Duke 2008 | ASHE | Equity and Efficiency in Health and Healthcare

Over the past half century, new medical devices have extended human life, reduced pain, risk, and disability and often, though not always, proven to be costly. This recognition has led to a growth of interest in evaluating the health and economic impact of medical devices. Increasingly, the U.S. FDA requires randomized trials of these devices for marketing approval. Payers across the world rely on cost-effectiveness analyses in making their coverage decisions, and there is growing movement to adopt these approaches among public and private payers in the U.S. In this paper, we argue that these analytic approaches, which have been shaped heavily by their use in the pharmaceutical realm, are ill-adapted to the medical device context. A first set of challenges are in the pre-marketing setting. In the case of novel devices, randomized clinical trials--the "gold standard" for evaluation of safety and efficacy of new drugs--can be more difficult to design and conduct. As previously argued, if trials compare a device to medical therapy, blinding is more difficult to achieve. Moreover, as vastly different treatment approaches may be associated with strong physician and patient preferences, it may be harder to achieve equipoise for randomization. Here we go further, providing evidence that because devices generally are embedded in clinical procedures, trials have to contend with a much more prominent learning curve phenomenon among users, as well as greater variations in provider skills, during the trial period. A final challenge is that the target populations for devices are typically much smaller than for pharmaceuticals, which argues for the need to reduce sample size. Even if these challenges are addressed, and well-controlled trial evidence becomes available, policy makers still have to decide make decisions in which they must balance the trade-offs between benefits, risks, and costs of new devices in a context of uncertainty. This uncertainty is, partly, shaped by the fact that pre-marketing trials are based on a sampling process, may have limited time frames, limited patient heterogeneity, and are conducted in specialized centers, which raises questions about their generalizability to the post-marketing setting. But there is also a more vexing source of uncertainty: in devices, there is much more post-marketing innovation and “learning by using” than with pharmaceuticals. This learning can both reduce costs and increase clinical effectiveness. With considerable learning and incremental innovation as technologies diffuse, useful evidence on both clinical effectiveness and economic costs of new devices is generally difficult to obtain until they are widely used. That is, learning and innovation require widespread adoption. This poses fundamental challenges to the evaluation enterprise as well as the policy making world, which are often neglected. For example, our literature review suggests that most cost-effectiveness analyses of devices fail to account for future technological change and learning. When such learning is unaccounted for, the use of strict cost-effectiveness thresholds in coverage decisions can eliminate valuable technologies before they have had a chance to become cost-effective. Using a wide range of implantable devices as case studies, we conclude with a discussion of methodological and practical opportunities for improving evidence, and a discussion of the most fruitful paths forward for policy making.