Health Affairs, 24, no. 4 (2005): 1022-1031 doi: 10.1377/hlthaff.24.4.1022 © 2005 by Project HOPE	New Online   Obama's Success & Next Steps   3 Health Policy Options for President Obama   A Hybrid Proposal   Bipartisan Compromise

This Article Abstract Figures Only Reprint (PDF) Erratum Erratum (v24,p1380) 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 ISI Web of Science 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 ISI Web of Science (24) Citing Articles via Google Scholar Google Scholar Articles by Stuart, B. Articles by Limcangco, M. R. Search for Related Content PubMed PubMed Citation Articles by Stuart, B. Articles by Limcangco, M. R. Related Collections Access To Care Health Reform Legal/Regulatory Issues Managed Care - Medicare Medicare Pharmaceuticals Chronic Care Health Spending Consumer Issues Elderly

Health Tracking

MARKETWATCH

Riding The Rollercoaster: The Ups And Downs In Out-Of-Pocket Spending Under The Standard Medicare Drug Benefit

Abstract

 
This study projects how much Medicare beneficiaries who sign up for the standard Part D drug benefit in 2006 will pay in quarterly out-of-pocket payments through 2008. In the first year we estimate that about 38 percent of enrollees will hit the benefit’s no-coverage zone, known as the "doughnut hole," and that 14 percent will exceed the catastrophic threshold. Because drug spending is highly persistent over time, beneficiaries who experience the biggest gaps in coverage are likely to do so year after year, with potentially serious financial consequences.

This paper examines the financial implications of the standard Part D benefit design for Medicare beneficiaries who enroll during the first three years of the program. Our findings are based on a simulation model calibrated to Part D coverage thresholds and applied to a cohort of beneficiaries likely to enroll. We then track quarterly out-of-pocket drug spending from 2006 through 2008 for the entire cohort and two subgroups dubbed "high spenders" (those with projected 2006 drug spending above the initial benefit limit of $2,250) and "catastrophic spenders" (those with 2006 drug spending above the catastrophic threshold of $5,100). This subgroup analysis provides information for policymakers about the persistence of drug spending among those most likely to face large quarter-to-quarter changes in their personal medication costs—what we call "riding the benefit rollercoaster."

A useful way to visualize the impact of this unique benefit design on enrollees’ out-of-pocket obligations is to compute the average number of months a beneficiary would be in a coverage gap as a function of annual drug spending (Exhibit 1). Our calculations are based on two assumptions: (1) that all spending on drugs counts toward the Part D cost-sharing thresholds; and (2) that enrollees’ drug spending is constant throughout the year and does not decline during gap periods.

To continue our example, a Medicare beneficiary spending $5,000 on drugs in 2006 will lack coverage for 60 percent of the year (7.2 months). By 2008 a beneficiary with the same annual drug bill will face 6.2 months in benefit gaps. Even beneficiaries whose spending exceeds the catastrophic coverage threshold will be without coverage for a considerable portion of each year, as indicated by the figures for annual spending of $10,000 (3.7 gap months in 2006, rising to 4.5 months in 2008).

Data And Methods

Top Data And Methods Results Discussion And Policy... NOTES

 
Our analyses are based on data from the 1998–2000 Medicare Current Beneficiary Surveys (MCBS), a longitudinal panel survey of a representative national sample of the Medicare population conducted by the Centers for Medicare and Medicaid Services (CMS) since fall 1991. Approximately 12,000 Medicare beneficiaries, both aged and disabled, living in the community and in institutions, are sampled from enrollment files. Respondents are surveyed three times a year for up to three years, using computer-assisted personal interviewing, about their demographic characteristics, health and functional status, Medicare supplemental insurance, and use of and spending for health services. The MCBS files link Medicare claims to survey-reported events and provide a complete source of payment data on all health care services, including prescription drugs.²

Study samples. The study sample was drawn from community-residing MCBS respondents observed in all three years from 1998 to 2000 (N = 3,370). From this cohort we selected three groups based on their characteristics in the base year (1998). First are "potential Part D enrollees," defined as those who are ineligible for subsidized coverage (having incomes above 150 percent of the federal poverty level) and have characteristics that would make Part D a reasonable option. This group includes all respondents with no drug coverage or coverage under a self-purchased Medigap policy or Medicare health maintenance organization (HMO), along with those who had any type of drug coverage for just part of the year. The group does not include beneficiaries with full-year, employer-sponsored drug coverage or Medicaid. These criteria resulted in a cohort of 1,333 potential Part D enrollees, or 40 percent of the sample pool.

