Learning from Medicare: Coverage Policy

July 2011

The rapid rate of technological innovation in the U.S. health care system is often cited as a leading driver of both escalating health care costs and supply-induced demand for services.1 As the insurer of more than 47 million Americans,Medicare’s decisions about the coverage of new technology have the potential to impact patterns of care across the country. From the perspective of policymakers, Medicare’s coverage determination process is also a potential tool in promoting evidence-based medicine, reducing geographic variations in care, and decreasing the amount of money spent on unnecessary or unproven care.Because of its size, Medicare’s actions also exert significant influence on the commercial insurance market. Many private insurers adopt Medicare’s coverage decisions.4 HCFO-funded research provides valuable insights to policymakers on the nature and impact of Medicare’s coverage process as well as its impact on the broader health care landscape.

Background: The Evolution of the Coverage Determination Process

When creating Medicare in 1965, Congress did not foresee the degree to which medical care and technology would rapidly change in the coming decades. The coverage determination process evolved out of Medicare’s statutory requirement to cover “reasonable and necessary” services.  This remains the standard for Medicare’s coverage of diagnostic and therapeutic services today, although some preventive services are now covered outright by statute. Private contractors, who process Medicare claims, assume the responsibility of determining if particular claims meet the “reasonable and necessary” standard. In the early years of the program, these contractors executed little oversight of claims, solving disputes informally and usually deferring to providers.5 Although the vast majority of coverage decisions are still made by local contractors,6 the Centers for Medicare and Medicaid Services (CMS) has developed a national coverage determination (NCD) process that it invokes in certain situations to determine if Medicare should cover particular technologies and procedures.

The NCD process is typically reserved for issues that affect a large number of Medicare beneficiaries or have the greatest overall impact on Medicare.The process is started by a manufacturer or professional organization appealing directly to CMS, or CMS determining that a particular treatment rises to the level of importance to set policy at a national level.8 In the case of topics that are controversial or have the potential for a major impact on Medicare, CMS may refer cases to the Medicare Evidence Development and Coverage Advisory Committee (MedCAC). The role of MedCAC is to provide independent, expert advice to CMS and make recommendations on the strength of scientific evidence. CMS selects up to 100 experts to serve on the MedCAC from fields such as the biologic and physical sciences, clinical medicine, health care economics, and medical ethics. The members review and evaluate medical literature, technology assessments, and data on the effectiveness of health services that are covered under Medicare or may be eligible for coverage.9 If requested by CMS, the Agency for Healthcare Research and Quality (AHRQ) provides external expertise in the form of technology assessments (TAs) to inform national coverage decisions.10 The TA process is reserved for coverage decisions that are too extensive or complex to be reviewed internally. Examples of assistance that AHRQ might provide are clinical knowledge of a specific area and expertise in the utilization of complex methods or modeling.11

At the conclusion of the NCD process, CMS posts a memorandum that explains how the agency reached its decision. The final decision contains formal instructions to contractors on how to process claims and outlines any changes to payment policy.12 An example of a recent NCD that attracted significant media attention was the approval of the drug Sipuleucel-T for men with advanced prostate cancer.13 During the period from 1999 to 2007, there were 119 coverage decisions. Of these, 60 percent were covered completely or covered with restrictions on clinical conditions, patient population, or care setting. Seven percent were left to local contractor discretion, and 33 percent were declared not eligible for Medicare coverage at all.  NCDs during this time addressed a wide array of technologies, with the most common being medical procedures (24 percent).14 

The Medicare Modernization Act of 2003 established the Council on Technology and Innovation (CTI) at CMS to coordinate coverage, coding, and payment processes for new technologies and procedures. The CTI supplements established processes and is intended to improve the quality of medical decisions and speed beneficiary access to effective new treatments. It is composed of senior CMS staff and clinicians and reports to the CMS Quality Council. The goal of the CTI is to ensure that the work of different entities—mainly other CMS offices involved in the NCD process—reflects CMS’s priority to provide high quality care and also to promote communication and coordination to keep stakeholders up to date.15

