Health Affairs, 23, no. 1 (2004): 113-122 doi: 10.1377/hlthaff.23.1.113 © 2004 by Project HOPE	New Online   House Health Reform Bill   Paying for Reform   Vetting AHIP's Report   HIV/AIDS Costs   Brief: Insurance Reform   HA Blog Top 10

This Article Abstract 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 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 HighWire Citing Articles via Web of Science (20) Citing Articles via Google Scholar Google Scholar Articles by Pauly, M. V. Search for Related Content PubMed PubMed Citation Articles by Pauly, M. V. Related Collections Insurance Coverage Medicare Pharmaceuticals Research And Technology

Medicare Challenge

Medicare Drug Coverage And Moral Hazard

Abstract

 
This paper explores the effect of more extensive drug coverage in Medicare on the use of and spending for prescription drugs and considers whether any additional use is likely to represent satisfaction of previously unmet needs or whether it represents yet more overuse. Reasonable estimates of the effect on spending strongly suggest that the spending increase will be small and that some of it will go to beneficiaries who do not face high financial barriers at present. Thus, from the viewpoint of improvements in health, national spending on drugs, or pharmaceutical firm revenues, effects are small. The effects of such programs on Medicare’s fiscal future are much more important.

The general theory that health economists apply to insurance is this: Health insurance causes people to behave differently, in terms of use of medical care and total spending, than they would behave without insurance. Insurance does more than just pay bills; it changes the amount and composition of bills. This is true even if the person has a reasonably high income and the potentially insured medical expense is small relative to a person’s wealth: People generally use more, and more costly, medical goods and services when insurance covers their cost than when it does not. To the extent that this increase in use or spending arises from the lower user price that health insurance brings about, it is called "moral hazard."³ It is generally viewed as undesirable and inefficient, because some of the use of medical care that insurance stimulates (compared with having no insurance, when a person pays the full market price) must by definition be care that is worth less to the person than its market price. If that price is close to cost, if no one other than the person is concerned about the person’s medical care use or health status, and if the person is thought to have correct information about medical benefits and side effects, care that is worth less than it costs is wasteful.⁴

Medicare prescription drug coverage would raise the number of people with drug coverage and increase the level of coverage for some of those already insured. One explicit objective of this change is to reduce "financial barriers" to beneficiaries, to facilitate the use of drugs that would not have been used in the absence of coverage. (The other strong rationale is to reduce drug spending’s displacement of other types of consumption spending by beneficiaries. In this paper I do not consider this risk protection–budget rationale.) The reason taxpayers might favor such additional use is because at least some of the stimulated use is thought to be "needed" or "appropriate" use that was previously forgone because of a cost or income barrier.⁵ Slightly more formally, there could be value perceived by others (for example, taxpayers not on Medicare) from a Medicare beneficiary’s greater use of care. In this sense, far from being inefficient, the "moral hazard" associated with such increased coverage might actually improve the attainment of social goals.⁶

However, concrete policy design and analysis require more than just the vague sentiment that at some points increasing the use of some drugs by some beneficiaries over what they would choose on their own is a good thing. Between the rate of use consumers and their doctors would choose if consumers had no insurance and the rate they would choose if drugs were completely free, there is a wide gap for almost all conditions and almost all types of beneficiaries. The fundamental design question, then, is how to get the actual rate of use to come close to the appropriate "happy medium," at which formerly unmet needs (where benefits are greater than cost) are met but overuse (where benefits are less than cost) has yet to begin.

This paper discusses these design and evaluation issues in the context of legislative proposals now being considered by Congress. I address two separate but related questions. One is descriptive: How large an increase in use and spending should the coverage being discussed be expected to stimulate? The second is evaluative or normative: How much of this use represents the proper balancing of costs and benefits?