We also identified a "high spender" cohort of potential Part D enrollees with projected 2006 prescription spending above $2,250, and a "catastrophic spender" cohort of high spenders with projected 2006 drug spending above $5,100. The sample sizes for these two subgroups are 510 and 191, respectively.

Spending projections. We used data from the National Health Accounts (NHA) from the CMS to forecast actual per capita drug spending reported in the MCBS to the future (1998–2006, 1999–2007, and 2000–2008).³ These spending forecasts represent the "counterfactual" to Part D—that is, what one would expect to see had the benefit not been enacted. To compute the impact of Part D on out-of-pocket spending, we imposed the Part D benefit thresholds onto the counterfactual estimates. For example, if mean per capita drug spending in 2006 for a particular group is forecast to be $3,500, we calculated the average out-of-pocket obligation under Part D as the sum of the $250 deductible, $500 in coinsurance, and $1,250 in the doughnut hole, for a total out-of-pocket payment of $2,000. Following our assumption of uniform spending throughout the year, out-of-pocket spending is $406 in the first quarter of the year, $219 in the second, $500 in the third, and $875 in the fourth. These estimates implicitly assume that annual drug spending is insensitive to any changes in the generosity of drug coverage during the year.

Results

Top Data And Methods Results Discussion And Policy... NOTES

 
Demographics. Compared with the community-residing Medicare population as a whole, our "potential Part D enrollee" cohort contains fewer disabled beneficiaries under age sixty-five, more married people, a higher proportion of whites, and more beneficiaries with higher average incomes (Exhibit 2).

Spending. The counterfactual spending estimates for the entire community-dwelling Medicare population are $3,081 in 2006 and $3,891 in 2008 (Exhibit 3). Mean drug spending by the "potential Part D enrollee" cohort is lower, at $2,608 in 2006 and $3,230 by 2008 (reflecting the sample exclusion of Medicaid recipients and beneficiaries with continuous employer-sponsored coverage). The "high spender" cohort has projected mean drug spending of $5,534 in 2006 and about the same two years later. The "catastrophic spender" cohort, on the other hand, has projected drug spending of $9,106 in 2006 but $7,328 in 2008. The trends for the two high-spending groups reflect the net effect of projected spending increases over time and regression to the mean. For the high spenders, the two effects essentially cancel each other out. There is more regression to the mean among catastrophic spenders, but even after three years, drug spending for beneficiaries in this category exceeds that of the average Part D enrollee by more than two to one. In other words, spending is highly but not completely persistent.

Effects of coverage thresholds. About 10 percent of potential Part D enrollees are projected to have zero spending in all three years (Exhibit 4). About 40 percent are forecast to have some spending in the doughnut hole each year, and about 15 percent will exceed the catastrophic threshold each year. The proportions hitting the various benefit thresholds are remarkably stable from year to year for all three groups. This is attributable in part to the persistence phenomenon already described, but it also reflects the fact that the annual increases in the statutory benefit threshold levels were designed to rise at about the same pace as anticipated increases in drug spending.

The benefit rollercoaster. We estimate that quarterly out-of-pocket payments for a Part D enrollee with average prescription spending will vary from a low of $163 in the second and third quarters of 2006 to a high of $590 in the fourth quarter of 2007—some variation, to be sure, but not much of a benefit rollercoaster (Exhibit 5). By contrast, the average beneficiary in the high spender cohort is projected to face quarterly out-of-pocket payments that vary from $401 in the second quarter of 2008 to $1,391 in the third and fourth quarters of that year. Finally, the average catastrophic spender is projected to face out-of-pocket obligations that range from a high of $2,276 in the second quarter of 2006 to a low of $95 in the fourth quarter of 2007. In other words, the amplitude of the benefit roller-coaster increases with average drug spending, and the patterns persist over time.

