Health Affairs, 23, no. 4 (2004): 108-123 doi: 10.1377/hlthaff.23.4.108 © 2004 by Project HOPE	New Online   Obama's Success & Next Steps   3 Health Policy Options for President Obama   A Hybrid Proposal   Bipartisan Compromise

This Article Abstract Figures Only Reprint (PDF) Submit a response to this article Alert me when this article is cited Alert me when Comments are posted Alert me if a correction is posted Services E-mail this article to a friend Similar articles in this journal Similar articles in ISI Web of Science Similar articles in PubMed Alert me to new issues of the journal Add to My Personal Archive Download to Citation Manager Reprints & Permissions Citing Articles Citing Articles via ISI Web of Science (10) Citing Articles via Google Scholar Google Scholar Articles by Foote, S. B. Articles by Halpern, R. Search for Related Content PubMed PubMed Citation Articles by Foote, S. B. Articles by Halpern, R. Related Collections Insurance Coverage Legal/Regulatory Issues Medicare Business Of Health State/Local Issues Politics Elderly

Medicare

Resolving The Tug-Of-War Between Medicare’s National And Local Coverage

PROLOGUE: As a federal entitlement, Medicare theoretically offers equal benefits to all forty million enrollees. But the program’s design subtly allows some beneficiaries better benefits than others. A clear example is the design of its managed care program, under which differences in payment rates have allowed plans in higher-spending regions to offer better benefit packages than those in lower-spending regions, while beneficiaries in regions with no managed care options at all have been unable to take advantage of any of the enhanced benefits offered to managed care enrollees.

Less well known, and potentially more damaging to the program’s credibility, are the differences in coverage under the fee-for-service program that result from differences in local coverage policies developed by Medicare’s administrative contractors. Medicare makes only a few national medical technology decisions a year. Many policy experts argue in favor of national coverage policies as being more efficient and fair. Medicare, however, is supported by the medical technology industry in arguing against this policy change, believing that there is no proof of inequities and that the current national process is too cumbersome.

In this paper Susan Foote, one of the nation’s foremost experts on these policies, and her colleagues analyze the current local medical review policies and survey contractors to determine how many of such policies exist; what procedures they affect; whether the policies affect new technology or an extension of an existing technology, or are related to utilization management; and what data sources the contractors used to support their policies. The authors provide some policy recommendations to streamline the medical review process.

Foote is an associate professor and head of the Division of Health Services Research and Policy at the University of Minnesota’s School of Public Health in Minneapolis. She wrote about the evolution of Medicare’s coverage policy in the July/August 2003 issue of Health Affairs. Joining her as coauthors are three colleagues from the same division: Douglas Wholey, a professor; Todd Rockwood, an assistant professor; and Rachel Halpern, a research assistant.

Abstract

 
Medicare’s decentralized local coverage policy process leads to policy variation, raising serious equity and quality issues. The policy debate resembles a tug-of-war, with advocates favoring nationalization of all local policies or arguing for the status quo. We extensively analyzed thousands of local policies and surveyed Medicare’s contractors. We found that all local policies are not the same. We classified them based on where they fall on the diffusion curve. The classification by type allows for reallocation to the national or local process to improve the decisions and satisfy Medicare’s equity and quality goals.

The policy debate about local versus national coverage resembles a tug-of-war, with advocates lining up on either side. On one side, the Medicare Payment Advisory Commission (MedPAC) has called for the elimination of local policies, to reduce "current complexity, inconsistency, and uncertainty."³ A recent report from the U.S. General Accounting Office (GAO) criticized the divided authority and also concluded that Medicare should not develop any new local coverage policies.⁴ On the other side, the Department of Health and Human Services (HHS) was not persuaded by the GAO report, stating that there was an inadequate analytic base to link inequities to the local process or to justify a new, centralized one.⁵ AdvaMed, the device industry association, also disputed the findings, declared that the national process was too slow, and urged preservation of the status quo.⁶

The Medicare Prescription Drug, Improvement, and Modernization Act (MMA) of 2003 calls for HHS to develop a plan to determine the need for greater consistency and less duplication among local Medicare contractors. MMA provides an opportunity to resolve the tug-of-war. However, little systematic research exists to guide the process.

