Q&A with Bennett Levitan
An expert on benefit-risk assessment explains the importance of patient preferences in the development and regulatory review of drugs and medical devices.
Bennett Levitan, M.D., Ph.D. is a member of MDIC’s Patient Centered Benefit-Risk Assessment Steering Committee. He is Director, Epidemiology at Janssen Research and Development, a Johnson & Johnson company. A specialist in drug benefit-risk assessment, he has over 20 years of experience in decision analysis and modeling and simulation, both as a consultant and for pharmaceutical companies.
MDIC: What are benefit-risk assessments, and why do they matter?
Bennett: Drugs and medical devices obviously provide benefits, but they also can cause harms: side effects. Benefit-risk assessment is the formal evaluation of a drug or device, considering both the benefits and harms.
It’s important because, ultimately, benefit-risk assessment is the basis of a health authority decision about whether to allow a drug or device on the market. Once regulators trust the data from a clinical trial, they’re going to ask, “Compared to existing treatments, for an appropriate group of patients, does this new treatment help more than it hurts?”
MDIC: When people in the medical device community talk about measuring patient preferences, what do they mean?
Bennett: When researchers study patient preferences, they are measuring—either qualitatively or quantitatively—the relative desirability or acceptability of the features of different drugs or medical devices. In other words, how much do patients like or dislike those features?
For example, some people might be comfortable with a shot. Others might prefer a treatment to be implanted. Other people might prefer an oral medication. Some people might not mind a treatment lasting for years. Others might say, “I want this over as quickly as possible.”
And preference studies don’t just tell us whether patients like or dislike something; they tells us the degree to which they like and dislike it, so we can start assessing tradeoffs. And that’s where it begins to be important. For example, consider pacemakers. Implanting a pacemaker is a surgical procedure that carries a risk of infection or death. But people choose to have pacemakers implanted quite often. Why do they do that? Because they see the benefits of pacemakers in reducing cardiac rhythm problems as greater than the risks of the implantation.
It’s also important to understand patient preferences at a population level. Patients do not have the same preferences: there’s a distribution. And just as, in some clinical trials, you learn that benefit outweighs harm only for a subgroup of patients—just men, for example, or people over a certain age—the FDA and MDIC are exploring whether, for some treatments, benefit outweighs harm and a device should be approved for a subset of patients who share a certain set of preferences.
MDIC: What’s an example of a situation when that might be true?
Bennett: The FDA’s Center for Devices and Radiological Health (CDRH) did a study that illustrates the usefulness of patient preference data. The study included several hundred patients—all obese—who were asked to compare various hypothetical treatments for obesity. They were told that the treatments would result in a given amount of weight loss and potentially a reduction in their use of diabetes or blood pressure medications, but would also potentially involve various side effects that may require going to the hospital, and could potentially result in death from the procedure. The study measured how preferences varied based on the levels and types of risks and benefits. How much risk were patients willing to accept in order to lose a certain amount of weight, and keep it off for a certain amount of time? Some patients might say, “I am happy to accept a low chance of a side effect, because having that weight gone—and having the other health consequences of obesity gone—is that important to me.” Other patients might say, “I’d love to lose this weight, but I don’t think it’s worth the risk. So I will not have the treatment; I will not have gastric banding; I will continue to struggle with a diet.”
MDIC: How much are patient preferences currently measured in the realm of medical devices?
Bennett: Benefit-risk assessments are still primarily value judgments made by physicians or regulators, but that’s beginning to change. As time goes on, I expect to see greater movement toward documenting patient preferences with formal methods, and seeing a role for these methods in medical device regulatory submissions.
MDIC: How do you measure patient preferences, and what are the challenges of doing so?
Bennett: There are many ways to assess preferences: Simple structured surveys; decision conferencing methods with expert facilitators, which can be done in a few hours; full-fledged studies using techniques such as conjoint analysis and best-worst scaling, which can take several months.
Measuring patient preferences is challenging in a number of ways. The large variety of methods leads to confusion, because there’s limited guidance on which ones would be appropriate in a regulatory submission, and they vary considerably in terms of how much time they take, as well as cost, rigor, and transparency. The question “Is a preference study worth it?” can also be challenging. Without regulatory guidance, and with no clear role for a preference study in an application to a health authority, it’s often hard for companies to justify doing one.
Yet another challenge is that we don’t always know if preference studies are needed. Many times, benefit-risk assessments can be handled by clinical judgment. Death is worse than minor bleeding: you don’t need a formal study for that. Where it could get trickier is, suppose you’re comparing a heart attack to a disabling stroke. Which is worse? I happen to think the disabling stroke is much worse, but some patients may not.
MDIC: How is MDIC contributing to the medical device community’s efforts to better understand patient preferences?
Bennett: The fact that MDIC is researching this issue, and that CDRH has put so much attention and funding into it, has raised awareness of the importance of patient preferences and their potential role in regulatory review. The reports that MDIC will produce—a catalog of methods for measuring data on patient preferences, and a framework for using that data in the regulatory review process—are going to be valuable resources. Most of the methods and principles will apply to drugs and biological products too. And MDIC’s framework is intended to serve as an input to guidance that CDRH may consider issuing. If CDRH does put out guidance that addresses the role of patient preferences in device development and review, that will address many of the challenges that we were discussing earlier: Which method should you use? When should you use it? Is it worth the investment? Is there a role for a preference study in a regulatory application? MDIC’s efforts will provide direction that helps both sponsors and health authorities.