Why Don’t Americans Use Cost-Effectiveness Analysis?

Cost-effectiveness analysis (CEA) offers decision makers a structured,rational approach with which to improve the return onresources expended. But decades after its widespread promotion tothe medical community, policy makers in the United States remainreluctant to use the approach formally. Indeed, the resistance toeconomic evidence in the United States in an era of evidencebasedmedicine in healthcare is perhaps the most notable developmentof all. This paper examines the resistance to CEA in theUnited States and explores ways in which to advance the field.

((Am J Manag Care. 2004;10:308-312)

Cost-effectiveness analysis (CEA) logically shouldoccupy an important place in the health policymakingtoolkit. The technique promises toincrease the return on society's investment in healthcare.It provides a tool to inform decisions in an explicit,quantitative, and systematic manner. Not only domedical journals routinely publish CEAs,^1-3 but manycountries have incorporated CEA into their technologyassessment and reimbursement procedures.⁴

In theory, CEA should be of use to managed careorganizations, insurers, health departments, and stateand federal policy makers.⁵ But despite the widespreadpromotion of CEA to the medical community, US policymakers remain reluctant to use the technique.

After repeated attempts to incorporate cost effectivenessas a criterion for covering new medical technologies,Medicare abandoned the pursuit in the face ofpolitical opposition.⁶

In another example, Oregon Medicaid initially soughtto rank services based on cost effectiveness. But theplan was opposed on ethical, legal, and political groundsand was implemented only after officials removed theoffending cost-effectiveness provisions.⁷ In the yearssince, no other state Medicaid program has attempted toimplement Oregon-style priority setting.

Why have managed-care plans not embraced CEA?Given their turbulent finances and attempts to coordinatehealthcare under fixed budgets, one might expectthem to be eager consumers. Instead, studies suggestotherwise.^8,9 Health plans have employed evermoreaggressive and sophisticated processes for managing carefrom incentive-based formularies to utilization reviewpolicies. But few, if any, providers use cost effectivenessas a formal policy tool.

It is also difficult to find explicit use of CEA in othercorners of US health policy. Clinical guidelines incorporateeconomic analyses infrequently, even when highqualityevaluations are available.¹⁰ Health insurancecontracts use the term "medically necessary," but not"cost effective."¹¹ In general, researchers find little relationshipbetween the cost effectiveness of lifesaving programsand their implementation.¹²

In this paper, I examine resistance to CEA in theUnited States and offer advice to policy makers. Full disclosure:I receive grant funding to review and occasionallyconduct CEAs. However, my intention in thispaper is not to advocate CEA, so much as to understandits place in the health policy field and itsprospects for the future.

UNDERSTANDING RESISTANCE TO CEA

One explanation for resistance to CEA holds thatAmerican physicians fundamentally do not understand–and have not been taught to think deeply–about resource constraints and tradeoffs.^13,14

A second theory points to lack of a trust in the methods,rather than a lack of understanding per se.Researchers have long observed that many publishedCEAs do not adhere to recommended protocols for conductingand reporting,^3,15 which may contribute to asense that the methodology is not ready for primetime.^16,17

A related problem is a perceived lack of relevance.Observers point to the disconnect between the abstract,societal-perspective of recommended CEAs, and theshort-term horizons of actual decision makers.^18,19 Critics argue that CEA studies neglect the budgetimpact of the service in question.²⁰ Promised savingsare perceived as illusory.

increase

A more cynical explanation questions not the methodsof CEA but the motives of investigators and/or sponsors.On the one hand, CEA is seen as a smokescreenfor cost-cutting efforts. Physicians view CEA as anaccounting or cost-management tool used by managersto improve the bottom line.^14,18,21 On the other hand,CEA is perceived as a tool with which those with afinancial gain at stake, particularly the drug companieswho increasingly sponsor studies, can advocate to health expenditures.¹⁶ Yet another explanationpoints to legal and regulatory barriers as key factors.¹¹

But each of these explanations falls short. Lack ofunderstanding or mistrust of methods and motivesundoubtedly plays a role. Still, why have policy makersin the United States not funded or conducted their ownCEAs, or tailored them to their own needs as they havein other countries?

