Ethical Considerations in CMS’sCoverage With Evidence Development

Annemarie Relyea-Chew, JD, MS

This paper reviews the CMS coverage with evidence development policy, its manner of implementation, andkey ethical issues raised by the policy. The author describes ethical considerations and issues associated with theprocess of coverage with evidence development for generating evidence for novel or emerging technologies.

Key Words: Coverage with evidence development, CMS, Medicare coverage, ethics

J Am Coll Radiol 2011;8:838-841. Copyright © 2011 American College of Radiology

New biomedical technologies—diagnostics, drugs, anddevices—require scientific and regulatory assessment ofbenefits, risks, efficacy, and costs before being approvedfor patient care. Even after successfully navigating theextensive FDA approval process [1] and receiving cover-age approval by local Medicare contractors, novel tech-nologies may lack sufficient evidentiary support to justifyissuance of a national coverage determination (NCD) byCMS. During the NCD process, scientific and medicalevidence is gathered and weighed before a coverage deci-sion (ie, payment or nonpayment) is made for the tech-nology at issue. CMS will not authorize payment forservice or devices “not reasonable and necessary for thediagnosis or treatment of illness or injury or to improvethe functioning of a malformed body member” [2].There is no regulatory definition of “reasonable and nec-essary,” but in 2000, CMS clarified that the primaryfactors in making an NCD include whether a technologyis safe, effective, and appropriate and whether its useleads to improved meaningful health outcomes in theMedicare population [3]. Currently, cost-effectiveness isnot a consideration in the coverage decisions.

Medicare has struggled with how best to providetimely and fair coverage for new but unproven medicaltechnologies. On an ad hoc basis, some promising tech-nologies that lacked sufficient outcomes data have re-ceived CMS coverage predicated on patient participationin prospective clinical trials or registries (Table 1 [4-9]).For example, the National Emphysema Treatment Trial,a multicenter, randomized trial begun in 1996, evaluatedthe relative value of lung-volume reduction surgery as anadjunct to medical therapy in patients with advancedemphysema [10]. The study is considered a successfulcollaborative model among federal agencies: the Na-

Department of Radiology, University of Washington, Seattle, Washington.

Corresponding author and reprints: Annemarie Relyea-Chew, JD, MS,University of Washington, Department of Radiology, Seattle, WA 98104;

e-mail: archew@u.washington.edu.

838

tional Institutes of Health and the National Heart, Lung,and Blood Institute conducted the trial and CMS condi-tioned coverage on participation in their randomizedcontrolled trial (RCT). Most important, the evidencefrom the National Emphysema Treatment Trial demon-strated that lung-volume reduction surgery should not beused for high-risk patients in inappropriate settings [11].Since publication of the RCT’s results in 2003 and theissuance of an NCD limiting payment to patients meet-ing criteria derived from the RCT, the number of lung-volume reduction surgery procedures being performedhas dropped (Table 1).

IMPLEMENTING A POLICY OF COVERAGEWITH EVIDENCE DEVELOPMENTIn 2005, CMS issued draft guidance describing the pol-icy of coverage with evidence development (CED) as anoption whereby Medicare would cover new technologiesand interventions even when there is insufficient evi-dence for CMS to make an NCD. Coverage with evi-dence development is specifically conditioned on eligibleindividuals’ participation in relevant research or regis-tries; coverage, or payment, is available only to thoseeligible individuals who agree to participate in the re-search [12]. The same year, CMS conditioned coveragefor several medical technologies and interventions: off-label use of costly drugs for colorectal cancer in RCTssponsored by the National Cancer Institute, enrollmentin a registry for prospective data collection for prophy-lactic use of implantable cardioverter-defibrillators(ICDs), and enrollment in a prospective data collectionregistry for PET for cancers [13] (Table 1).