View larger version (25K): [in this window] [in a new window]   EXHIBIT 5 Projected Mean Out-Of-Pocket Spending On Prescription Drugs By Medicare Beneficiaries, By Level Of Spending, Per Calendar Quarter, 2006–2008

Top Data And Methods Results Discussion And Policy... NOTES

Second, because drug spending is highly persistent, enrollees will tend to face the same cost-sharing obligations from one year to the next. This is a plus for beneficiaries with relatively low drug spending, most of whom will face a constant 25 percent coinsurance rate after meeting the annual Part D deductibles. But for typical beneficiaries with high or catastrophic drug spending, it means that Part D will not pay for a sizable portion of their annual drug bills. Indeed, we estimate that the average high spender will accumulate a total out-of-pocket obligation of almost $11,000 during 2006–2008; an average catastrophic spender will accumulate $12,300. To put these numbers in perspective, the median income of households headed by someone age sixty-five or older in 2001 was $23,118.⁶ Assuming that income growth for this population segment follows historical trends, we estimate that if just one member of such a household with median income was a high spender as we have defined it, then that household would spend 12.6 percent of its income on out-of-pocket drug costs under Part D during 2006–2008.⁷ A household containing a catastrophic spender would spend 14.2 percent of its income on out-of-pocket drug costs. Under these circumstances, the combined out-of-pocket obligation for an aged couple under Part D could easily approach 20 percent of household income.

Third, beneficiaries with high and persistent drug spending can expect to incur steep and recurrent fluctuations in out-of-pocket drug costs over time. We do not know how this will ultimately affect beneficiaries’ health and well-being, but it is certainly a risk factor that deserves careful monitoring.

Caveats. When interpreting these results, it is important to keep in mind the simplifying assumptions upon which they are based. First, we assumed that all drugs are covered by Part D. MMA specifies that drug plans are obligated to cover only those products on their formularies and that off-formulary prescriptions do not count toward meeting beneficiaries’ Part D deductible and catastrophic coverage limits.⁸ Occasional off-formulary use is unlikely to have any major impact on our projections, but sustained use of costly off-formulary medicines could further raise beneficiaries’ out-of-pocket spending.

Our second assumption is that beneficiaries have uniform drug spending throughout the year. We recognize that individual beneficiaries are likely to have peaks and valleys in the use of medications used to treat acute conditions, but at least 60 percent of the medications consumed by people over age sixty-five are for chronic conditions that require continuous medication.⁹ People who adhere to such drug regimens tend to have stable utilization patterns. However, this stability may be affected by how beneficiaries respond to the financial incentives incorporated in the Part D benefit design.

The incentives cut both ways. Beneficiaries who obtain drug coverage for the first time under Part D (about 40 percent of the potential Part D enrollee pool, based on our estimates) can be expected to increase their overall use of drugs in response to the lower prices under the basic benefit.¹⁰ Those who shift from other sources of drug coverage (Medigap plans, for example) may also achieve net savings but could face higher effective prices during certain times of the year because of the Part D cost-sharing provisions.

The real unknown is how "high spending" beneficiaries will behave during extended periods in which they face full prescription drug prices. This is difficult to predict, because the doughnut hole is a new structure in health benefit design that lacks direct precedent. The closest parallel could be annual benefit caps—a common feature in many Medicare HMO plans. Beneficiaries with capped drug benefits begin each calendar year with coverage (such plans have typically not imposed deductibles), lose it after hitting the benefit ceiling, but then regain it at the start of the following year. Two recent surveys of Medicare HMO enrollees in plans with caps between $750 and $1,500 a year found that beneficiaries cut back their medications after exceeding the cap. In one study, those who exceeded the plan caps were three times more likely to reduce dosage or discontinue a medication altogether compared with those whose spending stayed under the cap.¹¹ The other study found that beneficiaries decreased their use of essential medications during gaps in coverage.¹² These findings cannot be directly generalized to the national experience of beneficiaries under the standard Part D benefit design, but they strongly suggest that some beneficiaries will cut back even essential medications while in the doughnut hole. The ironic effect of such behavior, if it does occur, is that it will extend the time that these beneficiaries are without coverage.