Legal and regulatory factors also fail as a full explanation.Nothing in federal statute explicitly barsMedicare from using CEA. Similarly, nothing preventsprivate health insurers from writing contracts that specifycovered services as those that are deemed "medicallynecessary and cost effective." Plans may fear lawsuitsif they use CEA openly. But there are also plausible reasonsto believe that health plans could withstand thesechallenges, as they have withstood challenges to othercost-containment initiatives.¹¹

The best explanation is that, at its roots, resistanceto CEA in the United States is grounded, not in methodologicalor legal barriers, but in Americans' deep-seateddistaste of limits and of the corporate or governmentofficials who impose them.

At some level, people do not believe that resourcesreally are limited, or they recoil from the explicit natureof the cost-effectiveness exercise itself–that it forcesthem to think consciously about stark tradeoffs betweenmoney and health that they would rather leave at a subconsciousprivate level.¹³ Physicians believe using CEAviolates clinicians' advocacy duties, destroying the trustnecessary for good doctor—patient relationships.¹⁴

Moreover, CEAs may not capture public preferencesfor allocation of limited resources. As Daniels and Sabinpoint out, while intended as an empirically based, transparentmethodology relying on evidence, CEA is notvalue neutral.²² It carries with it morally controversialand, to some, unacceptable assumptions. Empirical evidencesuggests that patients and their physicians desireequity and fairness in a general sense, rather than efficiencyper se.^23-25 They want to give priority to patientswith life-threatening treatments or to children or thedisabled, even when treatments are not cost effective.^14,26,27 Application of CEA is seen as violating thespecial moral importance of health or as breaching citizens'inherent rights to healthcare.^21,28

Objections to CEA may be particularly acute in theUnited States for several reasons. For one, survey datasuggest that compared to their counterparts in Europeor Canada, Americans are more concerned about accessto the most advanced medical technologies.29 Foranother, CEA may be seen as especially objectionable inthe United States' private and often for-profit healthcaresystem. Consumers and providers may believe thatthese economically driven, sometimes publicly tradedorganizations are not motivated or inspired by a desireto maximize the value of healthcare services to membersof their plans. They may believe that the overallhealth system reflects unwise and inefficient use ofresources, and they may therefore be unwilling toforego potentially beneficial care simply because of cost.People have demonized managed-care plans for tryingto limit patients' choice of provider and treatment basedon economic considerations; they may view CEA as aformal, explicit methodology to achieve the same end.

Other countries' acceptance of CEA confirms thatthe United States' failure to use CEA is driven more bythe country's own cultural, political, and institutionalconditions than by the technique's inherent methodologicalshortcomings.

DOES ANYONE USE CEA? EXAMPLESFROM THE FIELD

But how to reconcile the resistance toward CEA withthe rapid growth of analyses in mainstream Americanmedical journals? Why would analysts continue publishingstudies if the information was assiduouslyignored?

CEA may actually enjoy considerable influence inthe United States, not as an explicit instrument for prioritizinghealth services, but as a subtle influence inpolicy discourse. Over the years CEA has challengedprevailing wisdom and brought clarity to healthcaredebates–underscoring, for example, that preventionprograms usually do not produce cost savings ³⁰ andthat sometimes high-tech, cost-increasing interventionscan ultimately provide very good value for themoney.³¹

Most health plans and formulary managers have aninterest in CEAs; and almost all large medical-productscompanies now have sophisticated health economicunits, reflecting a widely held view that customers nowconsider this information when making purchasing decisions.Health plans have begun adopting the Academy ofManaged Care Pharmacy's (AMCP's) new evidence-basedformulary guidelines that call for drug manufacturersto submit dossiers of clinical and economicevidence about their products to support the listing ofnew pharmaceuticals.³²

CEA may also be influencing selected clinical guidelinesincreasingly. For example, CEAs have shownstatin drugs to be relatively cost effective as secondaryprevention in persons with existing heart disease, butconsiderably less cost effective as primary prevention.³³ While formal recommendations did not follow theseCEAs strictly, the economic analyses were cited in theguidelines and likely play a role in targeting therapy.^5,34

More importantly, several national initiatives nowconsider CEA formally, including the recent recommendationsof the third US Preventive Services TaskForce (USPSTF) and the Guide to Community PreventiveServices.^35,36

IMAGINING A FUTURE FOR CEA

Has CEA had any influence in the 25 years since itsintroduction to the American medical community?³⁷ Ofcourse. By its sheer existence, CEA changes the natureof conversations about the impact of investments inhealthcare. It forces and focuses discussions about thevalue of health and medical services within a clear theoreticalframework. It generates a more careful considerationof available evidence and sheds light on how totarget resources to particular clinical practices or subgroupsof patients.^38,39 The thousands of articles in peer-reviewedliterature over the years attest to its currencyamong researchers and thought leaders.