This paper presents the results of our analysis of local policies. We developed a comprehensive database of LMRPs, and we conducted a survey of contractor medical directors (CMDs) who administer local contracts. We found that policies differ depending on where the technology or procedure falls on the diffusion curve at the time of evaluation.⁷ We identified three policy types: new technology (NT), extensions of covered technology (TE), and utilization management (UM). Despite the nearly exclusive focus on new technology in the debate, we estimate that the vast majority of LMRPs deal with management of common procedures. Each policy type requires different expertise, evidence, and evaluation, and each raises different considerations in terms of equity for beneficiaries and quality of care.

Research Design

Top Research Design Study Results Discussion Policy Recommendations Editor's Notes NOTES

 
LMRP database. We studied policies issued by Part A fiscal intermediaries (FIs) and Part B carriers. We downloaded more than 8,000 carrier and FI LMRPs from the CMS-sponsored Web site on 31 May 2001, our snapshot date.⁸ We coded eleven data points for each LMRP.⁹ We excluded policies if they fell outside the time frame of the study, focused on process rather than on technology, had no recorded procedure codes, or were duplicates of existing policies.¹⁰ Our final data set of 6,900 policies comprised 5,213 carrier LMRPs and 1,687 FI LMRPs.

LMRP classification. We observed differences among policies and designed a classification scheme. Two physician consultants (MD 1 and MD 2) reviewed the carrier LMRP titles and classified each into one of our three categories (NT, TE, and UM). We provided the following definitional guidelines: (1) NT policies provide guidance for and limitations on the use of new clinical interventions; (2) TE policies provide guidance for and limitations on new uses of procedures or technologies already covered for other uses; and (3) UM policies provide guidance for and limitations on widely diffused technologies or procedures to avoid misuse or overuse.

We classified the LMRPs based on the policies’ original effective dates to capture their status at the time of implementation. The consultants first classified the carrier LMRPs, then classified FI LMRPs based on the carrier classifications.¹¹ Although the majority of FI policies were successfully classified, 264 FI policies contained procedure codes from multiple classes. Of those, 216 were matched with carrier policies and assigned consistent categories. The consultants then classified the remaining forty-eight policies.

MD 1 classified the LMRPs as 88.6 percent UM, 8.4 percent NT, and 3 percent TE. MD 2 classified LMRPs as 73.1 percent UM, 16.1 percent NT, and 10.8 percent TE. Where relevant in the analysis, we report the results based on MD 1 and MD 2 classifications separately, with the upper and lower range for each category. A cross-classification analysis revealed that the consultants agreed on classification for 83.8 percent of the policies in the database. We understood that the consultants might not agree on every classification because technology diffusion is dynamic. However, their agreement on 83.8 percent of the policies underscores the fundamental validity of the classification scheme.

CMD survey. We surveyed sixty-seven CMDs, the physicians who manage the contracts. There were forty-eight participating organizations (twenty-eight carriers and twenty FIs), with sixty-seven CMDs administering one or more contracts. We developed the initial draft with external consultants and conducted cognitive interviews to evaluate it.¹² The final instrument contained sixty-six questions and was administered by mail with telephone follow-up. The overall response rate was 60 percent (forty respondents). Sixteen of the twenty-eight FIs responded, for a 57 percent response rate; twenty-four of the thirty-nine carriers responded, for a 62 percent response rate.¹³ We investigated differences between respondents and nonrespondents by comparing them in size (claims processed), geography, and organizational form (multiple-state or single-state contractors). Our respondents reflected the proportion of FIs and carriers in the total population. Additionally, 80 percent of FIs and 73 percent of carrier respondents reported that their organizations had multiple contracts, which is also consistent with the population as a whole.