Some Basic Medicare Data

Top Editor's Notes Some Basic Medicare Data What Do We Want,... The Effect Of Coverage... Judging The Magnitude And... The Elephant In The... Concluding Comments Notes

 
If we compare seniors with and without drug coverage (of any type), there is an extraordinarily wide gap in spending and use. Compared with those lacking drug coverage, those with coverage fill about 28 percent more prescriptions per year (32 versus 25), and in 2003 will spend about twice as much ($2,700 versus $1,400).⁷ However, people without drug insurance differ in other ways (for example, having a lower income, not having other supplemental insurance) from those who are insured, and these other influences could account for some of this difference. What can we estimate with confidence to be the separate effect of drug coverage?

What Do We Want, And Will We All Know It If We See It?

Top Editor's Notes Some Basic Medicare Data What Do We Want,... The Effect Of Coverage... Judging The Magnitude And... The Elephant In The... Concluding Comments Notes

 
There are economic and noneconomic ways of thinking about the "ideal" rate of use of a drug. Both views agree that if the use of the drug lowers total medical cost, such use should occur. While there clearly are specific cases in which drugs produce cost offsets, I assume that in the aggregate, more drug coverage will raise total medical (and Medicare) spending but that there are net health benefits.⁸ Then someone has to say how much benefit is enough to justify the cost.

The economic approach asks us to think of benefits from greater use of some drug in monetary terms and to think of both the value the informed patient places on improved health from use of the drug and the value that the rest of us (altruistic as well as informed) place on that patient’s improved health. The sum of these two values is the total social value from additional use, and use should therefore be extended to the point at which the value of additional use equals its net cost.

The noneconomic "medicalist" approach assumes that health care experts know or can know, based on clinical expertise, how much use a person needs.⁹ Then slogans like "the right drug for the right patient at the right time" sound good. The problem is that the definition of medical need is notoriously difficult, even impossible, at least given the kind of information we have or can have.

There is, however, some support for the view that more insurance reduces barriers for seniors to beneficial drugs. Dana Safran and colleagues found, in a survey of eight states, that more elderly people without drug insurance reported having failed to fill a prescription because of cost, or having taken a smaller dose to reduce the frequency of refills, than was true among elderly people having drug insurance.¹⁰ While most people with or without drug coverage did not report any such problems, one-fourth of those without drug insurance reported not filling a prescription because of cost, compared with 11 percent of those with coverage reporting similar behavior. Of course, some prescriptions are not needed (such as antibiotics for a viral respiratory infection) or might be of small benefit compared with other remedies (such as brand-name prescriptions for headaches or hay fever). But surely some of the forgone use would have been beneficial.

Others have examined the impact of cost sharing on the use of drugs for illnesses or conditions for which medication is thought by clinicians to be more useful ("essential" drugs), relative to drugs for other illnesses.¹¹ Such studies find that cost sharing reduces the use of drugs in both essential and nonessential cases.

These studies are not controlled trials. The Safran study reports simple differences in means, controlling only for income. The "essential drug" studies control for some but not all personal and health characteristics that might affect use. The well-known RAND Health Insurance Experiment, in contrast, used a controlled trial design.¹² Participants were randomly assigned to high and low levels of insurance coverage. However, the experiment only looked at what might be called bundled comprehensive coverage: The level of coverage of all medical goods and services was changed together; in no experiment was drug coverage varied while the level of coverage of other things was held constant. Compared to plans with cost sharing, in the free-care plan there was greater use of and spending for outpatient prescription drugs. The experiment found that all of the identifiable effects of coverage on drug use came though the effect of coverage on episodes of outpatient care (with which prescriptions are frequently associated), not on a higher number of prescriptions per episode. The impact of increased coverage for drugs on health levels was significant for low-income people at high initial risk (especially of high blood pressure) but not for others. It also found that cost sharing reduced inappropriate use as well as use thought to be more clinically appropriate. The clinical differences sometimes go in unexpected directions; the reduced use of antibiotics for viral conditions in the RAND experiment associated with high patient cost sharing was estimated to lead to a 17 percent lower rate of adverse events than under complete insurance coverage.

In general, consumers subject to cost sharing alone seem to be unwilling or unable to distinguish precisely between prescriptions health professionals regard as clinically appropriate (despite the absence of an effect on health) or inappropriate. Based on this confused evidence, it seems impossible to hazard a quantitative guess as to how much of the additional use will represent "needed" versus "unnecessary" additional drug consumption.