Responding to criticism and debate of the 2005 draftguidance, CMS issued a revised document, “NationalCoverage Determinations With Data Collection as aCondition of Coverage: Coverage With Evidence Devel-opment” [14]. The 2006 guidance document formalizedand described two forms of CED, deriving statutory au-

thority from the Social Security Act [2]: (1) coverage with

© 2011 American College of Radiology0091-2182/11/$36.00 ● DOI 10.1016/j.jacr.2011.08.011

a

ot4ecabawptc

Relyea-Chew/Ethical Considerations in CED 839

appropriateness determination and (2) coverage withstudy participation (CSP). The intent of coverage withappropriateness determination is to provide coverage of atechnology or service that meets the criteria of “reason-able and necessary” yet requires additional data to deter-mine appropriateness criteria for the specific intervention[14]. Providers submit data to registries or databases spe-cifically designed to aggregate the information needed tomake an NCD, but the registries or databases are notnecessarily monitored or maintained by CMS. The guid-ance document states that the data collection will bescientifically valid and that there will be patient safetymonitoring, quality assurance, and data protection. Spe-cifically addressing patient privacy concerns, CED data-bases, registries, and research protocols must be in com-pliance with provisions of 5 USC § 552(a) pertaining torecords maintained on individuals [15], HIPAA [16,17],nd protections of human research subjects [18].

Coverage with subject participation is analogous to theriginal CED concept. Where insufficient evidence existso demonstrate that an item or service is adequate to meet2 USC § 1862(a)(1)(A) standards, it could be consid-red “reasonable and necessary” if the patient enrolls in alinical study designed to provide evidence of the risksnd benefits of that of the item or service. If new evidenceenefits diagnostic and therapeutic practice, and resultsre published in a peer-reviewed journal, the NCDould be reconsidered. Only patients who agree to beart of the study will be covered by CMS; nonparticipa-ion will result in noncoverage. Statutory authority is

Table 1. Coverage with evidence developmentTechnology Under Review�

LVRS (NCD 2003) 1996-200(NETT)

CREST (NCD 2001) CategorycoveraRCT

PET for dementia and neurodegenerative diseases(NCD 2004)

FDG-PETsuspecor regi

Chemotherapy for colorectal cancer (NCD 2005) Off-labelenrolle

Cochlear implantation (NCD 2005) CoveragICDs (NCD 2005) Coverag

ACC NRegist

PET for cancer (NCD 2005) CoveragNation

Home use of oxygen (LOTT) (NCD 2006) RCTs; LOArtificial heart (NCD 2008) RCT; cat

exemp

Sources: ClinicalTrials.gov [4], Brott et al [5], CMS [6,8,9,13], National OnNote: ACC � American College of Cardiology; CREST � Carotid Revadeoxyglucose; ICD � implantable cardioverter-defibrillator; LOTT � LoNCD � national coverage determination; NETT � National Emphysema Trof Neurological Disorders and Stroke; RCT � randomized controlled trial.�Current as of March 2011.

ited at 42 USC § 1862(a)(1)(E), which permits Medi-

care coverage for items or services used in research deter-mined to meet the priorities of Medicare by the Agencyfor Healthcare Research and Quality [19]. This secondarm of CED has met with criticism as raising significantethical questions.

ETHICAL ISSUES IN COVERAGE WITHEVIDENCE DEVELOPMENTThe federal policy for the protection of human subjects(the “common rule”) incorporates principles of the Bel-mont report, and federally funded research is governedby those principles. The Belmont report, drafted in1979, incorporates basic principles to guide the ethicalconduct of research using human subjects [20]. The 3prescriptive tenets of the Belmont report are justice orfairness, beneficence, and respect for persons or auton-omy. The primary ethical questions raised by CED thattouch on these tenets are (1) whether Medicare benefi-ciaries who either cannot, or will not, participate in aclinical trial are deprived of a right to coverage; (2)whether Medicare beneficiaries are an appropriate popu-lation to study and upon which to base coverage deci-sions; and (3) whether conditioned participation in aCED study is coercive.

Is CED equitable? Medicare beneficiaries are entitledto items or services that are “reasonable and necessary”and for which NCDs have been issued. Not all eligibleMedicare beneficiaries have access to or consent to par-ticipation in CED research as a condition of coverage.

