Multimodal therapies raise a number of important issues for both regulators and payers regarding combinations of regulated products, such as a drug plus a device, as well as approaches in which only one component is regulated, such as a drug or device in combination with psychotherapy.
From the perspective of the FDA, multimodal therapies represent an emerging technology area, and developers are encouraged to consult with the agency at early stages of development to help get products to patients who desperately need them as quickly as possible, according to Carlos Peña, director of the division of neurological and physical medicine devices at the Center for Devices and Radiological Health (CDRH). He said presubmissions are one of the best ways to gauge preliminary information needed from sponsors and developers for a particular product, including a combination product.
While Peña’s colleagues from other centers agreed on the need for early consultation, they also urged caution. Wilson Bryan at CBER commented that there may be a tendency to think a combination is necessary when, in fact, individual components may be just as effective on their own. Billy Dunn at CDER said the agency’s guidance states that there should be a compelling reason to codevelop two products and that they should be developed independently if this is not the case.
The FDA has specific regulations that must be followed, said Robert Califf, but that does not preclude the agency from thinking about broader principles. For example, with regard to novel outcome measures, the overarching principle is that the outcome should demonstrate a clinically meaningful impact on the disease, said Dunn. If sponsors are able to define what that effect is and demonstrate that they are able to measure it, the agency may accept the outcome measure, but may ask that it be included in the trial design along with other established measures to provide some grounding with experience.
Moreover, as the field evolves, the FDA pays attention to consensus within the scientific community, said Dunn. Patient-focused drug development meetings and workshops are also increasingly important. These meetings not only provide opportunities to talk about emerging technology issues and policy, said Peña, but they also typically involve a lot of cross-talk among staff at different centers.
Peña, Bryan, and Dunn represent the three FDA centers that may be involved in approval of multimodal therapies. While the centers want to cooperate, each has different user fees, different evidentiary standards for different components in the same product, different manufacturing standards, and different cultures. According to Califf, the agency is making an effort to modernize and adapt a system in order to increase efficiency, consistency, and predictability in review; encourage innovation; and support development of novel technologies through a lean management approach, improved information technology systems, and harmonization at the leadership level.1 For example, Dunn said, there is a workgroup across all centers involved in neurology products. The workgroup discusses crosscutting issues and tries to develop consistency in requirements and messaging.
Office of Combination Products
The Office of Combination Products (OCP) was established in 2002 as a specific entity within the FDA to serve as a focal point for combination products, which are defined as two or more differently regulated types of medical products, according to OCP deputy director Patricia Love. These different components may be a drug and device, drug and biological, or device and biological; these components may be physically or chemically combined in a single entity (e.g., drug-eluting stent or prefilled syringes), co-packaged as a kit (e.g., drug plus empty syringe), or sold separately and labeled for use together. Multimodal therapies may or may not be combination products, depending on whether they fit into this definition.
Typically, one investigational application is needed for a combination product, and according to Love, combination products are assigned to the center based on the primary mode of action by which the product achieves its intended purpose. For example, a product consisting of a drug and delivery device, in which the primary mode of action resides with the drug, would be assigned to CDER as the lead center, said Love.
1 For additional information see http://blogs.fda.gov/fdavoice/index.php/2016/04/developing-a-consensus-voice-the-combination-products-policy-council (accessed August 31, 2016); and http://blogs.fda.gov/fdavoice/index.php/2016/03/leaning-in-on-combination-products (accessed August 31, 2016).
When the primary mode of action is not clear, the product is typically assigned to the center with the most expertise in dealing with significant safety and effectiveness questions. Love said her office works hard to make sure that all the issues are addressed consistently and in a transparent manner, and that all of the relevant centers are at the table to address issues that arise. Regardless of whether a product qualifies as a combination product, when a treatment involves two modalities that are regulated in different ways, the scientific questions are the same, she said.
Drugs, Biologics, and Devices: CDER, CBER, and CDRH
While small molecule drugs and biological products, may, depending on the product, be processed by two separate centers, CDER and CBER, the different challenges are generally scientific not regulatory, said Dunn. The regulatory pathways and standards of evidence are the same, added Bryan. Moreover, some therapeutic biological products are regulated by CDER, said Dunn.