Study Results

Top Research Design Study Results Discussion Policy Recommendations Editor's Notes NOTES

 
All contractors must follow CMS mandates for the development of LMRPs. The Medicare Program Integrity Manual, along with regularly updated program memoranda, contains many provisions governing LMRP development. These include identification of services for which a new or revised LMRP is needed, when discretion may be used to develop LMRPs, techniques for writing LMRPs, evidence and coding requirements, process for public comment, and the format for posting policies. The CMS also governs the conduct of Carrier Advisory Committees (CACs). Carriers are required to have CACs; FIs have the option of creating similar advisory bodies.¹⁴ Although all contractors must follow these directives, there is considerable variation among them.

Variation in resources. The size and organizational structure of contractors vary.¹⁵ Many contractors service small, single-state contracts; others preside over large states or multistate networks.¹⁶ Size, as measured by volume of claims processed, also varies greatly (Exhibits 1 and 2). The largest carrier processed 82.6 million claims in 2001, while the smallest processed only 2.1 million. Because contractors are paid based in large part on numbers of claims processed, one can infer sizable resource differences among contractors.

However, respondents perceived variations in total available resources based on organizational differences. For example, 59 percent of respondents reported that organizations with multiple-state networks, which are generally larger than single-state contractors, facilitated LMRP development. More than half also reported that multiple-state contractors had more access to scientific or medical expertise than single-state organizations had. Similarly, CMDs with single-state contracts perceived themselves at a disadvantage relative to contractors who are part of networks: 80 percent reported inadequate staffing, compared with only 51.7 percent of CMDs in multistate organizations.

CMDs’ self-reported data supported these perceptions of the adequacy of trained personnel to develop LMRPs. When we asked how many full-time equivalents (FTEs) worked on LMRP development under the contract, the mean was 3.1, median was 2, and mode was 1. The range was 0.5 FTE to 9 FTEs. Respondents with single-state contracts frequently reported some fraction of their own time plus that of a nurse reviewer as their sole LMRP development resource. Respondents from multistate networks reported the largest numbers of FTEs available for LMRP development, and many noted the support of corporatewide medical policy committees to coordinate work on LMRPs.

CMDs supplement internal resources in several ways. Seventy-one percent reported that CACs were useful, and 26 percent reported that they were somewhat useful, in LMRP development. Similarly, 65 percent reported that CACs had "a lot" of impact on LMRPs, and 32 percent reported that they had "some" impact. Only 3 percent noted that CACs had "a little" impact.

Contractors also collaborate to overcome resource constraints. Seventy-three percent of respondents reported almost always reviewing other LMRPs when deciding whether to draft a policy in their own jurisdiction. All respondents reported some level of communication with other contractors in developing LMRPs.

Contractors do not have extensive access to external assessment expertise. In the private sector, insurers have internal assessment capacity and often purchase technology assessments performed by private organizations, such as ECRI or the Blue Cross/Blue Shield TEC.¹⁸ One former contractor commented publicly that her large multistate organization lacked the resources to purchase formal assessments and had no internal capacity to replicate these studies.¹⁹

Variation in productivity. Policy productivity among contractors differs (Exhibits 1 and 2). Contractors are listed in descending order by size, as measured by claims processed. For each contract, we show the number of policies effective in the jurisdiction. The data show variation in the number of policies within organizations and across organizations. It does not appear that size is closely correlated to number of policies. We also measured the number of unique Healthcare Common Procedure Coding System (HCPCS) codes to determine if the differences in the number of policies might obscure policy-writing styles. Because contractors have discretion whether or when to issue LMRPs, it is not surprising that there is variation in their style. For example, some contractors may issue policies that are narrow in scope, with each policy containing relatively few HCPCS codes. Others may bundle many HCPCS in one policy. The far right columns in Exhibits 1 and 2 show total HCPCS codes, indicating that there is extensive variation in the number of procedures subject to policy limitations or specifications.

Variation in use of evidence. One measure of rigorous evaluation of technologies is citation of evidence. The CMS requires contractors to cite the evidence used to support their coverage decisions. CMDs demonstrated familiarity with the hierarchy of evidence. They reported that scientific journals and research in the public sector had the highest credibility and found information from manufacturers the least credible. Seventy-two percent also reported that they almost always relied on scientific journals in developing LMRPs.