So the evidence thus far is only helpful in rejecting extreme hypotheses: Cost sharing does discourage some use that would have done some good, but some of the use it discourages might be worthless or (as in the case of antibiotics) socially harmful. It would be desirable to have better information on the relationship of cost sharing and health outcomes to plan the most effective coverage and to decide whether the best-designed coverage is worth its cost.

The problem is that the information needed to judge clinical appropriateness for drug coverage as a whole is missing. In many cases clinical detail is absent, or professional consensus is either weak or poorly supported by evidence. Such evidence as we have suggests that some of the additional use is appropriate and some is not, with the ratio largely unaffected.

The Effect Of Coverage On Spending

Top Editor's Notes Some Basic Medicare Data What Do We Want,... The Effect Of Coverage... Judging The Magnitude And... The Elephant In The... Concluding Comments Notes

 
What can studies of the effects of drug insurance coverage tell us about the impact of coverage on spending? Getting an idea of the size and configuration of moral hazard–induced increases from a Medicare drug program on spending and use is important. If the increase turns out to be small, that would mean, on the one hand, that the "inflationary" effect of the benefit is small and its financing less of a problem. But, on the other hand, a small increase might also imply that the program’s impact on the problem of "reduced access to needed drugs" for the Medicare population as a whole, which is one of the main rationales for the program, would have been negligible. A large increase in spending, in contrast, could either reflect the widespread provision of much-needed care or increased use of costly drugs with small benefit.

Regardless of the value judgment, are the effects of more insurance large enough to matter? The evidence suggests that the answer is "maybe, but it depends." Lee Lillard and colleagues, for example, found a positive effect on drug spending of supplementary insurance coverage that included drugs compared with no supplementary coverage at all.¹³ Their study estimated that providing drug coverage to Medicare beneficiaries at the average level of private drug coverage (in 1990, the study year) would raise spending by 20 percent. If we assume that the average private plan covered 60 percent of expenses at that time, the "average" elasticity would be 0.33.¹⁴

However, in this and similar research there is no explicit measure of the generosity of drug coverage (but only a binary variable reflecting the presence of coverage), and the obvious adverse selection prevailing in Medigap drug coverage (compared with Medigap without drug coverage) fails even to be detected—much less controlled—by the statistical methods used. The unfortunate conclusion is that although we do not have precise estimates of the impact of drug coverage alone compared with no drug coverage at all, whether combined with partial or generous outpatient physician services coverage, we do know that the effect will be more than trivial if the increase in coverage is large.

Somewhat more reliable are studies (often in the general population) of the impact of varying drug cost sharing, given that some drug coverage was present initially. The message from research on the privately insured population under age sixty-five is consistent: Drug demand responds strongly to user price, and the arc elasticity is 0.3 to 0.4. This was the result found by the RAND experiment; the implied price elasticity was the same for drugs as for outpatient care generally. This is also the figure used by the Congressional Budget Office (CBO) in its estimates of the cost of drug bills. Alan Hillman and colleagues and Theodore Joyce and colleagues find similar results in terms of total spending.¹⁵ Elasticity calculated over the range of user price is 0.28 in Hillman and colleagues’ work and 0.33 in that of Joyce and colleagues. Other studies, designed to look more generally at prescribing behavior, find significantly positive effects of more generous coverage.¹⁶

If supply behavior is limited (for example, by putting prescribing physicians at risk for the cost of drugs), the response to patient cost sharing is greatly reduced.¹⁷ In the literature, proportional copayment does not appear to affect the choice between inexpensive generics and costlier drugs, or to affect the outcomes of drugs used to treat a given illness, but targeted "triple-tier" plans with copayment for nonpreferred brand-name products that are several multiples of those for generics or preferred branded products do have strong "share" effects.¹⁸

Effects of large marketwide changes in coverage on prices of a given quantity or set of drugs are harder to identify. There will be no effect on the price of competitively supplied products as long as supply is elastic (which seems to be the case for generic drugs). For monopoly products, such as patent-protected drugs, whether insurance coverage has an effect depends on whether the patient demand curve with insurance is less elastic than that without insurance.¹⁹