As a final point, we note that our out-of-pocket spending estimates do not include premiums. Part D monthly premiums are projected to be approximately $35 per month in 2006, $37 in 2007, and $41 in 2008.¹³ For beneficiaries who formerly had no drug coverage, these payments represent new outlays that must be factored against the benefits offered by Part D.

Winners and losers. For many Medicare beneficiaries, the cost sharing associated with Part D will represent an improvement over what they have now. This is clearly true for most of the estimated nine million beneficiaries who have no drug coverage at all or who qualify for low-income subsidies that fill the Part D gaps.¹⁴ The basic Part D benefit is also better than the standard Medigap policies H, I, and J or current Medicare Advantage (MA) plans with low drug benefit caps.¹⁵ On the other hand, MMA will increase the out-of-pocket burden for some beneficiaries. In scoring the cost of Part D, the CBO estimated that 17 percent of beneficiaries with retiree benefits would lose them because of employer pull-backs in response to MMA, and more recent estimates by the Department of Health and Human Services (HHS) suggest that the actual number might be twice that high.¹⁶

Our findings show that winning or losing under Part D is also a function of how much beneficiaries spend on drugs and the persistence of that spending over time. Those who spend deep into the doughnut hole or hit the Part D catastrophic threshold in one year are likely to do so year after year. The cumulative amount of time they have meaningful protection from high out-of-pocket payments (that is, being within the catastrophic coverage zone) is thus relatively short compared with the time they spend in the doughnut hole with no coverage at all.

Adding protections. Can the Part D benefit be restructured to better meet the needs of these beneficiaries? The usual criteria for judging optimal insurance benefits are not particularly helpful here, because Part D is not insurance in any traditional sense. The value of true insurance flows from risk reduction, but the fact that Medicare beneficiaries’ drug usage patterns are highly persistent means that there is little risk to insure, at least in the short run.¹⁷ Rather, the protection that beneficiaries seek is relief from obligations that are largely known beforehand. Viewed from this perspective, the two main functions of cost sharing are, first, to assure that total program outlays fall within budgetary projections and, second, to fairly apportion financial rewards among groups of beneficiaries with very different expected outlays for medications.¹⁸

These two roles need not conflict. For example, we calculated that a uniform coinsurance rate of about 56 percent would result in program spending equal to the amount projected for the standard Part D benefit, but it would treat low spenders and high spenders proportionately the same. If fairness is judged on the basis of the share of income devoted to out-of-pocket payments, then alternative cost-sharing structures merit consideration.¹⁹ Although income-based coinsurance is impractical, the system could be improved by replacing the current catastrophic threshold with a cap on the proportion of beneficiaries’ household income paid out of pocket for drugs. How low the cap should be is open to debate, but it clearly should be set below the 12 percent or more of income that we project a significant share of Medicare beneficiaries will face in the first three years under Part D.

There are less radical options for protecting beneficiaries from high out-of-pocket costs. Private drug plans can be expected to offer discounted drug prices to enrollees during gap periods. However, if experience with the Medicare drug discount cards is any guide, the savings are likely to be variable and selective; also, discounts will not reduce beneficiaries’ out-of-pocket liability for reaching the deductible and catastrophic thresholds.²⁰ Real protection requires that Congress repeal the provision under current law that penalizes drug plans for filling in the gaps.²¹ Last, but not least, Congress and the CMS should consider a major public education campaign to help beneficiaries plan for the inevitable fluctuations in financial obligations they will face under the Part D rollercoaster.

Because of space considerations, we have not discussed the effects of fluctuating coverage on beneficiaries’ health. This important aspect of coverage reform should be the subject of future research, to ensure that the health of America’s elders is not endangered as an unintended consequence of MMA.

