November 7th, 2012
Better Evidence, More Comparators Needed in Device Research
Sanjiv Shah, MD
Medical devices play an important role in cardiac care and have helped many of our patients. As many cardiac devices are considered high-risk and are often implanted, it is essential to establish that they are safe and effective and better than the alternative before physicians can recommend their use and understand the best role for new devices on our patients. In a research letter published in JAMA, my colleagues and I looked at the last twelve years of cardiac device premarket approvals. We found that less than half (58 out of the 121) of devices were approved based on at least one study involving an active control, and 35% (42 of 121) were approved based on a single arm study without a control group or based on objective performance criteria or goals, which also don’t have any control group. Another 17% used a control group from a different study, known as historical controls. These findings mean that the majority of cardiac devices are approved without data showing that they are better than an alternative treatment.
These findings highlight an important opportunity to improve the evidence required to support high-risk medical devices. Clearly, we want to approve devices expeditiously when they can improve patient outcomes. However, we do not want to approve and start using devices that have not been shown to be better than already available devices or current medical therapy and in fact may be harmful. Comparative effectiveness studies are essential to allow us to be confident that the benefits outweigh the harms for new devices. This information is critically important as many devices are permanently implanted and removal can only be done at great risk to our patients. The quality of the evidence needs to keep pace with the technologic advances in and proliferation of new devices.
In this day and age of randomized, controlled trials, not good science–The medication approval process is obviously far more rigorous and stringent.