Judging The Magnitude And Appropriateness Of Increases In Use: Another Try

Top Editor's Notes Some Basic Medicare Data What Do We Want,... The Effect Of Coverage... Judging The Magnitude And... The Elephant In The... Concluding Comments Notes

 
How much of the increase from more generous coverage for all beneficiaries would be appropriate? I have already discussed the use of clinical judgment and find the evidence at present to be inconclusive. An alternative approach focuses on the financial barrier imposed by the lack of insurance coverage and tries to identify subpopulations that should be affected differently by that barrier.

Clearly the difference any drug coverage bill would make to total spending and use of drugs, compared with its absence, depends both on the precise provisions of the new program and the situation with regard to spending and insurance coverage that would otherwise have existed. Any estimates of either type will be imprecise at this time, but it still might be possible get some idea of the difference that coverage proposals now before Congress would make to spending on drugs. The two versions of Medicare drug coverage passed by the House and Senate differ in their specific coverage provisions, but the average level of coverage provided in either version is virtually identical, at a surprisingly low 47 percent of spending.²⁰

What effect would these bills have? Although there could be severe adverse selection against a new plan that is not heavily subsidized, let us assume here that everyone who would receive an improvement in coverage will take the newly offered coverage.²¹ Estimates should be scaled down if participation is lower. To be specific: Those currently with either no coverage or Medigap drug coverage are assumed to take up the plan. Those with Medicare+Choice coverage and those with Medicaid or employment-based coverage are assumed to remain with their current coverage, and it is also assumed that there is no change in the proportion with either type of coverage.

Exhibit 1 shows the proportion of Medicare beneficiaries (including institutionalized beneficiaries) at various income levels relative to the federal poverty level who are currently estimated to lack drug coverage and to pay for drugs out of pocket. These three groups, characterized here as "poor and near-poor," "tweeners," and "well-off," approximately divide the beneficiary population into thirds. The proportion without any coverage is fairly uniform across the three groups, as Medicaid for the poor group offsets the negative effect of low income on the choice of private drug coverage. This lowest group consists of the poorer and the sicker, which accounts for their higher proportion of total drug spending. Because it includes more people on Medicaid, the most generous type of coverage, the average percentage paid out of pocket is actually lowest for them.²²

How would this increase in spending be distributed across the three income categories? It would be distributed virtually equally, since the slightly higher percentage uninsured for drugs among the poor is offset by a higher percentage with less generous Medigap drug coverage among the other two groups. How might one evaluate this pattern of spending increase? If one assumes that the well-off would not face financial barriers so serious as to deter use that would matter to others even without coverage, and that the approximately 25 percent of spending below $2,000 per year would be affordable to the "tweeners," one would conclude that about 41 percent (0.33 + [0.33 x 0.25]) of the small increase in spending would have been wasted. The remaining 3.3 percent increase in use and spending would be regarded by this criterion as socially desirable.

Obviously these calculations use (necessarily) imprecise estimates and draw on only a very rough criterion of desirability or efficiency. Very generous catastrophic coverage might add more stimulus than the plans described here—but this use would also be easier for pharmacy benefit managers (PBMs) to target. More careful investigation (probably requiring a plan in place to look at) of the clinical effectiveness of the changes in use might yield a more refined, more defensible number.

But the clear message is one of how small an impact, for good or ill, these programs are likely to have. Their spending increases are swamped by the recent double-digit growth in drug spending. Their modest effect on total spending means that even in the most optimistic circumstances, the Medicare drug coverage proposals now under discussion are not likely to do much good in the aggregate in terms of increased access, greater efficiency in the use of medical care, improved health outcomes, drug company profits, or incentives for research. For the small part of the population whose use is affected, there may be improvements in health (but, one suspects, not as large as if the program were more targeted).