Editor's Notes

 
Bruce Stuart (BStuart{at}rx.umaryland.edu) is professor and director of the Peter Lamy Center on Drug Therapy and Aging at the University of Maryland, Baltimore. Becky Briesacher is an assistant professor at the University of Massachusetts Medical School, Division of Geriatric Medicine, in Worcester. Dennis Shea is professor and chair of the Department of Health Policy and Administration at the Pennsylvania State University in University Park. Barbara Cooper is former director of the Program on Medicare’s Future at the Commonwealth Fund in New York City. Fatima Baysac is a doctoral candidate in the Department of Pharmaceutical Health Services Research at the University of Maryland, Baltimore. Rhona Limcangco is a research analyst at the Agency for Healthcare Research and Quality in Rockville, Maryland.

The authors thank Thomas Shaffer and two reviewers for their helpful comments on an earlier draft. Funding for this study was provided by the Commonwealth Fund. The views expressed here are those of the authors and do not necessarily reflect those of the Commonwealth Fund or its directors, officers, or staff.

NOTES

Top Data And Methods Results Discussion And Policy... NOTES

 

1. J. Mays et al., Estimates of Medicare Beneficiaries’ Out-of-Pocket Drug Spending in 2006: Modeling the Impact of the MMA (Menlo Park, Calif.: Henry J. Kaiser Family Foundation, November 2004).
2. A pharmacy validation study conducted by the CMS indicates that prescription drug spending reported in the MCBS is underreported by 17 percent.See J.A. Poisal, "Reporting of Drug Expenditures in the MCBS," Health Care Financing Review 25, no. 2 (2003): 23–36.
3. Centers for Medicare and Medicaid Services, "Table 11: Prescription Drug Expenditures; Aggregate and Per Capita Amounts, Percent Distribution, and Average Annual Percent Change by Source of Funds: Selected Calendar Years 1990–2013," 17 September 2004, www.cms.hhs.gov/statistics/nhe/projections-2003/t11.asp (18 April 2005). We began by inflating beneficiaries’ reported drug spending in each baseline year by 17 percent to reflect underreporting (see Note 2). We then took NHA estimates for each pair of years (for example, $317 in 1998 and $884 in 2006) to determine the unadjusted increase in percentage terms (279 percent). Because changes in drug spending for the overall population may differ from changes specific to the Medicare population, we then adjusted the NHA projections by an estimate of the historical difference in per capita spending trends in the MCBS compared with the NHA from 1993 to 2000. We performed a second adjustment to remove the portion of NHA projected spending increase associated with demand inducement due to improved drug coverage. There is no evidence of any increase in drug coverage among Medicare beneficiaries after 2000, nor is any improvement expected until Part D coverage begins in 2006. A final adjustment accounted for the fact that not all segments of the Medicare population can be expected to increase drug spending at the same rate. Our analysis of MCBS trend data indicates that per capita increases in drug spending from 1993 to 2000 followed a U-shaped pattern, with higher spending growth at the top and bottom of the income distribution. To adjust for this, we estimated how much spending increased by each income category as a percentage of the overall spending growth in the MCBS over this period. These ratios were multiplied by the base inflation factor to create income-adjusted inflation factors for 2006–2008. We then multiplied each sample beneficiary’s actual drug spending in 1998, 1999, and 2000 by the appropriate adjusted inflation factor to produce the spending projections.
4. B. Stuart et al., "Patterns of Outpatient Prescription Drug Use among Pennsylvania Elderly," Health Care Financing Review 12, no. 3 (1991): 61–72; N.E. Coulson and B. Stuart, "Persistence in the Use of Pharmaceuticals by the Elderly: Evidence from Annual Claims," Journal of Health Economics 11, no. 3 (1992): 315–328[CrossRef][ISI][Medline]; and M.V. Wrobel et al., "Predictability of Prescription Drug Expenditures for Medicare Beneficiaries," Health Care Financing Review 25, no. 2 (2003): 37–46.
