July 31, 2003


By Leighton Ku

PDF of this report

Related Report:
Senate Prescription Drug Bill Would Exclude Millions of Low-income Beneficiaries
HTM version | PDF version

Press Release:
HTM version  |  PDF version

View Related Analyses

If you cannot access the files through the links, right-click on the underlined text, click "Save Link As," download to your directory, and open the document in Adobe Acrobat Reader.

Contrary to longstanding Medicare principles under which all elderly and disabled beneficiaries are provided access to Medicare’s basic package of health care benefits, the prescription drug bill the Senate has approved would bar low-income beneficiaries from receiving Medicare’s prescription drug benefit if they also are enrolled in Medicaid.[1]  Such a policy would affect millions of low-income seniors and people with disabilities eligible for both Medicare and Medicaid, a group that is known as the “dual eligibles” and that includes the poorest individuals whom Medicare serves.  The Congressional Budget Office (CBO) has estimated that in 2006, when the Medicare prescription drug benefit would be implemented, this Senate provision would exclude 6.4 million low-income Medicare beneficiaries from the new drug benefit.

The millions of low-income Medicare beneficiaries who would be blocked from eligibility for the Medicare drug benefit would receive prescription drug coverage solely through Medicaid.  By contrast, all other Medicare beneficiaries would receive drug coverage through Medicare.  The differences are significant.  Many state Medicaid programs impose limits on the availability of prescription drugs in order to reduce their Medicaid expenditures.  As a consequence, many low-income Medicaid beneficiaries likely would not have access to the same range of prescription drugs as their more affluent peers.

Under the Senate bill, states would continue to bear a major share of the cost of prescription drugs for those low-income Medicare beneficiaries who also are enrolled in Medicaid.  These individuals would receive drug coverage through Medicaid, and states pay an average of 43 percent of Medicaid benefit costs.  As a consequence, the Senate bill would create a perverse incentive for some states to pare back Medicaid eligibility.[2]  By scaling back Medicaid eligibility, these states would lower their Medicaid expenditures and shift prescription drug costs for people who would lose Medicaid eligibility to Medicare, and hence to the federal government.[3]  The low-income seniors and people with disabilities who would no longer be eligible for Medicaid would receive prescription drug coverage through Medicare, but they could lose coverage for vital services that only Medicaid covers, such as home and community-based care, nursing home care, and acute care services such as vision, dental and hearing services. 

The Medicare prescription drug legislation the House of Representatives has approved contains a number of provisions that raise serious concerns, including inadequate premium and cost-sharing subsidies for low-income elderly and disabled people, dubious privatization proposals and costly tax provisions that are largely unrelated to Medicare and prescription drugs.  However, the House bill maintains the longstanding principle of the universality of Medicare benefits by ensuring that all beneficiaries, including the dual eligibles, are eligible for the Medicare drug benefit.

House and Senate conferees have the opportunity to adopt the House provision providing universal Medicare benefits, while adopting other, more beneficial provisions from the Senate bill.  Taken together, such changes would ensure that millions of the poorest Americans have access to Medicare prescription drug benefits in the future.


State-by-State Estimates

This analysis provides estimates of the number of low-income elderly and disabled Medicare beneficiaries in each state and the District of Columbia who would be denied Medicare drug benefits under the Senate bill because they are “dual eligibles.”  We provide estimates of the number of dual eligibles in each state in 1999, based on administrative data, as well as projections of the number of dual eligibles in each state in 2006 (see the table below).

The 1999 estimates are based primarily on analyses by Marilyn Ellwood and Brian Quinn of Mathematica Policy Research, Inc.[4]  Ellwood, one of the nation’s foremost experts in research about Medicaid eligibility, and her colleague Brian Quinn analyzed millions of detailed 1999 Medicaid eligibility records to identify dual eligibles.  Their analyses indicated that about 5.4 million Medicare beneficiaries, or about 15 percent of Medicare beneficiaries, were fully covered by Medicaid in 1999.  These individuals would have been denied access to the Medicare drug benefit if a policy such as that in the Senate drug bill had been in effect in 1999.

Under both the Senate and House bills, the Medicare prescription drug benefit would not be implemented until 2006.  By that time, the number of dual-eligible beneficiaries will be considerably larger.  CBO projects there will be 6.4 million such beneficiaries in 2006, a figure 19 percent greater than the estimate for 1999.  We trended the 1999 state estimates forward to generate state projections for 2006, based on both Census Bureau projections of changes in each state’s elderly population and the CBO estimate that there will be 6.4 million dual eligibles in 2006.

Under the Senate bill, thousands of low-income Medicare beneficiaries in every state would be denied the Medicare prescription drug benefit.  We project that 100,000 or more Medicare beneficiaries would be ineligible for the Medicare drug benefit (in 2006) in 23 states.  These states are:  Alabama, Arkansas, California, Florida, Georgia, Illinois, Indiana, Kentucky, Louisiana, Massachusetts, Michigan, Mississippi, Missouri, New Jersey, New York, North Carolina, Ohio, Pennsylvania, South Carolina, Tennessee, Texas, Virginia, and Wisconsin.


Technical Notes

1999 Estimates.  None of the standard Medicaid statistical reports identify the number of dual eligibles. The Ellwood/Quinn estimates are based on analysis of administrative data from the Medicaid Statistical Information System (MSIS) for the first quarter of 1999, which includes millions of eligibility records from nearly all states in the nation.[5]  Their detailed analyses help separate those dual eligibles who are “full” dual eligibles from those who have Medicaid coverage just for Medicare cost-sharing (i.e., Medicaid pays their Medicare cost-sharing charges but provides no other coverage).[6]  Under the Senate bill, people with full Medicaid coverage would be ineligible for the Medicare prescription drug benefit.  Those covered through Medicaid for Medicare cost-sharing only would be eligible for the Medicare drug benefit.