At the same time it has fallen short of expectations.Visions of using CEA as a formal tool for prioritizinghealth services never materialized. But excessive pessimismabout the prospects of CEA would be misplacedtoo.

US policy makers have gravitated toward a policy of"cost effectiveness once removed." The rules of engagementallow use of CEA but only at a safe distance.Rationing is permitted under the radar. Physicians preservethe fiction that everything is being done for theirpatients, even as they play a supporting role in apportioningscarce societal resources. At the same time,patients safeguard their trust in their doctors. Payersmaintain legal protection.

Ubel observed recently that "cost-effectivenessanalysis has had, at best, a troubled youth… but it willgive way to a successful adulthood."¹⁴ He went on topredict that administrators, policy makers, and otherswill become more adept at using CEA.

This may yet happen. The United States may witnessa broad cultural shift in the way policy makers thinkabout priority setting.⁴⁰ Practicing physicians may beginto embrace CEA openly as a tool for bedside rationing.Payers may begin appealing more directly to evidencefrom CEAs when writing contracts or targeting servicesfor coverage. The public may become more acceptant ofthe technique, especially if alternative approaches torationing, such as bureaucratic obstacles and longqueues, are seen as even more objectionable.

But these scenarios seem unlikely. If anything, recenthistory points to a backlash against the imposition oflimits. The public's appetite for managed care has longbeen eroding.⁴¹ Unlike our neighbors to the north oracross the ocean, the United States seems destined tokeep CEA at bay, defying the trend as we have resistedthe metric system.

The more probable future is one in which cost-effectivenessinformation is produced in abundance butrarely used in explicit fashion to cover or deny services.Instead, employers and health insurers will likely continueto impose limits in other ways.

ADVICE FOR THE US POLICY MAKERAND POLITICIAN

More and better training about CEA in schools ofmedicine, public health, pharmacy, and nursing wouldhelp the field and improve the quality of CEAs. So toowould renewed attempts to improve the peer-reviewprocess.^17,42

A strong, independent organization to conduct orevaluate CEAs would also be beneficial. Public organizations,such as the USPSTF, have made some progress.Government agencies, such as the Agency for HealthCare Research and Quality and the National Institutesof Health sponsor selected CEAs.⁴³ However, a clearlyarticulated policy for coordinating CEA research acrossagencies has never existed.

A stronger role for the FDA is possible, with theagency given authority to examine the cost effectivenessof prescription drugs before approval. But changing theFDA's approval authority would be a mistake and wouldembroil the agency in debates about value that are betterleft to the marketplace. Greater responsibility for theCenters for Medicare and Medicaid Services (CMS)seems natural, but experience has shown the enormousdifficulties the agency would face.⁴⁴

A new agency within the US Department of Health andHuman Services could be established, but political forceswould conspire against it. A better idea is to create aquasi-public entity like the Institute of Medicine to judgethe cost effectiveness of new therapies, though it toowould be hard-pressed to weather the political storms.

In the end, the best hope may be a decentralizedreform like the emerging Academy of Managed CarePharmacy formulary guidelines. Such reform could permitexplicit considerations of value without the falloutof centralized government assessments.

US policy makers and politicians can also learn severallessons from the past.

CEA Should Not Be Used Rigidly

Leaders in the field have always warned against usingCEA mechanically, but it took Oregon's misadventuresto drive home the lesson. Oregon was viewed as being inthe vanguard of health policy making, but the rest of thecountry never followed.⁴⁵ Expectations for CEA shouldbe modest. In countries that have implemented CEA,subsequent decisions have been consistent with economicefficiency. The Australian reimbursement authoritieshave been unlikely to recommend a drug if thecost-effectiveness ratio exceeded AU $76 000 per lifeyearsaved and unlikely to reject it if less than $42 000per life-year saved.⁴⁶ An upper threshold of about£30 000 per quality-adjusted life-year seems to haveemerged at NICE.⁴⁷ But exceptions have also beenmade, as in the case of riluzole for amyotrophic lateralsclerosis and beta interferon for multiple sclerosis.⁴⁸

CEA Will Not Save Money

CEA should never be conceptualized or promoted asa cost-containment tool but rather as a technique toobtain better value.⁵ Paradoxically, using CEA tends toincrease health spending because it reveals more underthanovertreatment.³⁹ This truth will likely present anadditional challenge for those advocating the methodology,but transparency on the matter will help manageexpectations.

