Note: Regulatory approval by the FDA is considered a critical goal by manufacturers and their investors, which is understandable, since there can be no commercialization without it. However, adequate payer reimbursement is clearly the final goal. Without it, commercialization stalls. The following blog is the second in a series of posts that examines the relationship between reimbursement and the type of regulatory approval for medical devices. Last month’s blog focused on the 510(k) process. The following discussion focuses on the premarket approval (PMA) process.

FDA’s PMA approval process is required for those medical devices “that support or sustain human life, are of substantial importance in preventing impairment of human health, or which present a potential, unreasonable risk of illness or injury.” Given the associated risk of these devices, the PMA approval process typically requires extensive clinical data, often in the form of a randomized trial. A PMA is considered the most rigorous of the various FDA regulatory pathways.

Every year, FDA reviews an estimated 3,000 to 4,000 devices through the 510(k) process. In contrast, only about 30 to 40 devices receive FDA approval annually through the PMA process.

As noted in last month’s blog, payers are often skeptical about 510(k) cleared devices, particularly when accompanied by minimal clinical data. However, this does not mean that approval through the PMA process will immediately result in a positive payer policy. This is a particularly frustrating situation when the manufacturer has conducted a high-level, well-designed randomized controlled (even placebo-controlled) trial that meets both its primary and secondary outcomes. How can this be?

It is important to understand that the payers can set their evidence requirement wherever they want, and payers may be interested in different outcomes than FDA, which is focused more on safety than effectiveness. If the device is an implantable device, for example, the payers may state that long term outcomes – up to seven years in some cases – would be required to validate safety. In other situations, FDA might have accepted an intermediate outcome, while payers may require final health outcomes. For example, for diagnostic technologies, the FDA may have accepted clinical validity data demonstrating that the diagnostic data is reliably related to an underlying disease. In contrast, payers may go one step further and demand clinical utility data, i.e., data demonstrating that the information has been actively used to direct patient management, with resulting improvements in final health outcomes. Here, payers have attached a therapeutic outcome to the diagnostic test, and this therapeutic outcome may not become apparent until years and multiple interventions after the original diagnostic test.

Case study: Artificial intevertebral discs

The experience with artificial intervertebral discs is an instructive case study. Artificial lumbar discs were the first to receive FDA approval in 2004 on the basis of randomized non-inferiority trials comparing the implantable discs with standard fusion. The key outcomes included pain relief, restoration of function and absence of adverse effects. While the randomized trials met their primary endpoint, payers resoundingly rejected FDA-approved non-inferiority trial design. Specifically, payers, particularly the Blue Cross Blue Shield Association, felt that a non-inferiority trial was inappropriate since there was still controversy regarding the outcomes of standard fusion.

This was a somewhat ironic stance since standard fusion, supposedly controversial, was routinely covered at the time. Payers were essentially imposing a higher evidence standard on the new technology, while standard fusion was “grandfathered” in. The result was an evidence requirement of a superiority trial with long term outcomes. From the manufacturer point of view, the size of a trial powered to demonstrate superiority would not have been feasible, and patients would certainly be lost to follow-up over the course of years, another limitation that payers would be quick to point out. While some payers have finally started to cover artificial lumbar discs on the basis of five to seven-year follow-up, for many years this implantable technology was “dead in the water.”

Why did payers impose such a high evidence requirement given FDA approval through the PMA process? The answer may relate to the overall reimbursement environment for spine technologies based on the increasing demand of fusion procedures in general. In addition, payers are very wary of implantable devices, concerned about their up-front cost, durability, and need for revision surgery. A payer nightmare would consist of a high percentage of implantable artificial discs that required further surgery or intervention following their failure. The long term outcomes of an implantable device are one reason why the evidence requirement is typically higher for devices compared to drugs, also approved on the basis of randomized trials through the FDA’s drug approval process. Drugs can simply be discontinued if complications or ineffectiveness become apparent. While the long term costs of drugs can certainly exceed the cost of a device, these costs are spread out gradually over years, compared to the up-front cost of a device.

Positive payer policies following a PMA approval are certainly a cause for celebration. However this may soon turn to dismay when the policy duplicates the exact patient selection criteria used in the pivotal trial. These patient selection criteria were created from a research point of view in order to create homogenous populations in a randomized trial, and manufacturers may have anticipated that broader populations would be candidates once the technology was commercialized. After all, the FDA-labeled criteria are typically broader than the specifics of the pivotal trial. However, presumably in an effort to control utilization, payers typically rely on the clinical trial data as the basis of their coverage policies.

What can manufacturers do to address the payers’ stiff evidence requirements? Unfortunately, the reality is that often data cannot speak for itself. Even a high-quality randomized study may need the additional support of providers, key opinion leaders and most importantly specialty societies. Support may come in the form of a preliminary methods paper published in the peer-reviewed literature explaining the key considerations and clinical validity of the trial design. Once the technology is commercialized, support can come in the form of a consistent pattern of overturned appeals, key opinion leaders willing to interact with the payers in their market, and support from specialty society guidelines.