New Review Brings Antidepressant Effectiveness into Question

A New England Journal of Medicine study reviewed numerous antidepressant trials up to 2004, in an attempt to determine how well the data support their efficacy. The important difference in this study was that the reviews covered data submitted to the FDA, not data published in journal articles.

It may surprise some to learn that while the FDA requires all data to be submitted to it when reviewing a medication, there is no requirement that all data get published in a medical journal. Anywhere from 30-50% of all clinical trials never get published.

You may not be surprised to learn that, generally, the studies that get published reflect favorably on the study medication, while those that do not get published did not show robust efficacy.

Unfortunately, the media has taken these results and exaggerated them. Contrary to the common reporting, the study did not find that antidepressants were ineffective or no better than placebo. Even when the unpublished, negative studies were taken into account, the meds were still found to be better than placebo, but their overall efficacy was 32% less than the result of looking at published studies alone.

What the study did find, however, was that negative studies were selectively unpublished. There are a number of possible explanations for this, the most obvious being that drug companies suppressed them.

But an important question is why the FDA-which receives all data, published or unpublished-simply does not release this information itself. It seems fairly obvious that if the FDA needed the data to determine whether a drug should win approval, doctors and patients should have access to the data for their own safety and treatment.

One reason the FDA does not publish the data itself is because it may be illegal. Pharmaceutical companies have used Exemption 4 of the Freedom of Information Act, which prevents "trade secrets" and other similar information from being released. (For example, Coca Cola is obligated to print the ingredients on its bottles, but it cannot be forced to publish its secret formula, nor can anyone demand it from the FDA.)

Recently, the government has mandated that the data get published. A website, clinicaltrials.gov, is a registry of all clinical trials. While generally useful, it still remains the case that a number of clinical trials never make it into the database.

So what should patients do? First, do nothing without the supervision of your doctor. Second, understand that even if the NEJM had learned that studies had shown that antidepressants were not effective, it does not mean that your experiences on a medication are any less real or valid. Studies are used because they provide the easiest way we can quantify effects on populations and make generalizations about efficacy and safety. However, studies are necessarily artificial; they measure changes that may not reflect your symptoms. (For example: the Hamilton Rating Scale, the most commonly used clinical trial scale, measures insomnia in depression, but not hypersomnia; so if your depression makes you sleep 15 hours a day, and Prozac reduces this, not only will the Hamilton not detect your depression, but it will not detect Prozac's efficacy, either.)