88 percent of Medicare HMO enrollees reported having drug benefits, compared with only 27 percent of Medigap policyholders.

While nearly all beneficiary groups saw important net gains in drug coverage over the period, not all groups had the same increase (Exhibit 2). Most notably, those with the highest incomes had virtually no net increase in drug coverage. As discussed above, the highest income group experienced declines in both Medigap and employer-sponsored coverage, without commensurate increases in Medicare HMO enrollment, which appears to explain this pattern.

Beneficiaries who reported having no chronic conditions also had more moderate net gains in drug coverage rates than others did. Beneficiaries without chronic conditions had greater rates of decline in Medigap coverage and lower rates of enrollment in Medicare HMOs than was true for beneficiaries with one to four chronic conditions. Although persons with five or more chronic conditions also joined Medicare HMOs at relatively low rates, they appear to have gained drug benefits primarily through increased Medicaid coverage.

Rural beneficiaries also had relatively small gains in drug coverage, experiencing only a 3.2-percentage-point increase over the period. Their modest gains were likely attributable to limited access to Medicare HMOs and some decline in Medigap coverage.

Snapshot of supplemental insurance coverage, 1999. Lack of supplemental insurance. Although almost 88 percent of Medicare beneficiaries had some type of supplemental insurance in the fall of 1999, 4.3 million beneficiaries (12.5 percent) had only traditional FFS Medicare when the survey was conducted. This figure contrasts with the 2.6 million beneficiaries (6.8 percent) with no supplemental coverage in 1998 reported by John Poisal and Lauren Murray. Their estimates apply to beneficiaries who had no supplemental coverage during all of 1998, while our estimates apply to beneficiaries with no supplemental coverage at the time of their fall 1999 interview. The difference between estimates suggests that large numbers of beneficiaries have gaps in coverage at some point during the year.

A disproportionately large share of beneficiaries who were under age sixty-five and disabled or were black lacked supplemental coverage in the fall of 1999 (Exhibit 3).
Discrepancies in coverage also were strongly evident by income, rural residence, health status, and gender.

When health is measured by the number of chronic conditions a person has, the number of reported chronic conditions raised a beneficiary's likelihood of having supplemental insurance. This appears to be explained by these beneficiaries' higher rates of Medicaid coverage (Exhibit 4).

Multivariate methods to examine the relationship between beneficiaries' characteristics and the absence of supplemental coverage found that age, race/ethnicity, income, urban/rural status, gender, self-reported health status, and number of chronic conditions were all statistically significant and in the same directions as the bivariate findings. Black non-Hispanic beneficiaries, for example, were 2.1 times as likely as white non-Hispanics were to lack supplemental coverage; beneficiaries under age sixty-five and disabled were almost twice as likely as were those ages sixty-five to seventy-four to have no supplemental insurance in 1999; and beneficiaries in the most poor ($10,000 or less) and near-poor ($10,001-$20,000) income groups were 2.6 times as likely as beneficiaries in the highest income group were to be without any source of supplemental insurance.

Sources of supplemental insurance. One-third of all beneficiaries had supplemental coverage through an employer in the fall of 1999 (Exhibit 4).⁸ One-fourth of all Medicare beneficiaries had a Medigap policy, and one-sixth were enrolled in a Medicare HMO. Another 11 percent had supplemental coverage through Medicaid. Nearly 2 percent of beneficiaries reported having supplemental coverage through other public programs.

Sources of supplemental insurance varied by beneficiary characteristics: Beneficiaries who were under age sixty-five and disabled were far less likely than were those age sixty-five or older to have a Medigap policy, to be enrolled in a Medicare HMO, or to have employer-sponsored retiree insurance. Instead, younger disabled beneficiaries relied much more heavily on Medicaid to supplement Medicare. Rural beneficiaries were more likely than urban beneficiaries were to hold a Medigap policy and much less likely to be in a Medicare HMO. This is largely explained by rural beneficiaries' limited access to Medicare HMOs.

Black non-Hispanic beneficiaries were much less likely than their white counterparts were to have employer-sponsored or Medigap coverage, again depending more heavily on Medicaid. Beneficiaries reporting Hispanic or "other races and ethnicities" were somewhat more likely than black beneficiaries were to have Medigap, Medicare HMO, and Medicaid coverage.