It Matters How You Say It

cost effectiveness

Research shows that physicians understand thatresources are limited, but they are not willing to admitto rationing.¹⁴ Similarly, health plan managers denythat they ration care but admit that their budgets areconstrained.¹⁸ These responses are instructive. It suggeststhat the term is part of the problem.Insisting on an "eat your broccoli" approach–thatAmericans must simply come to grips with the need toration–is unlikely to work.

Incentives First

Debates about the use of CEA cannot be divorcedfrom debates about the underlying health system andthe incentives the systems embody (eg, formularieswith tiered copayments that create incentives forpatients to seek lower cost drugs). Value-for-moneyarguments may be more acceptable if used in a decentralizedfashion by competing plans rather than centrallyby CMS. But reconfiguring the incentives that faceproviders and patients is critical.

Think Broadly Across Sectors

A final message involves the importance of thinkingexpansively about applications of cost-effectivenessanalysis. Cost effectiveness analysis should not focussimply on health and medical interventions but morebroadly on interventions to reduce environmental,motor vehicle, and occupational risks as they too affecthealth outcomes.

Anticipating the Medicare Drug Benefit

To Medicare, CEA has been an elephant in the livingroom, officially ignored despite its obvious importance.A decade of failed attempts to integrate CEA hasrevealed the strength of reluctance in the United Statesto openly confront resource constraints. If Medicareofficials–and politicians–learned anything from theordeal, it is the political folly of trying to ration honestly.Despite rising Medicare expenditures and theimpending retirement of the baby boomers, prospectsfor the adoption of CEA remain dim. Physicians arealready angered and frustrated by CMS. Politicians willnot risk antagonizing elderly voters.

At the same time, Medicare cannot avoid difficultdecisions about costly new technology. What's a $400billion-agency to do? Most likely, it will stumble alongwith one hand tied behind its back, evading the toughchoices by delegating authority, cutting payment rates,and handling expensive new technology through codingand payment policy, all the while reassuring beneficiariesthat they will maintain access to important advances.

Recent high-profile national Medicare coverage decisionson implantable cardioverter defibrillators and lungvolume reduction surgery are instructive. Despite vowsthat it does not use CEA, Medicare in both instancessought to limit its budget exposure by selecting populationsfor coverage, consistent with the notion thatMedicare appeals to economic factors, if not explicitCEA, in making decisions.

All of this anticipates the newly enacted Medicaredrug benefit that will take effect in 2006. Many questionsremain about how private plans will make formularydecisions about new drugs. A key question iswhether they will use CEA explicitly. History suggeststhat it will be difficult for them to do so. Tellingly, newlegislation gives the Agency for Health Care Researchand Quality new funds for cost-effectiveness researchbut explicitly forbids CMS from using the information towithhold coverage of new drugs.

CONCLUSIONS

Various factors explain the resistance to CEA in theUnited States, including a lack of understanding aboutthe conceptual approach, a mistrust of methods andmotives, and regulatory and legal barriers. But theloudest message of all pertains to Americans' distaste oflimits.

CEA's proponents have always assumed thatincreased spending and the public's appetite for medicaltechnology will eventually force us to recognize limitsmore directly and that inevitably we will embrace CEAas the best solution to our dilemma. But Americans'reluctance to acknowledge limits–and their resistanceto CEA–has proven remarkably durable. In manyways, CEA has emerged as a technical success but apolitical failure.

There are a few encouraging signs. Where use of thetechnique has emerged, however, it has done so not asthe Oregon plan's architects envisioned but in morenuanced fashion. The most plausible scenario for thefuture is one in which coverage decisions reflect anuntidy combination of evidence- and politics-basedmedicine and in which CEA plays a behind the scenesbut not explicit role.