Drug−drug combinations are not considered a combination product, but are referred to as fixed-combination prescription drugs for humans and have a separate regulation. Each component must make a contribution to the overall effect, and there should be a compelling reason to use a combination approach, said Dunn. He added that if all the criteria for a combination product are met, there remain implications for study design. For example, it is not usually sufficient to, in isolation, study the combination versus placebo in most circumstances or the combination versus only one of the components; a factorial design is often needed, if feasible, unless the contribution of each component is otherwise established.
Bryan went on to say that if the combination involves two approved products, either two drugs and/or biologics, it could be exempt from Investigational New Drug regulations; however, if the two products are investigational, the development program will have to demonstrate that each component contributes to the safety and/or effectiveness of the combination. The amount and the types of data needed and the study design will depend on the particular nature of the disease and the particular combination being tested, he said.
When OCP determines that the primary mode of action of a multimodal product derives from a device, it is assigned to CDRH, although as Peña noted, the centers engage in substantial cross-talk when a submission contains multiple components that cross the lines of responsibility of the respective centers. For combinations that include a drug or
biologic plus an unregulated intervention, such as a psychosocial intervention, the usual standards of evidence apply to the drug or biologic component, said Bryan: substantial evidence of effectiveness as well as evidence that the drug is safe. However, different sections of the label may be influenced, in particular the “indications and usage” statement and the “clinical studies section.” Whether the psychosocial intervention is specified in the indication statement may depend on the evidence of safety and effectiveness of the drug in the absence of the psychosocial intervention, as well as the overall benefit–risk assessment for the drug in the absence of the other intervention. Ultimately it will depend on the data provided. If data are unavailable or insufficient regarding effectiveness of the drug in the absence of psychosocial intervention, it will depend on the FDA’s level of concern about having these data before putting the drug on the market, said Bryan.
When a therapy is approved by the FDA, many government plans automatically allow that drug or that device to be covered under Medicare as well as under individual group plans and Medicaid, said Robinson-Beale of Blue Cross Idaho. However, the largest insurers or payers are not these governmental programs; they are private health plans, which do not automatically approve coverage and which scrutinize the evidence differently. The major dilemma for payers is that health costs are intolerably high, said Robinson-Beale, and multimodal therapies may contribute to escalating costs. She added that many other factors contribute to the rising cost of health care, including practice variation, uninformed preferences, and supply and demand of new technologies. As a result, between 2000 and 2013, overall health insurance costs increased 66 percent, while take-home pay increased by only about 3 percent. The dramatically increased cost of care also reflects the fact that whereas 20 percent of the population used to drive 80 percent of costs, today about 6 percent of the population drives 80 percent of costs. The reason is that treatments have become exceptionally complex and costly, not only in terms of specialty drugs, but combinations of specialty drugs and gene therapy. Another factor contributing to high costs, said Robinson-Beale, is excessive waste. Indeed, a report from the Institute of Medicine suggested that one-third of health care spending may be excessive and unnecessary (IOM, 2013). High health care costs are also driven
by the preponderance of off-label prescribing, particularly for behavioral health indications, said Robinson-Beale.
These costs raise ethical issues as well, said Robinson-Beale, in that the high cost of medical care is forcing difficult decisions to the forefront for payers, providers, and patients. This has become particularly challenging in oncology. In neurology, the decisions differ in part because the risk of mortality may be lower than in oncology, but the risk of lifetime disability may be very high. For example, the cost of treating depression with TMS may be about $6,000 per month compared to about $400 per month with a typical antidepressant, and combining the two may cost even more. Payers therefore have to constantly readjust policies to keep up with latest research and evidence available. Nonetheless, Sarah H. Lisanby of NIMH said that in practice, many clinicians are now delivering psychotherapy during the delivery of TMS because the patient is there every day with the clinician. Although this may be more resource efficient, whether payers will approve reimbursement will depend largely on having reliable, measurable endpoints with good enough follow-up to ensure that the investment is paying off, said Steven Pizer, associate professor of health economics at Northeastern University and chief economist, health care financing and economics at the U.S. Department of Veterans Affairs (VA).