The lowest income group and those in poorest self-reported health, while more likely to rely on Medicaid, had much lower rates of Medigap, Medicare HMO, or employer-sponsored coverage than did their higher-income or healthier counterparts. Women relied more heavily on Medicaid than men did, although no significant differences were observed for other sources of supplemental coverage.

Snapshot of outpatient prescription drug benefits, 1999. Lack of drug benefits. Although most beneficiaries had some supplemental coverage in the fall of 1999, 37.7 percent reported having no prescription drug benefits at that time and, thus, had to pay the full cost of their outpatient drugs out of pocket (Exhibit 5). This figure contrasts with Poisal and Murray's findings that 27 percent of beneficiaries had no drug coverage throughout 1998. The primary reason for the difference most likely arises from gaps in drug coverage for a share of beneficiaries during the year.

Several groups of traditionally vulnerable beneficiaries were more likely to lack drug coverage, including beneficiaries living in rural areas, those age eighty-five and older, and the near-poor. Beneficiaries who reported no chronic health conditions were also more likely to be without drug benefits.

These findings were supported by multivariate analysis that examined the independent effect of beneficiaries' characteristics on the lack of drug coverage. For example, near-poor beneficiaries were 1.7 times as likely as were those in the highest income group to lack drug coverage; rural beneficiaries were 1.7 times as likely as were urban beneficiaries to lack drug coverage; and those age eighty-five and older were 1.6 times as likely as were those ages sixty-five to seventy-four to lack drug coverage.

The multivariate analysis also indicates that poor beneficiaries were more than twice as likely as those in the highest income group were to be without drug coverage, and black non-Hispanic beneficiaries were 1.3 times as likely as their white counterparts were to lack drug benefits. The multivariate analysis also indicated, although less strongly, that disabled beneficiaries under age sixty-five, men, and persons with fair or poor self-reported health status were less likely to have drug benefits in 1999.

Sources of prescription drug coverage. The primary sources of drug coverage in fall 1999 were employer-sponsored plans, followed by Medicare HMOs, with only 7 percent of all beneficiaries reporting drug coverage through a Medigap policy (Exhibit 6). Most beneficiaries who had an employer-sponsored plan or who were enrolled in a Medicare HMO had drug benefits in 1999 (83 percent and 88 percent, respectively), compared with only 27 percent of beneficiaries who had a Medigap policy (not shown).

Again, there were noteworthy differences in sources of drug coverage by beneficiary subgroup that help to explain differences in lack of drug benefits for these subgroups. Rural beneficiaries were far less likely than their urban counterparts were to have drug coverage from a Medicare HMO, which is explained to a large degree by their limited enrollment in Medicare HMOs. However, even when enrolled in a Medicare HMO, rural HMO enrollees were less likely than urban HMO enrollees were to have drug benefits through their HMO. Rural beneficiaries with Medigap policies were also somewhat less likely than urban beneficiaries with Medigap were to have drug benefits from their policies (not shown).

Persons age eighty-five and older were less likely than persons ages sixty-five to eighty-four were to have drug benefits from a retiree health plan or Medicare HMO, and much less likely than the under-age-sixty-five disabled were to get drug coverage through Medicaid. A relatively large share of the poorest beneficiaries had drug benefits under Medicaid, but this group was much less likely than higher-income beneficiaries were to have drug benefits from all other supplemental insurance sources. Near-poor beneficiaries had lower rates of drug coverage through Medigap and employer-sponsored plans than higher-income beneficiaries had, which were not offset by Medicaid drug benefits. Finally, beneficiaries with more reported chronic conditions (who had higher rates of Medicaid coverage) had higher rates of drug coverage than did those with few or no chronic conditions.

Discussion And Implications

The shift in supplemental insurance and prescription drug coverage between 1996 and 1999 occurred during a period characterized by both a rapid increase in the availability and popularity of Medicare HMOs and double-digit increases in Medigap premiums. Medicare payments to Medicare HMOs fostered new market entry and allowed plans to offer desirable additional benefits, such as prescription drugs, at low cost to beneficiaries.⁹ At the same time, rapidly rising Medigap premiums, especially for policies that include some drug coverage, made Medigap a less attractive option.¹⁰ Our findings suggest that Medicare HMOs may have cushioned the fall for many beneficiaries who dropped their Medigap coverage between 1996 and 1999. Medicare HMOs also appear to be largely responsible for the nearly six-percentage-point increase in the share of beneficiaries with outpatient drug coverage over the same period.