Acknowledgment

I am grateful to Milton Weinstein, PhD, and an anonymousreviewer for comments on earlier drafts of this manuscript.

From the Program on the Economic Evaluation of Medical Technology, Center for Risk Analysis, Harvard School of Public Health, Boston, MA.

The author is suported by grants from the Agency for Health Care Research and Quality and The Robert Wood Johnson Foundation’s Changes in Health Care Financing and Organization (HCFO) Initiative. The views expressed are those of the author.

Address correspondence to: Peter J. Neumann, ScD, Harvard School of Public Health, 718 Huntington Ave, 2nd Floor, Boston, MA 02115. E-mail: pneumann@hsph.harvard.edu.

Med Care

Med Care.

Ann Intern Med.

Value Health.

Cost-effectiveness in Health andMedicine.

64Federal Register

Med Decis Making.

HealthAffairs.

Health Affairs.

J Gen Intern Med.

J of Health Polit Policy Law.

Risks, Costs, and Lives Saved: Getting BetterResults from Regulaton.

JAMA.

Pricing Life: Why it's Time for Health Care Rationing.

Ann InternMed.

Ann Intern Med.

JAMA

Am J Manag Care.

JQual Improve.

Am J Manag Care.

Am JManag Care

Setting Limits Fairly: Can We Learn to Share MedicalResources?

Health Policy

Med DecisMaking.

Med Decis Making

JAMA.

JAMA.

Pricing the Priceless: Cost-benefit Analysis ofEnvironmental Protection.

Health Affairs.

Is Prevention Better Than Cure?

AnnIntern Med.

HealthAffairs.

JAMA.

Circulation.

Am J Prev Med.

Am J Prev Med.

N Engl J Med.

BMJ.

Int JTechnol Assess in Health Care.

JournalMed Ethics

JAMA.

BMJ.

Med Care.

Harvard Health Pol. Rev.

Med Care Res Rev.

PharmacoEconomics.

BMJ.

Cost-Effectiveness Thresholds: Economic andEthical Issues.