The Elephant In The Living Room: Insurance And The Growth Rate Of Spending

Top Editor's Notes Some Basic Medicare Data What Do We Want,... The Effect Of Coverage... Judging The Magnitude And... The Elephant In The... Concluding Comments Notes

 
Although we have evidence that increases in drug coverage cause more moral hazard and more inefficient spending, the history of the past decade poses a more serious question: dealing with the long-term rate of growth (rather than just the current-period level) of drug spending. Suppose coverage increased for the Medicare population to the same level as that of the population under age sixty-five and then stopped growing. Would the new quantity and type of drugs that have driven drug spending just be "high" (relative to the efficient level) but no longer growing rapidly in real terms? And if the rate of growth of spending was more rapid when the level of coverage was high, would that higher growth be a bad thing?

We know the answers to neither of these crucial questions. There is some evidence for medical spending in general of a connection between the rate of growth of spending and the level of insurance coverage.²³ Total spending on new technology will definitely be higher with more generous coverage than without it, but, paradoxically, because the previous-period spending will also be higher, the percentage rate of increase in spending might not be very high.²⁴ So long as insurers offer plans that can refuse to cover new technology, we know that technology will be financed only if, on balance, it makes consumers better off.

This relatively benign attitude toward spending growth is less plausible for Medicare, because higher premiums to pay for it mean higher taxes and greater attendant economic distortion. Medicare might eventually refuse to pay for a beneficial but costly new technology.²⁵ The problem might not be the value of the new product, but the capacity of the Medicare fiscal system that taxes nonbeneficiaries to support it.

Concluding Comments

Top Editor's Notes Some Basic Medicare Data What Do We Want,... The Effect Of Coverage... Judging The Magnitude And... The Elephant In The... Concluding Comments Notes

 
By the narrow spending and efficiency considerations associated with the idea of moral hazard in insurance, current proposals for Medicare drug coverage improvement seem to be much ado about relatively little. Because these plans increase coverage for relatively few, there should not be a large increase in either the volume of drugs consumed or the revenues to drug companies. Relative to total Medicare spending, current and (especially) future, these proposed programs make even less difference.

However, it is important to note that this conclusion might well not apply to the public finance aspects of the program. In current proposals, most of the tax finance and the beneficiary premium goes to pay for drug expenditures that people are already making—but they shift those expenses from out of the beneficiary’s pocket at the point of use to other pockets at other points in time. That is, the main effect of these programs is one of transfers of money, not changes in medical care or health.

Not all of these transfers need be zero-sum; moral hazard might be an inevitable although undesirable side effect of a valuable program. If (as might be the case) the combination of nearly free original Medicare and adverse selection against stand-alone drug benefits deterred drug coverage that we will all value highly in retirement, there might be a gain. The real cost-benefit calculus ought to value any financial risk reduction as a gain and, as I have argued elsewhere, set it against the potentially massive efficiency losses from tax financing of this insurance as the next generation swells the ranks of beneficiaries.²⁶ Since a sizable portion of the new coverage envisioned by the bills being discussed would not provide this valuable catastrophic protection, but rather "coverage" of much lower value for relatively small and certain drug expenses, one might wonder whether the current proposals are the best (or even a good) way to add drug coverage.

Editor's Notes

Top Editor's Notes Some Basic Medicare Data What Do We Want,... The Effect Of Coverage... Judging The Magnitude And... The Elephant In The... Concluding Comments Notes

 
Mark Pauly is the Bendheim Professor and Chair, Department of Health Care Systems, at the University of Pennsylvania’s Wharton School, in Philadelphia.

The author thanks Amol Shah for excellent research assistance.

Notes

Top Editor's Notes Some Basic Medicare Data What Do We Want,... The Effect Of Coverage... Judging The Magnitude And... The Elephant In The... Concluding Comments Notes

 