While we found no overall net change in supplemental coverage rates for the noninstitutionalized Medicare population and found net increases in prescription drug coverage, these trends did not occur consistently across all beneficiary groups. Between 1996 and 1999, for example, beneficiaries in the highest income group saw the greatest net loss in supplemental coverage and virtually no net gain in drug coverage. Beneficiaries in this income group appeared to be less likely than lower-income beneficiaries were to enroll in Medicare HMOs, which might have compensated for declining Medigap and employer-sponsored coverage. Beneficiaries in better health–as measured by excellent or good self-reported health status and fewer chronic conditions reported–also experienced somewhat higher net losses in supplemental insurance and smaller net gains in drug coverage, primarily as a result of greater net declines in Medigap coverage rates. If rising premiums drive healthier beneficiaries away from the Medigap market, adverse selection could drive up premiums and make Medigap policies increasingly unaffordable for persons with modest incomes. Rural beneficiaries were also at a disadvantage compared with those living in urban areas, most likely because they lacked the same access to Medicare HMOs.

Despite generally positive trends observed between 1996 and 1999, there is evidence of shifting coverage patterns since then. Some of these shifts present challenges for beneficiaries who need help paying for drugs and filling the gaps in Medicare's benefit package. Between December 1999 and December 2001 the number of Medicare HMOs dropped considerably, disrupting coverage for about 1.3 million Medicare beneficiaries.¹¹ Another fifty-eight plans withdrew from the program or reduced their service areas in 2002, terminating coverage for more than 536,000 beneficiaries.¹² Meanwhile, Medigap premiums have continued to climb, rising by 16 percent on average between 1998 and 2000 for the seven Medigap plans that do not cover prescription drugs and by 37 percent for plans that offer limited drug benefits.¹³

With increasingly unaffordable premiums, the Medigap market is unlikely to absorb the recent and serious declines in Medicare HMO availability. It is also improbable that employer retiree plans or Medicaid will pick up the slack. Numerous studies show trends in declining retiree coverage.¹⁴ Although the MCBS data presented in this paper suggest that employer-sponsored coverage was relatively stable during 1996-1999, declines in retiree coverage are expected in the future. Medicaid will likely remain a critical source of supplemental coverage for the lowest-income beneficiaries but is unlikely to help middle- and higher-income beneficiaries, who may lose coverage from other sources.

The uncertain future of supplemental insurance for Medicare beneficiaries threatens to increase the number of seniors and disabled adults without drug coverage, perhaps reversing the gains for many traditionally vulnerable beneficiary groups seen during 1996-1999. This uncertainty places even greater pressure on policymakers to add a prescription drug benefit to Medicare.

The authors greatly appreciate the valuable comments of Kathryn Langwell, John Poisal, Bruce Stuart, and Frank Eppig on this paper. The contents of the paper are the sole responsibility of the authors and do not necessarily represent the views of either the Henry J. Kaiser Family Foundation or Barents Group of KPMG Consulting.