Description OutcomeVRS covered during NIH RCT 2003 publication of RCT;

LVRS completedinvestigational device exemption;for enrollees in NINDS and NIH

RCT ongoing

vered for enrollees withdementia and enrolled in RCT

PET registry trialsongoing

of chemotherapy drugs forin NCI-sponsored RCTs

RCTs ongoing

r enrollees in RCTs No RCTs proposedr subgroup; prospective data innal Cardiovascular Data

Registry ongoing

r providers, patients in registry;ncologic PET Registry

Registry ongoing

RCTs ongoingry B investigational device RCT ongoing

gic PET Registry [7].larization Endarterectomy Versus Stenting Trial; FDG � 2-[18F]fluoro-2-erm Oxygen Treatment Trial; LVRS � lung volume reduction surgery;ent Trial; NIH � National Institutes of Health; NINDS � National Institute

3 L

Bge

coted

stryuseese foe foatio

rye foal OTT

egotion

coloscung-Teatm

Those beneficiaries lacking access to conditioned cover-

rchmCvrsaMiptenccraearn

ipIMopTdFcw

840 Journal of the American College of Radiology/Vol. 8 No. 12 December 2011

age through registries or CSP do not receive a benefitother members of their Medicare cohort receive. Fairdistribution and access to benefits is an issue, especially asall potential CED research participants do not reside ingeographic regions where new medical technology isstudied. The disparity is exacerbated by the vagaries ofUS health insurance coverage for the same interventions(ie, covered, not covered, or covered under CED).

Proponents have rebuffed the concern that CED isunfair, arguing that when patients are not otherwise en-titled to a particular treatment, there can be no harm tothose who are unable to participate in CED research[21]. Participation in a CED study and receipt of cover-age is not an entitlement, it is claimed, but a policydeveloped to support early access to innovative medicaltechnologies. For those Medicare beneficiaries who doparticipate in a CSP clinical trial, randomization andblinding pose the risk of not receiving the purportedbenefit but having to undergo the burden of the process,follow-up, and monitoring [22].

Are Medicare patients appropriate subjects for CEDesearch? The principle of beneficence incorporates twoomplementary obligations: (1) refrain from or do noarm (nonmaleficence), and (2) maximize benefits andinimize risks. When reviewing requests for an NCD,MS examines evidence indicating the product or ser-

ice will improve health outcomes and the magnitude ofisks and benefits. Only individuals who qualify as re-earch subjects may participate in a CED clinical trial,nd the sample should include representatives of the

edicare population with the health condition describedn the NCD [12]. Without an appropriate sample size ofatients who meet the inclusion criteria, generalizabilityo the population in question is unlikely. Medicare ben-ficiaries are primarily adults aged � 65 years, predomi-antly women, and many have comorbidities that pre-lude selection for CED studies [23]. Enrollment inlinical trials for this population is low historically. Theecent CMS decision, for example, which denied cover-ge for CT colonography screening on the basis of cost-ffectiveness analyses, has this limitation [24]. The meange of subjects enrolled in studies of CT colonographyeviewed by CMS in making its determination was sig-ificantly younger than that of the Medicare population.A recent NCD will test the ethical principle of benef-

cence. Known contraindications to the use of MRI inatients with implanted permanent pacemakers andCDs have precluded the issuance of an NCD for clinical

RI in this population [25]. Injuries during MRI doccur in patients with ICDs or implanted permanentacemakers and also trigger sentinel event alerts [26].he topic has been repeatedly reviewed and coverageenied by CMS because of patient safety concerns. Onebruary 24, 2011, CMS released a revised NCD foronditioned coverage of MRI for Medicare beneficiaries

ith implanted permanent pacemakers and ICDs [27].

Evidence offered to support the revision was not highquality. Medtronic, the manufacturer of a newly ap-proved ICD marketed as safe for the MRI environment(with some conditions), opposed the CED and obtainedexpedited reconsideration by CMS for unrestricted cov-erage of MRI, proposed that coverage be extended whenthe Medtronic or analogous produce is implanted inMedicare patients. On July 7, 2011, CMS issued its finaldetermination and will cover MRI when a product ap-proved by the FDA as safe for the MRI environment–such as Medtronic’s–is implanted in Medicare patients[28]. The new NCD creates incentive to implant orreplace ICDs with a new product that lacks a safetyrecord save one study submitted to the FDA for expe-dited review for product approval [29].