Despite reported knowledge of evidence and assertions about frequency of use, our findings reveal that evidence citations vary. Exhibit 3 shows the percentage of policies in a local jurisdiction citing an evidence source at least once. Policies may cite more than one type of evidence. Our coders classified as "journal articles" citations to peer-reviewed professional and academic publications; "texts" are medical textbooks, and "references," other professional sources. "Workgroups" refer to informal collaborations among contractors to draft policies.²⁰

Exhibit 4 shows citations based on policy type. NT policies focus on undiffused new technologies. To get approval from the Food and Drug Administration (FDA), these technologies must be supported by rigorous, peer-reviewed scientific evidence. By contrast, UM policies focus on commonly used procedures. Support for a UM policy is likely to be based on information in textbooks or practice guidelines. TE policies may use one or both types of evidence. NT and TE policies more frequently cite journal articles, while UM policies rely more heavily on different sources of evidence. Also, fewer than half of all NT policies cited at least one journal article. We infer from this that local contractors do not conduct extensive assessments but instead appear to rely on other LMRPs.

Timing comparisons are challenging because of the decentralized local system. Do we count only the first local policy covering a specific technology, or do we wait until all local jurisdictions have a policy in place, or until a majority does? Whatever the measure, the multiple processes take time. Each jurisdiction must follow CMS-prescribed steps. One contractor reported that expeditious development time for a noncontroversial LMRP is six to nine months, with additional time required for publication, comment, and promulgation. More complicated or controversial policies can take much longer.²² A policy diffusion curve plots the spread of LMRPs across jurisdictions (Exhibits 5 and 6).

View larger version (25K): [in this window] [in a new window]   EXHIBIT 5 Diffusion Of New Technology (NT) And Technology Extension (TE) Policies, 1989–2001

View larger version (22K): [in this window] [in a new window]   EXHIBIT 6 Diffusion Of Utilization Management (UM) Policies, 1989–2001

Some contractors appear more likely to lead in developing NT policies. We identified all NT HCPCS codes by whether they were typically mentioned in NT policies. Then, for every contractor, we determined the earliest data on which an LMRP was initiated for an NT HCPCS code. Clear leaders emerged (Exhibit 7). We found that contractor size and NT policy leadership were positively correlated. The top five FIs shown in Exhibit 7 have the largest claims volume. While the correlation between claims volume and LMRP leadership is somewhat less strong with carriers, three carrier leaders are also in the top five in claims volume. Size appears to be related to leadership in new technology policy development.

Top Research Design Study Results Discussion Policy Recommendations Editor's Notes NOTES

Quality is affected by policy type. High-quality care depends on rigorous evaluation of new procedures, including new technology and new uses of approved technology.²⁴ There is a well-articulated hierarchy of evidence based on the credibility and integrity of research designs.²⁵ Comprehensive technology assessment draws upon formal methods such as meta-analysis and decision analysis and is subject to high standards of evaluation.²⁶ We found that local contractors do not have sophisticated technology assessment expertise and fail to consistently cite evidence to support NT and TE policies. While larger contractors have more resources, even they lack resources to acquire commercially available technology assessments.

Local contractors are more likely to have adequate resources to manage utilization. Contractors uncover utilization issues by reviewing claims data. There is little formal evidence for many commonly used procedures, although there is a growing body of data on best practices. CACs include representatives from statewide medical specialty societies familiar with clinical practice guidelines.²⁷ Variations in UM policies may reflect differences in underlying practice patterns. When contractors perceive problems through claims review, their local policy interventions clarify appropriate use. Ironically, the use of different policies may actually reduce variation in care, not exacerbate it, by bringing regional practitioners more in line with accepted practice. However, some UM policies may accommodate inappropriate practice patterns, raising the need for some oversight and review.

Equity is affected by policy type. Although contractors can approve claims case by case, access to a new technology generally requires a coverage policy because providers are reluctant to perform services if payment is uncertain. Access for beneficiaries will be affected by failure to issue a policy or delays in doing so. UM policies do not raise the same access issues because UM LMRPs address services that are already widely diffused. The issues are less about access and more about appropriate use of an available intervention.