Two technical problems arise with the MSIS data.  First, in many states, the specific status of some dual eligibles cannot be determined from these data; it is not clear in these cases whether the individual was a full dual eligible or a “partial” dual eligible (i.e., an individual with Medicaid coverage only for the Medicare cost-sharing charges).  We examined data from states with a very low proportion of these “unknown” cases and found that in these states, an average of 87.5 percent of all dual eligibles were classified as full dual eligibles.  In the four states where the portion of “unknown” cases was high (Georgia, North Dakota, Ohio and Rhode Island), we applied this national percentage to impute the proportion of dual eligibles in the state that were full dual eligibles.  That is, we assumed that 87.5 percent of dual eligibles in these states were full dual eligibles.  For states with a moderate number of “unknown” cases, we assumed that all unknown cases were full dual eligibles.[7]  Using this approach, we estimated that an average of 88.8 percent of dual eligibles in these states were full dual eligibles.  This is close to the 87.5 percent estimate derived for states with few unknown cases, leading us to conclude that the level of error in our national estimates is minimal.  (These estimates are very close to those of CBO, which estimates that roughly 90 percent of dual eligibles are “full” dual eligibles.)

The Ellwood/Quinn study also lacked data on the overall number of dual eligibles in a few states (Arkansas, Hawaii, Kentucky and Pennsylvania).  To provide a rough approximation, we took data on the enrollment of aged and disabled people in Medicaid in 1999, as reported by CMS or the Kaiser Commission on the Future of Medicaid and the Uninsured, and used these data to impute estimates of the number of full dual eligibles in these states.  The estimates for states for which data have been imputed are less accurate than the estimates for the other states. 

Data on the number of Medicare beneficiaries in each state are based on CMS data on the number of Medicare Part B enrollees in each state in 1999.

2006 Projections.  To project the number of dual eligibles in each state in 2006, we trended the 1999 estimates forward, using the CBO estimate of 6.4 million full dual eligibles as a national target.  State-specific trend factors were based on Census Bureau projection of changes in the number of  persons 65 or older in each state.[8]  We prorated the growth rates the Census Bureau has projected so that the 2006 state estimates sum to 6.4 million at the national level.

The actual number of dual eligibles in each state in 2006 will vary from these estimates because of demographic, economic or policy changes that will occur over the next few years, but these projections provide a credible approximation of the number of people who would be affected when the Medicare prescription drug benefit takes effect in 2006, if the Senate’s approach of excluding full dual eligibles is adopted.

End Notes:

[1] For a more detailed discussion, see Melanie Nathanson, Edwin Park and Robert Greenstein,  “Senate Prescription Drug Bill Would Exclude Millions of Low-income Beneficiaries,” Center on Budget and Policy Priorities, July 31, 2003.  Also see Robert Reischauer, “When More Means Less,” New York Times, July 16, 2003.

[2]  As of 2001, the Medicaid eligibility criteria for seniors and people with disabilities in 19 states were broader than the federal minimum Medicaid eligibility criteria.  The federal minimum income limit for Medicaid for the elderly and disabled is only about 74 percent of the poverty line for an individual.  Most of the 19 states with broader eligibility criteria extend Medicaid coverage to elderly and disabled people with incomes between 74 percent and 100 percent of the poverty line.  National Association of State Medicaid Directors, Aged, Blind and Disabled Eligibility Survey, June 27, 2002.

[3]  States could find shifting such costs to the federal government particularly attractive because a key factor driving state Medicaid expenditures upward is the rising cost of care for seniors and people with disabilities, which is attributable in large part to gaps in Medicare coverage for services such as prescription drug costs and long-term care.  Because of these gaps in Medicare coverage, Medicaid has been bearing an increasing share of the total costs of health coverage for seniors and the disabled, while Medicare’s share has been falling over time.  See Leighton Ku, “The Medicaid-Medicare Link: State Medicaid Programs Are Shouldering a Greater Share of the Costs Of Care for Seniors and People with Disabilities,” Center on Budget and Policy Priorities, February 25, 2003.

[4] Marilyn Ellwood and Brian Quinn, “Background Information on Dual Eligibles in MSIS, FY 1999,” Cambridge, MA: Mathematica Policy Research, updated Feb. 28, 2002.

[5] These 1999 MSIS data are the most recent data that have been published to date.  The Urban Institute is now analyzing 2000 MSIS data and may release enrollment estimates at some point. 

[6] The “partial dual eligibles” include groups such as Qualified or Specified Low-income Medicaid Beneficiaries (QMBs and SLMBs). These are beneficiaries for whom Medicaid pay some or all of the Medicare cost-sharing expenses.  Individuals who receive more limited Medicaid coverage would  be eligible for Medicare drug benefits under the Senate bill.

[7] The underlying assumption is that most state data systems can typically identify partial dual eligibles, such as QMBs or SLMBs, because these groups receive far fewer benefits than full dual eligibles, but sometimes have  problems distinguishing between specific categories of full dual eligibles, who receive more similar benefits.   Thus, the unknown groups are nearly all full dual eligibles whose specific Medicaid eligibility codes are missing.

[8] Census Bureau, Population Projections for States by Selected Age Groups and Gender: 1995 to 2025, 2003.