1. Elixhauser A, Luce BR, Taylor WR, Reblando J. Health care CBA/CEA: an updateon the growth and composition of the literature. . 1993;31:JS1-JS138.2. Elixhauser A, Halpern M, Schmier J, Luce BR. Health care CBA and CEA from1991 to 1996: an updated bibliography. 1998;36:MS1-MS9.3. Neumann PJ, Stone PW, Chapman RH, Sandberg EA, Bell CM. The quality ofreporting in published cost-utility analyses, 1976-1997. 2000;132:964-972.4. Hjelmgren J, Berggren F, Andersson F. Health economic guidelines–similarities,differences, and some implications. 2001;4(3):225-250.5. Gold MR, Siegel JE, Russell LB, Weinstein MC. New York, NY: Oxford University Press; 1996.6. Medicare program: procedures for making national coverage decisions, 22619-22625 (1999).7. Tengs TO, Meyer G, Siegel JE, et al. Oregon's Medicaid rankings and cost-effectiveness:is there any relationship? 1996;16:99-107.8. Luce BR, Lyles AC, Rentz AM. The view from managed care pharmacy. 1996;4:168-176.9. Titlow K, Randel L, Clancy CM, Emanuel EJ. Drug coverage decisions: the roleof dollars and values. 2000;19(2):240-247.10. Wallace JF, Weingarten SR, Chiou CF, et al. The limited incorporation of economicanalyses in clinical practice guidelines. 2002;17:210-220.11. Jacobson PD, Kanna ML. Cost-effectiveness analysis in the courts: recent trendsand future prospects. 2001;26(2):291-326.12. Tengs TO, Graham JD. The opportunity costs of haphazard social investmentsin life-saving. In: Hahn RW, ed. New York, NY: Oxford University Press; 1996:167-182.13. Eddy DM. Clinical decision making: from theory to practice. Cost-effectivenessanalysis. A conversation with my father. 1992;267:1669-1675.14. Ubel PA. Cambridge,Mass: MIT Press; 2000.15. Udvarhelyi IS, Colditz GA, Rai A, Epstein AM. Cost-effectiveness and costbenefitanalyses in the medical literature: are they being used correctly? 1992;116:238-244.16. Evans R. Manufacturing consensus, marketing truth: guidelines for economicevaluation. 1995;123:59-60.17. Rennie D, Luft HS. Pharmacoeconomic analyses: making them transparent,making them credible. . 2000;283:2158-2160.18. Prosser LA, Koplan JP, Neumann PJ, Weinstein MC. Barriers to using costeffectivenessanalysis in managed care decision making. 2000;6:173-179.19. Berger ML. The once and future application of cost-effectiveness analysis. 1999;25(9):455-461.20. Langley PC. Is cost-effectiveness modeling useful? 2000;6(2):250-251.21. Asch DA, Jepson C, Hershey JC, Baron J, Ubel PA. When money is saved byreducing health care costs, where do US physicians think the money goes? . 2003;9(6):438-442.22. Daniels N, Sabin JE. New York, NY: Oxford University Press; 2002.23. Ubel PA, Loewenstein G, Scanlon D, Kamlet M. Value measurement in costutilityanalysis: explaining the discrepancy between analog scale and person tradeoffelicitations. . 1998a:43:33-44.24. Ubel PA, Spranca M, DeKay M, et al. Public preferences for prevention versuscure: what if an ounce of prevention is worth only an ounce of cure? 1998;18:141-148.25. Perneger TV, Martin DP, Bovier PA. Physicians' attitudes towards health carerationing. . 2002;22:65-70.26. Hadorn DC. Setting health care priorities in Oregon: cost-effectiveness meetsthe rule of rescue. 1991;265:2218-2225.27. Hadorn DC. The problem of discrimination in health care priority setting. 1992;268:1454-1459.28. Heinzerling L, Ackerman F. Washington, DC: Georgetown Environmental Law andPolicy Institute; 2002.29. Kim M, Blendon RJ, Benson JM. How interested are Americans in new medicaltechnologies? a multicountry comparison. 2001;5:194-201.30. Russell LB. Washington, DC: BrookingsInstitute; 1986.31. Weinstein MC . High-priced technology can be good value for money. 1999;130(10):857-858.32. Neumann PJ. Evidence-based and value-based formulary guidelines. 2004;23(1):124-134.33. Goldman L, Weinstein MC, Goldman PA, et al. Cost-effectiveness of HMGCoAreductase inhibition for primary and secondary prevention of coronary heartdisease. 1991;265:1145-1151.34. National Cholesterol Education Program (NCEP). The second report of theExpert Panel on Dectection, Evaluation, and Treatment. 1994;89:1329-1495.35. Saha S, Hoerger TJ, Pignone MP, et al. The art and science of incorporatingcost effectiveness into evidence-based recommendations for clinical preventiveservices. 2001;20(3S):36-43.36. Carande-Kules VG, Maciosek MV, Briss PA, et al. Methods for systematicreviews of economic evaluations for the Guide to Community Preventive Services. 2000;18(15):75-91.37. Weinstein MC, Stason WB. Foundations of cost-effectiveness analysis forhealth and medical practice. 1977;296:716-721.38. Sculpher M, Drummond M, O'Brien B. Effectiveness, efficiency, and NICE[editorial]. 2001;322:943-944.39. Mitchell AS. Antipodean assessment: activities, actions, and achievements. 2002;18(2):203-212.40. Weinstein MC. Should physicians be gatekeepers of medical resources? . 2001;27:268-274.41. Robinson JC. The end of managed care. 2001;285:2622-2628.42. Jefferson T, Demicheli V. Quality of economic evaluations in health care. 2002;324:313-314.43. Power EJ, Eisenberg JM. Are we ready to use cost-effectiveness analysis inhealth decision making? 1998;36:MS10-MS17.44. Neumann PJ. Medicare, cost-effectiveness, and new medical technology. In press.45. Vladeck BC. Commentary on the Oregon Health Plan. 2000;57:235-242.46. George B, Harris A, Mitchell A. Cost-effectiveness analysis and the consistencyof decision making: evidence from pharmaceutical reimbursement in Australia(1991 to 1996). 2001;19(11):1103-1109.47. Taylor D. Funding medicines for people with multiple sclerosis. 2001;323:1379-1380.48. Towse A, Pritchard C. Does NICE have a threshold? An external view. In:Towse A, Pritchard C, Devlin N, eds. London: Office of Health Economics;2002:25-30.

Download PDF: Why Don"t Americans Use Cost-Effectiveness Analysis?