[ajarmst]

Not necessarily. We do not need to know the mechanism of an illness to test the efficacy of a treatment. In fact, identifying an efficacious treatment is frequently a key step in developing a model of the underlying illness. There are a number of treatments for depression that have shown promise across multiple studies (and most studies control for placebo effect, and good studies control for active placebo where possible). SSRIs and SNRIs have shown promise for short-term alleviation of symptoms depression, with some serious concerns about long-term use. Cognitive therapy has proven to be quite effective. Some studies of mindfulness (although there are concerns about study quality) indicate great promise there as well.

One thing that is starting to become clear is that effective long-term treatment is likely to require more than one therapy, tailored to the specific situation and needs of the individual patient. Any study, like this one, that implies that depression has a simple mechanism and an obvious cure is about the only one that you can be fairly sure has no validity.