CMS further refined its coverage process in 2006 by introducing formal guidance through a new policy called “coverage with evidence development” (CED). CED does not offer a final “yes” or “no” answer on Medicare coverage, but conditions payment for a new technology or treatment on the collection of additional evidence.16 CED arose from the tension between strict evidence-based coverage decisions and the need to be responsive to innovation and emerging technologies that might benefit patients.17 Additionally, CED partially arose out of CMS frustration over evidence that did not meet its needs for coverage determination. The Medicare population—the elderly and disabled—are underrepresented in clinical trials. Also, most trials do not compare treatments with currently covered options. This need to generate more relevant evidence for coverage and other health policy decisions led CMS to use CED to promote the development of evidence relevant to its primary populations.18 Many health care stakeholders such as policymakers, health plans, and purchasers support CED to encourage the development of stronger evidence regarding the use of new technologies.19 Since 2006, CED has only been used twice. Medicare’s use of this tool is limited by specific statutory language giving Medicare the authority to limit coverage to patients enrolled in studies.20

HCFO Research on the Coverage Determination Process

HCFO-funded research has evaluated several facets of both the local and national coverage determination process.  Susan Bartlett Foote, J.D., of the University of Minnesota, examined coverage determination in two separate HCFO studies. In the first, she sought to inform future Medicare coverage policy by examining variations in local medical review policies (LMRPs) for new technologies and procedures. Foote and her colleagues examined the policies and classified them into three categories: new technology, extensions of new technology, and utilization management of widely used procedures. They identified policies addressing the same procedure or technology and required a sample of at least 20 separate policies from different contractors in order to make comparisons. They found substantial similarity among the coverage decisions by different contractors on the same technology or procedure in the categories of new technology and technology extension. There was significantly more variation in the decisions relating to utilization management.21 These results illustrated the differences between local contractors’ decisions and raised questions about the effect of this variation. At a practical level, the variation in LMRPs means that even beneficiaries in close proximity geographically might not have access to the same treatments, raising questions about fairness in the program.22

Foote also conducted an in-depth review of the evolution of the local coverage process to inform the debate on the merits of the local system.23 She found that during the 1990s the local contractor market was characterized by increased consolidation and departures of firms. These changes in the market led to large networks without any regional coherence. Her analysis raised important questions about the local system. First, while proponents see advantages in having coverage decisions decentralized through local contractors, Foote concluded that the current state of the market is no longer truly “local” and that the presence of multiple evaluators of technology may detract from quality standards. Second, is it appropriate to have different standards of evidence for the adoption of new technology, or does that introduce too much variation? Third, while proponents of the local process point to its responsiveness and timeliness, is the local coverage process only perceived as timelier because evidence standards are lower?24 

In her second HCFO-funded project on the coverage process, Foote answered some of the questions raised in her first project by examining the impact of Medicare coverage policies on health care utilization. She and her colleagues examined Medicare claims from 1999 to 2002 for eight different treatments and technologies where there was significant variation across local coverage decisions. They found no evidence that coverage policies systematically affect utilization.25 In a separate review of LMRPs, Foote and her colleagues examined variations in the use of scientific evidence in coverage decisions. They found that the frequency and type of citations varied widely, indicating that contractors take different approaches to evaluating scientific evidence. This result suggested a potential role for CMS in triaging coverage decisions at the national level to improve decision making and satisfy the quality and equity goals of Medicare.26 In 2003, CMS instructed contractors to create local coverage decisions (LCDs) in addition to LMRPs. LCDs contain only reasonable and necessary information, while LMRPs also contained benefit category and statutory exclusion provisions.27

While Foote’s work documented variations in the local process, another HCFO project undertaken by Peter Neumann, Sc.D. (at the time at Harvard University) examined the same issue for NCDs. Reviewing 69 CMS coverage decisions between January 1998 and August 2003, Neumann and colleagues found that coverage decisions have generally been consistent with scientific evidence, and that the coverage determination process is increasingly evidence-based and transparent. However, their research also revealed that there are gaps in the evidence base which limit the quality of evidence available to CMS. Neumann and his colleagues suggested that there is a need for greater investment in “pragmatic” clinical trials to aid decision makers working in the areas of coverage policy and technology assessment.28 Improvement in this process and filling in the evidence gaps is also valuable for the private sector. HCFO-funded research led by Linda Bergthold, Ph.D., of Stanford University, investigated medical necessity decision-making in private health plans. She and her colleagues surveyed plan medical directors from across the country and found that they favor the use of technology reports from private organizations and Medicare, and consult randomized controlled trials when possible.29


The ongoing rapid development of medical technology and treatments will keep the NCD process in the spotlight. Research on Medicare’s coverage determination policy can provide information to policymakers the evolution of the coverage process and its consequences for the U.S. health care system. Details on HCFO studies and related publications can be found at http://www.hcfo.org.