Policy Recommendations

Top Research Design Study Results Discussion Policy Recommendations Editor's Notes NOTES

 
In MMA Congress directs HHS to address the allocation of local and national decisions. The recommendations that follow provide a framework for that task.

The CMS should classify all policies. Because of the importance of Medicare coverage, the CMS should take a proactive role to triage coverage decisions. Classifying policies into NT, TE, and UM provides a critical first step in appropriate allocation of responsibility. Classification is a commonly used technique in approaching the wide range of medical technologies.²⁸

NT and TE policies should be national. Quality and equity require national coverage policies for new technologies. It is more efficient to have one national assessment than thirty to forty local ones, and local resources and expertise are limited. Our findings also suggest that local decisions are not necessarily timely. MMA includes provisions to speed up the national coverage process, addressing concerns expressed by device manufacturers.

HHS has expressed concerns over workload if all policy making were centralized. However, only a small percentage of policies are NT or TE, and many of those address the same technologies. The CMS could defer a TE national decision, allowing for some flexibility at the local level for a limited period of time. Any TE decision, however, should be based on empirical studies of effectiveness.

The CMS should oversee and assist local UM policy development. UM policies raise fewer issues about equity than about quality. Local contractors often discover utilization issues when they review claims. It is logical that they retain this management function to administer policies, pay claims, and educate providers on appropriate utilization issues. However, local UM policies would benefit from greater CMS oversight and guidance. HHS should convene an expert panel to evaluate all existing UM policies. Using clear criteria such as program cost, safety, and quality, the experts could recommend consistent national UM policies where there are reliable data on appropriate parameters for use. The expert body could be available for consultations with local contractors, as well as periodic reviews of future UM policies to ensure consistency where appropriate.

Redesign contracts to ensure equitable allocation. We have found that much variation in contractors’ performance is tied to differences in contractors’ size and resources. MMA contains provisions for contracting reform and indicates a movement toward regional management of health plans. We recommend that regional contractor jurisdictions be redrawn so there are reasonably equivalent resources for LMRP development. The CMS should also provide greater assistance and oversight, including facilitating access to data on best practices and technology assessment where available.

MMA opens an opportunity to resolve the tug-of-war between local and national coverage policy. Our research provides the analytic base needed to accomplish the task. These reforms can eliminate inappropriate variation in Medicare, offering improvements in equity and quality for all beneficiaries.

Editor's Notes

Top Research Design Study Results Discussion Policy Recommendations Editor's Notes NOTES

 
This study was funded by the Robert Wood Johnson Foundation through its Changes in Health Care Financing and Organization (HCFO) initiative.

NOTES

Top Research Design Study Results Discussion Policy Recommendations Editor's Notes NOTES

 