Is CED coercive? The main concern expressed by crit-ics of CED is that the ethical principle of respect forpersons is compromised by offering Medicare patientsmedical coverage in exchange for participating in re-search [22,30]. The transaction, it can be argued, is co-ercive. A research subject’s informed consent should bevoluntary and knowing (competent); if given under co-ercion, or duress, it is invalid [31]. When an elderly, illpatient lacks alternatives to treatment save enrollment ina coverage with appropriateness determination registryor CSP, arguably consent is not volitional. “Physicians,”states an authority, “should be aware of how vulnerablepatients may be to the coercive influence of unrealistichope, especially those suffering from chronic, life-threat-ening disorders” [32]. Vulnerability is context specific. Inthe context of research, vulnerable populations specifi-cally entitled to protection include elderly individualswho are economically or educationally disadvantaged[33]. Nevertheless, proponents of CED argue that Medi-care beneficiaries could pay privately for access to inno-vative technology as an alternative [34]. Realistically, in-terventions using new health care technology are likelyexpensive, and the cost would be prohibitive for virtuallyall Medicare recipients.

When elderly patients are recruited and may qualify asstudy subjects, an ethical issue raised is whether thesepatients comprehend the difference between receivingclinical medical care and being a study subject. Thera-peutic misconception occurs when patients conflate thegoals of medical treatment with those of research, trust-ing that they will receive individual medical benefitrather than serve as mere data points for new knowledge[35]. In the quest for sufficient Medicare subjects, phy-sician-researchers should be cognizant of basic ethicalprinciples and not exploit the trust of vulnerable, elderlypatients who cannot distinguish differences among theroles of physician, researcher, patient, and subject.

CONCLUSIONSCMS’s CED paradigm, although laudable, is imple-

mented on an irregular basis. Ethical considerations and

1

1

1

1

1

1

1

2

2

2

2

2

Relyea-Chew/Ethical Considerations in CED 841

resistance to the process have slowed efforts by CMS togenerate sound evidence to inform appropriate coveragedecisions for novel or emerging technologies.

REFERENCES

1. Makower J, Meer A, Denend L. FDA impact on U.S. medical technologyinnovation: a survey of over 200 medical technology companies. Availableat: http://www.advamed.org/NR/rdonlyres/040E6C33-380B-4F6B-AB58-9AB1C0A7A3CF/0/makowerreportfinal.pdf. Accessed November 10,2011.

2. Social Security Act. 42 USC § 1395y(a)(1)(A) (2011). Exclusions fromcoverage and Medicare as secondary payer.

3. Medicare program: criteria for making coverage decisions (42 CFR Part405; Part II, section A, page 31127). Available at: http://www.gpo.gov/fdsys/pkg/FR-2000-05-16/pdf/00-12237.pdf. Accessed November 10,2011.

4. ClinicalTrials.gov. Carotid Revascularization Endarterectomy VersusStenting Trial. Available at: http://clinicaltrials.gov/ct2/show/NCT00004732. Accessed October 6, 2011.

5. Brott TG, Hobson RW II, Howard G, et al. Stenting versus endarterec-tomy for treatment of carotid artery stenosis. N Engl J Med 2010;363:11-23.

6. Centers for Medicare and Medicaid Services. Medicare approved facili-ties/trials/registries: overview. Available at: https://www.cms.gov/MedicareApprovedFacilitie. Accessed October 19, 2011.

7. National Oncologic PET Registry. Home page. Available at: http://www.cancerpetregistry.org. Accessed October 6, 2011.

8. Centers for Medicare and Medicaid Services. Medicare approved facili-ties/trials/registries: artificial heart clinical study approvals. Available at:http://www.cms.gov/MedicareApprovedFacilitie/06_artificialhearts.asp.Accessed October 19, 2011.

9. Centers for Medicare and Medicaid Services. Medicare approved facili-ties/trials/registries: Long Term Oxygen Treatment Trial. Available at:http://www.cms.gov/MedicareApprovedFacilitie/02_o2trial.asp. AccessedOctober 19, 2011.

0. Fishman A, Martinez F, Naunheim K, et al; National Emphysema Treat-ment Trial Research Group. A randomized trial comparing lung-volumereduction surgery with medical therapy for severe emphysema. N EnglJ Med 2003;348:2059-73.

1. Tunis SR, Pearson SD. Coverage options for promising technologies:Medicare’s “coverage with evidence development.” Health Aff (Mill-wood) 2006;25:1218-230.

2. Centers for Medicare and Medicaid Services. Draft guidance for thepublic, industry, and CMS staff: factors CMS considers in making adetermination of coverage with evidence development. Available at:http://www.cms.gov/coverage/download/guidanceced.pdf. Accessed Octo-ber 19, 2011.