1. D. Altman, "Some Doubts about the Logic of Senate Plan for Drug Aid," New York Times, 14 June 2003.
2. M.V. Pauly and P.M. Danzon, "Insurance and New Technology: From Hospital to Drugstore," Health Affairs (Sep/Oct 2001): 86–100.
3. For unexplained reasons, the Congressional Budget Office (CBO) calls the effect of insurance "induced demand," the term in health economics usually used to describe suppliers (physicians, hospitals, or drug companies) who talk people into consuming more of their product at a given user price. In the Washington, D.C., area the moral hazard phenomenon is also often termed the "woodwork effect."
4. This is the best economic interpretation of the "waste" that insurance can bring about.
5. Exceptions exist to this simple view. If supply is constrained (as it is in some other countries but probably would not be for Medicare drug coverage), limited increases in spending and therefore limited moral hazard will occur. More relevant to drugs, if price is much above cost, the extra use might not be inefficient from a short-run societal perspective (since the marginal cost of drugs is usually much less than their price), although it will be undesirable for the consumers and taxpayers footing the bill and possibly in the long-run cost perspective that includes research and development (R&D) costs.
6. A secondary argument is that the patient might have undervalued the drug.
7. D. Crippen, "Designing a Twenty-first Century Medicare Prescription Drug Benefit," Statement before the House Energy and Commerce Subcommittee on Health, U.S. House of Representatives, 8 April 2003 (Washington: Federal Document Clearinghouse, 2003); and Henry J. Kaiser Family Foundation, "Medicare and Prescription Drug Spending Chartpack," June 2003, www.kff.org/content/2003/6087/6087v4.pdf (25 June 2003).
8. For a look at aggregated data, see F.R. Lichtenberg, "Are the Benefits of Newer Drugs Worth Their Cost? Evidence from the 1996 MEPS," Health Affairs (Sep/Oct 2001): 241–251. The great bulk of detailed studies of drug coverage explore the effects of introducing cost sharing for the atypical Medicaid population, a step not contemplated by any of the drug coverage proposals. See S.B. Soumerai et al., "Payment Restrictions for Prescription Drugs under Medicaid: Effect on Therapy, Cost, and Equity," New England Journal of Medicine (27 August 1987): 550–556; and S.B. Soumerai et al., "Effects of Limiting Medicare Drug-Reimbursement Benefits on the Use of Psychotropic Agents and Acute Mental Health Services by Patients with Schizophrenia," New England Journal of Medicine (8 September 1994): 650–655.
9. S. Glied, Chronic Condition: Why Health Reform Has Failed So Far (Cambridge, Mass.: Harvard University Press, 1998).
10. D.G. Safran et al., "Prescription Drug Coverage and Seniors: How Well Are States Closing the Gap?" 31 July 2002, www.healthaffairs.org/WebExclusives/Safran_Web_Excl_073102.htm (22 October 2003).
11. R. Tamblyn, "The Impact of Pharmacotherapy Policy: A Case Study," Canadian Journal of Clinical Pharmacology (Fall 2001): 39A–44A; and A.D. Federman et al., "Supplemental Insurance and Use of Effective Cardiovascular Drugs among Elderly Medicare Beneficiaries with Coronary Heart Disease," Journal of the American Medical Association (10 October 2001): 1732–1739.
12. E.B. Keeler et al., The Demand for Episodes of Medical Treatment in the Health Insurance Experiment (Santa Monica, Calif.: RAND, 1988). The arc (average) elasticity was about 0.3 over the range of coinsurance from 25 to 95 percent. The RAND experiment did not include the elderly as participants.
13. L.A. Lillard, J. Rogowski, and R. Kington, "Insurance Coverage for Prescription Drugs: Effects on Use and Expenditures in the Medicare Population," Medical Care (September 1999): 926–936.
14. This means that for every 10 percent drop in the out-of-pocket price, use and spending rise by 3.3 percent.
15. A.L. Hillman et al., "Financial Incentives and Drug Spending in Managed Care," Health Affairs (Mar/Apr 1999): 189–200; and G.F. Joyce et al., "Employer Drug Benefit Plans and Spending on Prescription Drugs," Journal of the American Medical Association (9 October 2002): 1733–1739.
16. J.K. Hellerstein, "The Importance of the Physician in the Generic versus Trade-Name Prescription Decision," RAND Journal of Economics (Spring 1998): 108–136; and N.E. Coulson and B.A. Stuart, "Insurance Choice and the Demand for Prescription Drugs," Southern Economic Journal (April 1995): 1146–1157.
17. Hillman et al., "Financial Incentives."
18. Ibid.; Hellerstein, "The Importance of the Physician"; J.P. Newhouse and the Insurance Experiment Group, Free for All? Lessons from the RAND Health Insurance Experiment (Cambridge, Mass.: Harvard University Press, 1993); T.S. Rector et al., "Effect of Tiered Prescription Copayments on the Use of Preferred Brand Medications," Medical Care (March 2003): 398–406; and Joyce et al., "Employer Drug Benefit Plans."
19. Such a reduction in elasticity need not necessarily occur; the introduction of aggressive private pharmacy benefit managers (PBMs) might well increase elasticity as coverage expands. Estimates therefore compromise by assuming an effect but making it modest. D. Holtz-Eakin, "Cost Estimate for S. 1, Prescription Drug and Medicare Improvement Act of 2003," 17 June 2003, ftp://ftp.cbo.gov/43xx/doc4330/s1.pdf (30 June 2003).
20. CBO, "Cost Estimate for H.R. 1, Medicare Prescription Drug and Modernization Act of 2003, and S. 1, Prescription Drug and Medicare Improvement Act of 2003," 22 July 2003, ftp://ftp.cbo.gov/44xx/doc4438/hr1s1.pdf (1 August 2003), 28. Precise estimation of the percentage covered is difficult because that depends on how spending responds to coverage; even if the net percentage is somewhat higher than this forecast, the conclusions reported below will still follow.
21. M.V. Pauly and Y. Zeng, "Adverse Selection and the Challenges to Stand-Alone Prescription Drug Insurance," NBER Working Paper 9919, August 2003, www.nber.org/papers/w9919 (12 November 2003).
22. Among those in lowest income group not on Medicaid, both the proportion uninsured for drugs and therefore the percentage paid out of pocket tend to be higher than for the other two groups.
23. E.A. Peden and M.S. Freeland, "Insurance Effects on Medical Spending (1960–1993)," Health Economics (December 1998): 671–687.
24. It is possible that even a small increase in net spending from Medicare coverage might be associated with a change in how payment levels for drugs are determined. If prices are reduced below what they would otherwise have been, those reductions will (prospectively) discourage research.
25. M.V. Pauly, "What if Technology Never Stops Improving? Medicare’s Future under Continuous Cost Increases," Washington and Lee Law Review (forthcoming).
26. M.V. Pauly, "Medicare and Prescription Drugs," Statement before the Committee on Ways and Means, U.S. House of Representatives, 9 April 2003 (Washington: Federal Document Clearinghouse, 2003).

