The Application of Post-Market Registries and other Evidence for Medical Devices - Part II: US Center For Medicare & Medicaid Services (CMS) Perspective: What Is CMS’s View On The Role Of Post-Market Evidence In Coverage For Medical Device?

This is a summary of the presentation given at the ISPOR Medical Device and Diagnostics Council Forum at the ISPOR 10th Annual International Meeting, May 17, 200, Washington, DC, USA.

The US Center for Medicare & Medicaid Services (CMS) wants data that results in good information and allows physicians and patients to make good, informed decisions about healthcare. CMS collects data in many ways. We analyze claims data which has some utility. We sponsor clinical trials. We have paid for services within clinical trials for a number of years. We have coordinated clinical trials with U.S. National Institute of Health and U.S. Agency for Health Care Research and Quality. The new initiative, coverage with evidence development, is another way of collecting data, where we condition our payment on additional data collection.

The term, post-market, is really less appropriate for CMS than it is for FDA because our issue is more a coverage issue, than a marketing issue. There are four scenarios where we have interest in information: 1) pre-market, pre-coverage; 2) pre-market, post-coverage (for example, a new labeling request for a particular technology at FDA and we’re already paying for it); 3) post-market, pre-coverage (where FDA has approved the technology and CMS does not cover it); and 4) post-market, post-coverage. CMS is interested in evidence as a condition for coverage. The steps to develop Medicare reimbursement are: 1) FDA approval; 2) a health care benefit category approved by Congress (i.e. according to Social Security Act 1861, Congress determines the services that CMS covers such as hospital services, physician services, colorectal cancer screening); 3) coverage by CMS/CAG (Coverage and Analysis Group), coding by CMS/CMM (Center for Medicare Management); and 4) payment by CMS/CMM.

The National Coverage Decision (NCD) is the decision we make to pay for a technology. We can use post-market evidence to support that decision or we can, in our decision, pay for a technology but only in the context of collecting some post-marketing evidence. The Medicare national coverage process is as follows:

The process begins with a request for coverage of a technology, which may begin with preliminary discussions before FDA approval. There is a fairly extensive assessment of evidence, which may include internal assessment or an external assessment by a health technology assessment organization or an advisory committee. A draft decision memorandum is issued with a comment period. All public comments are considered. They are part of the evidence process. We are looking for kinds of evidence that will help us in this decision. This is where the FDA status can be important. If, in fact, we have not made a coverage decision and the FDA has approved a product, we request evidence. We use the same hierarchy of evidence as the FDA as given below:
• Randomized, double-blinded, controlled trial
• Randomized, unblinded clinical study
• Observational (conditions of approval) studies.
• Non-randomized registry study with formal l follow-up and data collection (single arm trial)
• Informal registry study (“open enrollment”) with less stringent follow-up anddata collection
• Sentinel reporting centers (“MedSun” contracted clinical centers)
• Medical Device Reporting (MDR) system
• Tabulation of adverse events reported to product manufacturer
• Product complaint reports sent to manufacturer
• Voluntary reports by healthcare professionals/users

We are looking for good quality evidence including post-approval studies. If there is a lack of good quality evidence, we consider all evidence. We then apply the appropriate weight to each source of evidence, depending on its scientific rigor and validity.

Three criteria are used to determine coverage. The technology: 1) improves net health outcomes using standard principles of evidencebased medicine; 2) is generalizable to Medicare population; and 3) is generalizable to the general provider community beyond clinical trial providers. This is where post market studies may be helpful. Frequently devices are approved by the FDA, but the studies are not applicable to the Medicare population. It is a much younger population or just barely our population such as a study with an average age of 63 where patients may be in the upper 60s but rarely into the mid-70s or older.

A good use of appropriately designed and collected post-approval studies is how well does this technology work as it diffuses into the general population? The generalizability to the general provider community is another good use of post- approval studies, particularly for technically difficult devices. Pre-approval studies are conducted in select centers by well trained providers. After FDA approval, the device diffuses into the general practice community and, for some, there is concern that, as has happened frequently in the past, outcomes decrease markedly. We want assurance that this is not going to occur. Post-approval information is helpful in demonstrating satisfactory diffusion into the community. If not, some standards need to be applied to those facilities before reimbursement can occur.

The final step in the Medicare National Coverage process is that the final decision is issued with instructions. The coverage decisions are national coverage, national non-coverage, national coverage with restrictions such as coverage for: a) specific populations; b) coverage for specific providers or facilities such as the recent carotid artery stent decision where a facility must certify that they meet certain standards before they are reimbursed for carotid stents because of the high risk that is involved in this particular procedure and short term data we have for this particular procedure; or c) coverage with evidence development.

The coverage with evidence development initiative is intended to enable Medicare to provide payment for items and services under conditions that help assure significant net benefits of the treatment for beneficiaries, and to also assist doctors and patients in better understanding the risks, benefits and costs of alternative diagnostic and treatment options. To state simply, CMS is going to cover a technology, but we are not quite sure that the evidence rises to the level that we would normally say yes. Typically, we would say that we were not going to cover the technology, since the evidence does not rise to a certain level. Now, we will cover it, while the data is being collected to answer questions on the technology, with the main goal of assisting patients and doctors. CMS wants good quality data however.

The coverage with evidence development has limitations. It is only used for coverage decisions. It is an expansion of current information and is rarely used (i.e. only when there is a likelihood of evidence collection is small or when the hypotheses are limited by type of evidence collection). In addition to the cover with evidence evidence development studies, there are other data collected at all times (by CMS or the industry) that are also important.

With coverage with evidence development, there is the potential of changing the coverage decision in the future. If the data supports a change in the coverage decision, we will change the decision. All of our coverage decisions are reviewable, can be reconsidered, and any evidence we collect can be used to make different coverage decisions.

CMS has set up several of registries recently where we think we can get some good quality data. That data is mainly about appropriate patients getting care and/or appropriate providers providing that care. We don’t think registries are the answer to very detailed clinical questions. We do think the field of evidence collection is an area that ISPOR needs to spend more time, as you are doing, to define how we can more simply collect good quality information, how we can work with trial design in such a manner that data collection is simpler and it’s still valid and less costly (i.e. $30 million for a clinical trial of 180 patients is unreasonable). There needs to be a better way to do this. We don’t have the answers to that. We’re asking for the answers to that and we encourage ISPOR to be participants in this endeavor.

  Issues Index | 2005 Issues Index