Gaps, Tensions, and Conflicts in the FDA Approval Process: Implications for Clinical Practice

Richard A. Deyo, MD, MPH


J Am Board Fam Med. 2004;17(2) 

In This Article

Common Misconceptions

Many consumers and physicians may have misconceptions about the FDA approval process. For a new drug to win approval, the FDA does not require it to be better than products already available—only that it be effective (better than nothing) and fairly safe. For high-risk devices, demonstration of safety and efficacy are also required. But for moderate risk devices, only safety and "substantial equivalence" to a previously marketed device are required. The benefit of a new drug or device must be judged to outweigh the risks. This is all Congress has allowed the FDA to require.

In some cases, the definition of "effective" is narrow and may not address the end results of therapy. A drug that achieves a "surrogate outcome" may be approved if it lowers cholesterol, lowers high blood pressure, or improves heart rhythm—without knowing if it improves life expectancy.

In some cases, approved drugs were later found to increase rather than decrease mortality. The antiarrhythmic drugs encainide and flecainide were examples. The company-sponsored trials that led to their approval showed they were effective in suppressing ventricular arrhythmias. However, by one estimate, these drugs produced a death toll of 50,000 before their toxicity was demonstrated in a large National Institutes of Health-sponsored clinical trial, in which the mortality with active treatment was twice that with placebo.[8,9] Furthermore, studies used for drug approval are not designed with sufficient statistical power for detecting important but infrequent safety problems.

Drug approval generally requires rigorous testing of clinical efficacy, in the form of at least 2 randomized controlled trials. However, the regulations for medical devices are quite different. Medical devices include anything from contact lenses to cardiac pacemakers and MRI scanners. Most new devices are approved by demonstrating "substantial equivalence" to a product that was marketed more than 25 years ago (before 1976). For this type of approval, a device need only do technically what it claims and be reasonably safe. A device that delivers electric current to the skin can be considered "effective" without asking if it relieves symptoms. Devices that do not claim substantial equivalence to an older device (a tiny fraction of new submissions) are required to undergo more rigorous review. This may or may not require randomized trials.