Robinson-Beale said payers must make coverage determinations regarding what is medically necessary according to the generally accepted standards of medical practice. The information they consider must include data from large randomized clinical trials, large observation studies, and practice guidelines from subspecialty societies, particularly those based on formal consensus methodology. She noted that the clinical application of a treatment should be clearly outlined in terms of identifying the appropriate indications (e.g., illness, injury, or disease) as well as the frequency, extent, site, and duration of treatment. In addition, for practice guidelines, evidence should be sufficient that a clinician not familiar with a particular treatment has a very clear outline of dosage, what to expect, how to measure the treatment effect, and how to know when the treatment is failing, said Robinson-Beale.
Also playing an increasingly important role for reimbursement decisions are comparative effectiveness studies, in which a treatment is compared not only to placebo, but also to what is called the standard of care. Finally, payers want a demonstration that treatment effects are reproducible with fidelity, that they preserve life and functionality, and that they do not cost more than an alternative intervention, Robinson-Beale added.
Robinson-Beale described the evidence that will be reviewed in evaluating new technologies, including the number of people tested, the outcomes observed, gradations of response, and the demographic details of the responsive population. The transition from phenomenologically-based diagnostic categories such as those embodied in the Diagnostic and Statistical Manual, 5th Edition, to neurobiologically-based dimensions promoted by the RDoC initiative, may further complicate these assessments, she said, due to lack of standardization in how research studies are conducted and reported. For multimodal therapies, she added that clear indications for use are needed, including whether a step therapy protocol should be used, when treatment starts with one drug before initiating treatment with the device or vice versa. If this approach is used, evidence should also demonstrate that the combination of both is much better than either alone.
Generating evidence of comparative effectiveness is currently part of the European Medicines Agency (EMA), but not the FDA framework, noted Karl Kieburtz. Dunn commented that comparative effectiveness data can be challenging to interpret. However, Robinson-Beale said U.S. payers will consider European data in their decision-making process. The problem, she said, is that devices and drugs may not be comparable to what is available in the United States. Pizer noted that another difference between health care systems in the United States and national health systems in Europe is that rather than remaining within a single national system, Americans typically change payers every few years as they change jobs. Insurers in the United States thus have less incentive to invest in something that might be costly in the short term, but will eventually result in cost savings. This is changing in the United States as federal and state governments are paying a larger share of the total health care budget, said Pizer.
Public payers manage costs differently than private payers do, said Pizer, who is also the chief economist for PEPReC (Partnered Evidence-Based Policy Resource Center), a new VA resource center made up primarily of economists and health services researchers to forecast the demand for care through timely, rigorous data analysis, and improve connections between providers and researchers to get closer to the ideal of a learning health care system. According to Pizer, the VA is in an unusually good position to learn more about drugs and devices because it comprises a large community of researchers as well as care providers. Pharmacy services are centrally managed, with an evidence-based national formulary that can negotiate good prices for drugs. In comparison,
decisions about devices are made in a decentralized way, resulting in reduced bargaining power. With regard to multimodal therapies, Pizer said the incremental value from the combination will be of foremost importance, which means that studies will have to be large enough to detect an incremental benefit.
In both Europe and the United States, there is a push to generate an evidence base that will serve three purposes—labeling, practice guidelines, and reimbursement—despite the fact that the criteria by which decisions will be made in those three areas are very different, said Califf. The label is the most fundamental source of information about a medical product, he said, adding that the label should include evidence about the population in which it is going to be used, including in multimodal therapies. Moreover, the label and practice guidelines are complimentary pieces of information, although the label is constrained by a high level of evidence that is required, while the practice guidelines can reflect nuance, filling in the gaps between knowledge and professional views of how to put the information together, said Califf. Jeffrey Nye, vice president of neuroscience innovation and scientific partnership strategy at Janssen Research and Development, noted that a license from the FDA does not assure widespread use in practices, and that other barriers, such as restrictive reimbursement, may limit utilization.
According to Peña, some efforts have been made to bring the FDA and the Centers for Medicare & Medicaid Services (CMS) together to try to design studies that can address the needs of both agencies. Martha Morrell noted that from the clinician’s standpoint, postmarket and open-label data may be the most compelling and most representative of the naturalistic environment. Other potential sources of data include electronic medical records and patient registries, noted Robinson-Beale. She even suggested that a new governmental organization might be needed to look at the both cost efficiency and effectiveness of treatments using these and other sources of data.