According to this article, a new meta-analysis of recent research studying the effects of SSRI antidepressants (such as Prozac) on depression has concluded that unless a person is suffering from severe depression, there is no clinically significant difference between the medication and a placebo. Even when the depression was severe, the difference was not all that much. Let's sort through all of this...
First, a meta-analysis is a powerful research tool that allows researchers to analyze areas of interest by combining similar research studies in one large "super study." An important facet of research is that, the larger your sample size (i.e. number of participants), the more faith one can have in the results. What a meta-analysis does is combine smaller studies that are ostensibly measuring the same thing, and combining them in order to increase the sample size, and therefore the strength of the conclusions.
Next, what this study highlights is the power of the brain (though that was not its primary intent). Even being in this field as long as I have, I am still amazed, at times, at the concept of the placebo effect, and just how strong it can be. Really interesting stuff. In this case, simply believing one was taking a pill that should make them feel better leads to some improvement. To my mind, this also makes the case for cognitive-behavioral therapy (CBT): If you change your perception, you change your feelings and behaviors.
What's also important to take away from this is that the pharmaceuticals are selling something their products can't deliver. No, I am not trashing antidepressants; I think they can be an important part of one's overall treatment for depression. However, this idea that if somebody takes a pill, they can treat such a complex issue, an issue with biological, psychological, and social/interpersonal aspects is absurd. Now, more than ever, it should be apparent that individuals with depression ought to be participating in therapy to address the areas of their mood disorder that their medication does not impact. Again, treating depression, anxiety, etc. is a multifaceted effort - period.
A couple of other good summaries of this research can be found here and here. Bottom line - if you're on an antidepressant, don't discontinue it without working with your doctor on getting off of it. In the meantime, if you are experiencing mild or moderate depression, some sort of therapy or skill-building ought to be part of your treatment. Ditto for severe depression, though the type and scale of the therapy will likely differ at that level. If you suffer from any level of depression, or work with clients who suffer from depression, do not expect that they will be fine because of a prescription.