3. Centers for Medicare and Medicaid Services. Coverage with evidence develop-ment. Available at: http://www.cms.gov/medicare-coverage-database/overview-and-quick-search.aspx. Accessed October 19, 2011.

4. Centers for Medicare and Medicaid Services. National coverage determi-nations with data collection as a condition of coverage: coverage withevidence development. Available at: https://www.cms.gov/center/covera-ge.asp. Accessed November 9, 2011.

5. Records maintained on individuals. 5 USC § 552a (2011).

6. General administrative requirements 45 CFR 160 (2011).

17. Security and privacy. 45 CFR 164 (2011).

18. Protection of human subjects. 45 CFR 46 (2011).

19. Research on outcomes of health care services and procedures. 42 USC §1320b-12 (2011).

20. National Commission for the Protection of Human Subjects of Biomed-ical and Behavioral Research. The Belmont report: ethical principles andguidelines for the protection of human subjects of research. Available at:http://www.hhs.gov/ohrp/humansubjects/guidance/belmont.html. No-vember 9, 2011.

1. Miller FG, Pearson SD. Coverage with evidence development: ethicalissues and policy implications. Med Care 2008;46:746-51.

2. Noah L. Coerced participation in clinical trials: conscripting humanresearch subjects. Admin L Rev 2010;62:329-66.

3. Dhruva SS, Redberg RF. Variations between clinical trial participants andMedicare beneficiaries in evidence used for Medicare National CoverageDecisions. Arch Intern Med 2008;168:136-40.

4. Centers for Medicare and Medicaid Services. Cost-effectiveness of CTcolonography to screen for colorectal cancer. Available at: http://www.cms.gov/medicare-coverage-database/details/technology-assessments-details.aspx?TAId�58&bc�BAAgAAAAAAAA&. Accessed October 19,2011.

5. Kanal E, Barkovich AJ, Bell C, et al. ACr guidance document for safe MRpractices: 2007. Available at: http://www.acr.org/SecondaryMainMenuCategories/quality_safety/MRSafety/safe_mr07.aspx. Accessed Novem-ber 9, 2011.

26. The Joint Commission. Sentinel event alert: preventing accidents andinjuries in the MRI suite. Available at: http://www.jointcommission.org/assets/1/18/SEA_38.PDF. Accessed October 19, 2011.

27. Centers for Medicare and Medicaid Services. Pub 100-03 Medicare na-tional coverage determinations: magnetic resonance imaging (MRI) inMedicare beneficiaries with implanted permanent pacemakers (PMs) orimplantable cardioverter defibrillators (ICDs). Available at: http://www.cms.gov/transmittals/downloads/R132NCD.pdf. Accessed October 19,2011.

28. Centers for Medicare and Medicaid Services. Decision memo for mag-netic resonance imaging (MRI) (CAG-00399R3). Available at:https://www.cms.gov/medicare-coverage-database/details/nca-details.aspx?NCAId�252&ver�3&NcaName�Magnetic�Resonance�Imaging�(MRI)�(3rd�Recon)&bc�BEAAAAAAAA&. Accessed October 19,2011.

29. Wikoff BL, Bello D, Taborsky M, et al. Magnetic resonance imaging inpatients with a pacemaker system designed for the magnetic resonanceenvironment. Heart Rhythm. Available at: http://www.heartrhythmjournal.com/article/S1547-5271%2810%2901079-9. Accessed October 19,2011.

30. Wadman M. Medicare compels heart patients to enlist in follow-upresearch. Nature 2005;433:341.

31. Beauchamp TL, Childress JF. Principles of biomedical ethics. 6th ed.New York: Oxford University Press; 2009.

32. Berg JW, Appelbaum PS, Lidz CW, et al. Informed consent: legal theoryand clinical practice. 2nd ed. New York: Oxford University Press; 2001.

33. Carnahan SJ. Medicare’s coverage with study participation policy: clinicaltrials or tribulations? Yale J Health Pol L Ethics 2007;7:229-72.

34. Pearson SD, Miller FG, Emanuel EJ. Commentary: Medicare’s require-ment for research participation as a condition of coverage. JAMA 2006;296:988-91.

35. de Melo-Martin I, Ho A. Beyond informed consent: the therapeuticmisconception and trust. J Med Ethics 2008;34:202-5.