1. S.B. Foote, "Focus on Locus: Evolution of Medicare’s Local Coverage Policy," Health Affairs 22, no. 4 (2003): 137–146.[Abstract/Free Full Text]
2. L. Baker et al., "The Relationship between Technology Availability and Health Care Spending," Health Affairs, 5 November 2003, content.healthaffairs.org/cgi/content/abstract/hlthaff.w3.537 (16 March 2004).
3. Medicare Payment Advisory Commission, Reducing Medicare Complexity and Regulatory Burden (Washington: MedPAC, 2001).
4. U.S. General Accounting Office, Medicare: Divided Authority for Policies on Coverage of Procedures and Devices Results in Inequities, Pub. no. GAO-03-175 (Washington: GAO, 2003).
5. Ibid., Appendix VI.
6. AdvaMed, "AdvaMed Rejects GAO Findings on Medicare Local Coverage; Process Remains Key Avenue for Patient Access to New Medical Technologies," Press Release, 13 May 2003.
7. Institute of Medicine, Assessing Medical Technologies (Washington: National Academies Press, 1985).
8. At the time we downloaded the information for our study, the LMRP Web site was www.lmrp.net; the site is now www.cms.hhs.gov/mcd. Our description relates to the old Web site. The date is important because contracts may be transferred from one carrier organization to another as these organizations merge or leave the program. There were contractors either not listed on the LMRP Web site or whose policy sets were empty on our "snapshot" date. We revisited the Web site and found that most policy sets previously missing on 31 May 2001 were posted. We downloaded these "missing" policy sets on 2 April 2002. We included only those policies with original effective dates on or before 31 May 2001, and we excluded two carriers (Nebraska and Vermont) and one FI (Wyoming) that were posted after the second download.
9. Data points in our coding process included policy title; description; focus (drug, surgery, medicine); start date of comment period; start date of notice period; original effective date; Healthcare Common Procedure Coding System (HCPCS) or Current Procedural Terminology (CPT) codes covered; evidence cited; revision dates and reasons; states covered; and dates and locations of CAC meetings.
10. Excluded categories were anesthesia; ambulance and transportation; professional service components; psychological services "incident to" provision; psychiatric partial hospitalization programs; evaluation and management; consultations; observation care; care plan oversight; discharge planning; concurrent care; emergency room; nursing home; hospice; independent laboratory and diagnostic testing facilities; x-ray supplier; collection of specimens; and payment and facility administration.
11. Medicare’s automated payment systems rely on codes to identify procedures used. Contractors use the HCPCS codes. There are miscellaneous codes if no established codes exist. The HCPCS contains three sets of codes. Level I codes include CPT codes used to identify medical services and procedures furnished by physicians. Level II codes include products, supplies, and services not included in CPT codes, and Level III codes are "local" codes developed by contractors for local use (eliminated in December 2003 pursuant to the Health Insurance Portability and Accountability Act, or HIPAA).
12. S. Sudman, N.M. Bradburn, and N. Schwarz, Thinking about Answers: The Application of Cognitive Processes to Survey Methodology (San Francisco: Jossey-Bass, 1996).
13. Eight CMDs (12 percent) provided written refusals, citing time and resource constraints; two also wanted written approval from the CMS. Others refused to participate when called or did not return calls.
14. CMS, Medicare Program Integrity Manual (Manual Transmittals through Transmittal Number 52, 10 October 2003), www.cms.hhs.gov/manuals/108_pim/pim83toc.asp (16 March 2004).
15. W.R. Scott, Organizations: Rational, Natural, and Open Systems, 5th ed. (Upper Saddle River, N.J.: Prentice Hall, 2003), 263–267.
16. Foote, "Focus on Locus."
17. The activities were LMRP development; provider relations; claims processing, payment, and recovery; administrative; fraud; beneficiary inquiries/appeals; and other.
18. A.M. Garber, "Evidence-Based Coverage Policy," Health Affairs 20, no. 5 (2001): 62–82.[Abstract/Free Full Text]
19. Statement of Charlotte Yeh, former contractor for National Heritage Insurance (New England), Medical Technology Leadership Forum, 29 January 2003.
20. Contractor workgroups develop model policies that are shared with other contractors. See "Device Makers Ask HCFA to Scrap ‘Clandestine’ Policymaking Process," Reuters Health, 4 January 2001.
21. T.G. Thompson, Report to the Congress on National Coverage Determinations (Washington: DHHS, 2002).
22. National Heritage Insurance Company, "Local Medical Review Policy Strategy FY 2002" (Mimeo, 2002).
23. J.E. Wennberg, E.S. Fisher, and J.S. Skinner, "Geography and the Debate over Medicare Reform," Health Affairs, 13 February 2002, content.healthaffairs.org/cgi/content/abstract/hlthaff.w2.96 (16 March 2004).
24. MedPAC, Report to the Congress, Medicare Payment Policy (Washington: MedPAC, March 2004).
25. S.D. Ramsey et al., "The Limited State of Technology Assessment for Medical Devices: Facing the Issues," American Journal of Managed Care 4 (Special Issue, September 1998): SP188–SP199.
26. Garber, "Evidence-Based Coverage Policy."
27. R.A. Rettig, Health Care in Transition: Technology Assessment in the Private Sector (Washington: RAND, 1997), 11.
28. 21 U.S. Code, sec. 321. See the 1976 Medical Device Amendments to the Food, Drug, and Cosmetic Act.

                      What's this?