5. Were the basic benefit thresholds to remain the same from year to year, we estimated that 85 percent of high spenders in 2006 would remain in the doughnut hole in 2007 and 2008 (as opposed to the 82 percent reported in Exhibit 4). Likewise, were the catastrophic threshold held constant at 2006 levels, we estimated that 69 percent of catastrophic spenders would exceed the threshold in 2007 (rather than 65 percent), and 63 percent would still exceed it in 2008 (as opposed to the 53 percent shown in the exhibit).
6. U.S. Census Bureau, "Current Population Reports," in Statistical Abstract of the United States: 2003, Table no. 685, www.census.gov/prod/2004pubs/03statab/income.pdf (15 May 2004).
7. According to MCBS data, beneficiaries’ mean annual income grew 3.8 percent per year from 1993 through 2000. Assuming that this growth rate persists, we projected that the median income of aged households will reach $27,857 in 2006, $28,916 in 2007, and $30,014 in 2008.
8. In the parlance of MMA, off-formulary purchases by beneficiaries do not count as true out-of-pocket costs.
9. C. Roe, A. McNamara, and B. Motheral, "Use of Chronic Medications among a Large, Commercially-Insured U.S. Population," Pharmacoepidemiology and Drug Safety 11, no. 4 (2002): 301–309.[Medline]
10. See, for example, M.V. Pauly, "Medicare Drug Coverage and Moral Hazard," Health Affairs 23, no. 1 (2004): 113–122[Abstract/Free Full Text]; L.A. Lillard, J. Rogowski, and R. Kington, "Insurance Coverage for Prescription Drugs: Effects on Use and Expenditures in the Medicare Population," Medical Care 37, no. 9 (1999): 926–936[CrossRef][ISI][Medline]; and J. Blustein, "Drug Coverage and Drug Purchases by Medicare Beneficiaries with Hypertension," Health Affairs 19, no. 2 (2000): 219–230.[Abstract]
11. E.R. Cox et al., "Medicare Beneficiaries’ Management of Capped Prescription Benefits," Medical Care 39, no. 3 (2001): 296–301.[CrossRef][ISI][Medline]
12. C.W. Tseng et al., "Cost-Lowering Strategies Used by Medicare Beneficiaries Who Exceed Drug Benefit Caps and Have a Gap in Coverage," Journal of the American Medical Association 292, no. 8 (2004): 952–960.[Abstract/Free Full Text]
13. Congressional Budget Office, Letter to the Honorable Don Nickles providing "additional information about CBO’s cost estimate for the conference agreement on H.R. 1," 20 November 2003, www.cbo.gov/showdoc.cfm?index=4814&sequence=0 (18 April 2005).
14. Mays et al., Estimates.
15. The typical MA policy has an annual benefit cap of less than a $1,000. See CMS, "Medicare+ Choice: Policy Concerns, Implications, and Prescription for Change," 16 September 2004, www.cms.hhs.gov/healthplans/benpkg/mc00anal.asp (18 April 2005).
16. CBO letter to Don Nickles; and R. Pear, "Medicare Law Is Seen Leading to Cuts in Drug Benefits for Retirees," New York Times, 14 July 2004.
17. There is, of course, greater risk over the long run as even healthy beneficiaries face the prospect of deteriorating health status and growing need for pharmaceutical therapy.
18. See J. Nyman, The Theory of Demand for Health Insurance (Stanford, Calif.: Stanford University Press, 2003).
19. It is worth noting that high spenders also have above-average incomes, which might justify some of the differential in out-of-pocket payments between low and high spenders.
20. The CMS has recently signaled that it will permit MA plans to use any Medicare rebates they receive for providing low-cost medical benefits to fill in the doughnut hole without risk of losing reinsurance payments. However, this ruling affects MA plans and not those offering only drug benefits.See "CMS: MA Plans Can Fill Drug Benefit Gap with Rebates from Low Bids," Inside CMS 8, no. 4 (2005): 1–6.
21. T. Docksum and A. Hassol, Evaluation of the Medicare-Approved Prescription Drug Discount Card and Transitional Assistance Program, Final Report to CMS by Abt Associates (Washington: Abt Associates, 1 December 2004).

                      What's this?