                      What's this?

This article has been cited by other articles:

				S. Schneeweiss, A. R. Patrick, A. Pedan, L. Varasteh, R. Levin, N. Liu, and W. H. Shrank The Effect Of Medicare Part D Coverage On Drug Use And Cost Sharing Among Seniors Without Prior Drug Benefits Health Aff., March 1, 2009; 28(2): w305 - w316. [Abstract] [Full Text] [PDF]

				W. Yin, A. Basu, J. X. Zhang, A. Rabbani, D. O. Meltzer, and G. C. Alexander The Effect of the Medicare Part D Prescription Benefit on Drug Utilization and Expenditures Ann Intern Med, February 5, 2008; 148(3): 169 - 177. [Abstract] [Full Text] [PDF]

				F. R. Lichtenberg and S. X. Sun The Impact Of Medicare Part D On Prescription Drug Use By The Elderly Health Aff., November 1, 2007; 26(6): 1735 - 1744. [Abstract] [Full Text] [PDF]

				B. Stuart, B. A. Briesacher, D. G. Shea, B. Cooper, F. S. Baysac, and M. R. Limcangco Riding The Rollercoaster: The Ups And Downs In Out-Of-Pocket Spending Under The Standard Medicare Drug Benefit Health Aff., July 1, 2005; 24(4): 1022 - 1031. [Abstract] [Full Text] [PDF]

				L. Simpson Lost In Translation? Reflections On The Role Of Research In Improving Health Care For Children Health Aff., September 1, 2004; 23(5): 125 - 130. [Abstract] [Full Text] [PDF]