Related HCFO Grants

Title: Evaluation of Medicare’s Local Medical Review Policies for New Medical Technologies
Grantee Institution: University of Minnesota
Principal Investigator: Susan Bartlett Foote, J.D.
Grant Period: May 1, 2001-December 31, 2003

How does variation in coverage decisions affect access to new technologies and equity for beneficiaries in the Medicare program? Are Medicare’s Local Medical Review Policies (LMRPs) in need of reform? LMRPs are one of two ways that HCFA evaluates new technologies and procedures in order to make coverage determinations. According to the researchers, a small number of technologies are reviewed through HCFA’s national process, where decisions are made uniformly across the country. However, the majority of such coverage decisions are made by local carriers and intermediaries under contract to HCFA, with the LMRPs binding only in the local jurisdiction. The researchers evaluated variations in LMRPs by analyzing: 1) the players - who participates in and influence decisions; 2) the process - how decision-making procedures differ; 3) the evidence - how evidence of value is solicited and measured; and 4) the outcomes - measuring and mapping timing and content patterns. The objective of the study was to provide policymakers considering the virtues of a more uniform national coverage policy, relative to a more flexible local policy, with better information about the extent of LMRP variation, the sources of variation, and the implications for flexible decision making, beneficiary access, and Medicare equity.

Title: Impact of Medicare
Grantee Institution: University of Minnesota
Principal Investigator: Susan Bartlett Foote, J.D.
Grant Period: November 1, 2003-October 31, 2006

How do Medicare coverage policies, under national coverage determinations (NCDs) and local medical review policies (LMRPs), affect claims, access and cost? The researchers examined eight procedures that fall into three policy categories: new technology, extension of existing technology, and utilization management. The study included an examination of changes in the use of the eight procedures over the period 1999-2001 to answer the following research questions: 1) Do LMPRs or NCDs affect local practice patterns in Medicare? 2) Are there different effects for distinct categories of policies? 3) If an NCD applies to all providers and beneficiaries, can we expect consistency in utilization patterns in Medicare following the implementation of a national policy? 4) If LMRPs are consistent across local contracts, should we see consistency in utilization post-policy implementation? 5) To the extent that LMRPs applicable to the same procedures vary from region to region, should we expect to see variations in practice that reflect these policy differences? and 6) If variations in utilization persist in light of similar policies, how can those variations in practice patterns be explained? The objective of this study was to provide insights to policy makers on the effectiveness of coverage policies, the appropriate balance between national and local decisions, and implications for efforts to enact contractor reform.

Title: Cost Effectiveness, Quality, and the Future of Medical Technology Asssessment
Grantee Institution: Harvard School of Public Health
Principal Investigator: Peter Neumann, Sc.D.
Grant Period: July 1, 2002-June 30, 2004

How does Medicare assess and make coverage decisions for new medical technologies? First, the researchers conducted an in-depth descriptive and multi-variate explanatory analysis of 100 CMS coverage decisions over the past 12 years. They then compared Medicare's processes and decisions with those of other health technology assessment (HTA) organizations. Specifically, the researchers examined 6 key questions: 1) What technologies has Medicare formally assessed in the past decade? 2) What are the key determinants of Medicare coverage decisions? 3) Have coverage decisions been consistent with evidence of societal cost-effectiveness? 4) Have the same technologies been assessed by other HTA organizations? 5) What "best practices" for technology assessment surface from an investigation of key technology assessment organizations in the U.S. and abroad? and, 6) What role can cost-effectiveness play in future assessments in the U.S. (given data limitations, multiple societal objectives, and likely political opposition)? The objective of the study was to inform decision makers about HTA processes and to reveal "best practices" about technology assessment as they consider whether to cover new medical technologies.

Title: Understanding Medical Necessity Decision Making
Grantee Institution: Stanford University
Principal Investigator: Linda Bergthold, Ph.D.
Grant Period: June 1, 2000-October 31, 2001

How do policies regulating medical necessity decision-making influence national health plans? Researchers from Stanford University attempted to answer the following questions: 1) Who are the medical necessity decision makers? 2) How are the terms defined and what information do decision makers use in making their decisions, including what type of evidence and cost effectiveness information is considered? 3) What procedures do health plans use to communicate with physicians and patients throughout the decision making process and to track and use coverage decisions for quality improvement? 4) How can variation in terminology and application of guidelines be reduced? 5) What is the role of accreditation, regulation, legislation, and organizational policies and procedures in promoting clearer definitions and more consistent decision making? and 6) How do the answers to the first five questions vary by the size of the health plan, its tax status, degree of management of care, or geographic region? The objective of this project was to test the findings of a similar project recently completed in California and provide a better understanding of medical necessity decision-making to state and national policymakers.


