One question often asked, in reference to whether antidepressants are effective, is whether suicide rates have declined in response to widespread antidepressant use. Many authors have attempted to answer this question. One study attempted to demonstrate a time correlation between antidepressant use and declining suicides. Why did these authors cut off the data for the 1950s, and begin their analysis with data for the 1960s? Including the suicide rate data (very low, especially among women) for the 1950s would reduce confidence in the authors' conclusion that antidepressant use is time-correlated with a drop in suicides. (The 1950 data occurred long before SSRIs were available, but undercut the authors' claim that the 1960-1988 data represented the baseline condition from which suicide rates dropped.)
More importantly, given the data that antidepressant use is not associated with a drop in suicidal behavior, such as suicidal ideation and suicide attempts, even if time correlation were properly demonstrated, the effectiveness of antidepressants is not supported. It is difficult to imagine any way in which antidepressant medications could reduce suicides while having no effect on suicidal ideation, behavior, and attempts, especially given that the JAMA study found that "cry for help" insincere suicide gestures decreased slightly between the 1990-1992 and 2001-2003 studies.
Actually, there is a way that antidepressants could reduce completed suicides but not suicidal behavior or ideation: if they interfere with cognitive ability in general, rendering suicides clumsier or more poorly planned.
But the more likely explanation, given all the data, including the February study that the commonly prescribed antidepressants don't work better than placebo, is that some factor other than antidepressant use is responsible for any drop in suicides. The obvious answer is that, to the extent that there has been a drop in suicide rates, it's due to the drop in gun ownership over the past few decades. (Here's a graphic.) As I have previously noted, gun ownership is highly correlated with suicide. Reduction in gun ownership, in addition to reduction in availability of other reliably lethal methods of suicide, such as lethal pesticides (frequently used for suicide by men and women in poor countries) and barbiturate sleeping pills, is much more likely to be responsible for the drop in suicides than increased antidepressant use. These factors - what might be termed coercive suicide prevention methods - would easily explain a drop in suicide not associated with a drop in suicidal ideation or behavior.
Thanks to Overcoming Bias poster Peter McCluskey for the pointer to the PLoS article.
Edit: Cognitive behavioral therapy (CBT), much hyped for depression, apparently doesn't work either. From the study:
Stravynski and Greenberg suggested that all models of psychotherapy, including cognitive behavior therapy, may be "equally unsound scientifically but they energize the therapists and provide useful fictions to activate the patients to lead somewhat more satisfactory lives." [Citations omitted.]