I recently read this interesting article about the exaggeration of evidence supporting the use of anti-depressants to treat anxiety. The article is based on recent research out of Oregon State University that identified two primary ways in which such evidence was being exaggerated. The first is called publication bias and basically means that positive results got more publicity than negative results. In fact, studies of drugs that the FDA determined to have positive affects were five times more likely to be published in peer-reviewed journals than studies which found no impact or negative impacts. The second manner of exaggeration is called outcome reporting bias and basically consists of researchers putting a positive "spin" on neutral data. If a research team spends 3 years and $100,000 of grant money testing a drug to treat anxiety, there is a tendency to find a useful result even if it isn't supported by the data.
Those biases have resulted in an explosion of the prescription of second generation antidepressants by primary care physicians who read a few positive articles but, due to bias, do not necessarily understand the whole picture. Certainly, these anti-depressants can help many people with depression, and perhaps some people with anxiety and panic attacks, but as documented in this recent detailed recent study at Oregon State, that benefit has been exaggerated for over 20 years.