1. Bodenheimer, T. “High and Rising Health Care Costs. Part 2: Technologic Innovation,” Annals of Internal Medicine, Vol. 144, No. 11, 2005, pp. 932-37.
2. Kaiser Family Foundation, Medicare Chartbook, Fourth Edition, 2010. See also:
3. Foote, S.B., and Town, R.J. “Implementing Evidence-Based Medicine Through Medicare Coverage Decisions,” Health Affairs, Vol. 26, No. 6, 2007, pp. 1634-1642.
4. Vance, A. “Insurers Shun Multitasking Speech Devices,” The New York Times, A1, September 14, 2009. 
5. Foote and Town, 2007. 
6. Centers for Medicare & Medicaid Services (2010). Innovators’ Guide to Navigating Medicare, See also:
7. Medicare Payment Advisory Commission (2003). An Introduction to How Medicare Makes Coverage Decisions. Washington, DC: MedPAC. 
8. Gillick, M.R. “Medicare Coverage for Technological Innovations—Time for New Criteria?,” The New England Journal of Medicine, Vol. 350, No. 21, 2004, pp. 2199-2203. 
9. Centers for Medicare & Medicaid Services (2011). Medicare Evidence Development & Coverage Advisory Committee. See also:
10. Agency for Healthcare Research and Quality (2011). Technology Assessments. See also: http://www.ahrq.gov/clinic/techix.htm.
11. Centers for Medicare & Medicaid Services (2003). “Guiding Principles for When National Coverage Determination Topics are Referrred for External Expertise via Technology Assessment and/or the Medicare Coverage Advisory Committee.” See also:
12. CMS, 2010. 
13. Pollack, A. “Medicare Will Cover 2 Expensive Cancer Drugs,” The New York Times, p. B7, June 30, 2011. 
14. Neumann, P.J. et al. “Medicare’s National Coverage Decisions for Technologies, 1999-2007,” Health Affairs, Vol. 27, No. 6, 2008, pp. 1620-1631.
15. Centers for Medicare & Medicaid Services (2011). Overview: Council on Technology and Innovation. See also:
16. Kary, W. et al. “Coverage for Evidence Development: A Conceptual Framework,” The Center for Medical Technology Policy, Issue Brief, June 2009. See also:
17. Tunis, S.R. and Pearson, S.D. “Coverage Options for Promising Technologies: Medicare’s Coverage with Evidence Development,” Health Affairs, Vol. 25, No. 5, 2006, 1218-1230.
18. Mohr, P.E. and Tunis, S.R. “Access with Evidence Development: The US Experience,” Pharmacoeconomics, Vol. 28, No. 2, 2010, pp. 153-162.
19. Kary, 2009. 
20. Mohr and Tunis, 2010. 
21. Foote S.B. et al. “Variation in Medicare’s Local Spending Policies: Content Analysis of Local Medical Review Policies,” The American Journal of Managed Care, Vol. 11, No. 3, 2005, pp. 181-187.
22. Hawryluk, M. “Local commotion: The flip side of national Medicare policy-making,” American Medical News, August 5, 2003. See also:
23. Foote S.B. et al. “Variation in Medicare’s Local Spending Policies: Content Analysis of Local Medical Review Policies,” The American Journal of Managed Care, Vol. 11, No. 3, 2005, pp. 181-187.
24. Foote, 2003.
25. Foote, S.B. et al. “The Impact of Medicare Coverage Policies on Health Care Utilization,” Health Services Research, Vol. 43, No. 4, 2008, pp. 1285-1301.
26. Foote, S.B. et al. “Resolving The Tug-Of-War Between Medicare’s National and Local Coverage,” Health Affairs, Vol. 23. No.4, 2004, pp. 108-123.
27. CMS, 2010. 
28. Neumann, P.J. et al. “Medicare’s National Coverage Decisions, 1999-2003: Quality of Evidence and Review Times,” Health Affairs, Vol. 24, No. 1, 2005, pp. 243-254. 
29. Austin, B. (2003). How Do MCOs Decide if an Intervention is Medically Necessary?, Findings Brief, Washington, DC: AcademyHealth.