NOTES

1. N. Pourat et al., "Socioeconomic Differences in Medicare Supplemental Coverage," Health Affairs (Sep/Oct 2000): 186-196.
2. See, for example, J. Poisal and L. Murray, "Growing Differences between Medicare Beneficiaries with and without Drug Coverage," Health Affairs (Mar/Apr 2001): 74-85; J. Poisal and G. Chulis, "Medicare Beneficiaries and Drug Coverage," Health Affairs (Mar/APR 2000): 248-256; and M. Davis et al., "Prescription Drug Coverage, Utilization, and Spending among Medicare Beneficiaries," Health Affairs (Jan/Feb 1999): 231-243.
3. The Access to Care files also differ from the Cost and Use files in that they reference different beneficiary populations. The Cost and Use files include all individuals who had Medicare coverage at any time during the calendar year, while the Access to Care files include only beneficiaries who were Medicare-eligible for the entire year. In addition, the Access to Care files, in contrast to the Cost and Use files, do not permit reconciliation of self-reported drug coverage with drug payment data, leading to lower estimates of drug coverage rates. Davis et al., "Prescription Drug Coverage," estimates that such reconciliation contributes between four and nine percentage points to their drug coverage estimates. However, conversations (in January 2002) with John Poisal and Frank Eppig of the Centers for Medicare and Medicaid Services (CMS) suggest that this estimate is much lower for the more current releases of MCBS data.
4. B. Stuart, D. Shea, and B. Briesacher, "Dynamics in Drug Coverage of Medicare Beneficiaries: Finders, Losers, Switchers," Health Affairs (Mar/APR 2001): 86-99.
5. Based on personal communication with Poisal and Eppig (May 2001), the "other public insurance" category in the Access to Care files primarily includes beneficiaries enrolled in state pharmacy assistance programs and, less frequently, some beneficiaries who have access to Department of Veterans Affairs (VA) or other military benefits.
6. This methodology is adopted from Poisal and Murray, "Growing Differences between Medicare Beneficiaries with and without Drug Coverage," and Poisal and Chulis, "Medicare Beneficiaries and Drug Coverage," in which each beneficiary is assigned to one primary supplemental insurance category, but the individual may or may not obtain their drug insurance from that source. Assigning drug coverage to primary supplemental insurance, rather than to the secondary source that includes the drug benefits, changes the estimates only slightly. For example, the 1999 estimate of prescription drug coverage for Medicaid beneficiaries would decline by 0.1 percentage points if the alternative method were used.
7. Again, it is important to note that these drug coverage estimates were not supplemented by drug source-of-payment data. For example, Poisal and Murray, whose drug coverage estimates are supplemented by payment information, report that 73 percent of noninstitutionalized Medicare beneficiaries had drug coverage at some point during 1998. See Poisal and Murray, "Growing Differences between Medicare Beneficiaries with and without Drug Coverage."
8. Employer-sponsored coverage may be understated and Medicare HMO enrollment overstated, as a result of MCBS data collection methods. If a beneficiary is enrolled in a Medicare HMO, he or she is coded as having Medicare HMO coverage even if the Medicare HMO is offered through an employer.
9. L. Nelson et al., "Access to Care in Medicare HMOs, 1996," Health Affairs (Mar/APR 1997): 148-156.
10.Families USA, "The Crushing Costs of Medicare Supplemental Policies," October 1996, www.familiesusa.org/media/reports/famsup.htm (30 May 2001); Weiss Ratings, Inc., "Medigap Consumers Face Erratic Price Increases," 17 April 2001, www.weissratings.com/NewsReleases/Ins_Medigap/20010416Medigap.htm (31 May 2001); and American Academy of Actuaries, Report to the National Association of Insurance Commissioners (Washington: AAA, 8 June 2000).
11. M. Gold, "Medicare+Choice: An Interim Report Card," Health Affairs (July/Aug 2001): 120-138.
12. Centers for Medicare and Medicaid Services, "CY 2002 Nonrenewal Report, by State, County, Plan," 29 September 2001, www.cms.hhs.gov/healthplans/nonrenewal/markprintedoutnew.asp (23 January 2002).
13. Weiss Ratings, Inc., "Prescription Drug Costs Boost Medigap Premiums Dramatically," 26 March 2001,www.weissratings.com/NewsReleases/Ins_Medigap/0010326Medigap.htm (31 May 2001).
14. Hewitt Associates, Retiree Health Coverage: Recent Trends and Employer Perspectives on Future Benefits, report prepared for the Henry J. Kaiser Family Foundation (Washington: Kaiser Family Foundation, October 1999); and Kaiser Family Foundation and Health Research and Educational Trust, Survey of Employer-Sponsored Health Benefits, 2001, as reported by J. Gabel et al., "Job-Based Health Insurance in 2001: Inflation Hits Double Digits, Managed Care Retreats," Health Affairs (Sep/Oct 2001): 180-186.

Mary Laschober is a senior manager in the Health Practice of Barents Group of KPMG Consulting, Inc., in Washington D.C., where she focuses on Medicare policy and beneficiary education. Michelle Kitchman is a senior policy analyst with the Henry J. Kaiser Family Foundation, and Patricia Neuman is vice-president and director of the foundation's Medicare Policy Project. Allison Strabic was a senior consultant in the Health Practice of Barents Group, focusing on Medicare managed care performance indicators.

©2002 Project HOPE–The People-to-People Health Foundation, Inc.