Here is what the Cochrane Collaboration actually said from their own systematic review in 2005:

Authors' Conclusions

Implications for practise
The results of this review have important practical consequences. There is evidence to support drug therapy for this condition, which is associated with disability, comorbidity and high use of health services. Dysthmic patients can improve on antidepressants ...Doctors do not need to wait two years to consider pharmacological treatment for their patients

With regards to Peter Gotzsche: his position on anti depressants should not be taken seriously, it is so extreme it is hard to find any support in the mainstream medical community. The Lancet and the BMJ were just some of those that spoke out against him.
http://www.behaviorismandmentalhealt...6614702329.pdf

He actually wants to implement a policy in which large scale withdrawal clinics are set up and the vast bulk of patients taken off their psychotropic medications. He wrote a book which was reviewed here in the link below. Take special note of where the reviewer found an entire chapter unreadable

Book Review: ?Deadly Medicines and Organised Crime: How Big Pharma Has Corrupted Healthcare? by Peter Gøtzsche - Speaking of Medicine

Really you need to come up with some better sources. All you do is reference mad in America. Whitaker was wrong on almost every page of his book. I mean to get even BASIC things wrong is particularly disturbing

Anatomy of a Non-Epidemic - a Review by Dr. Torrey - Treatment Advocacy Center

Gotzsche relies on the Fournier studies but the GIbbons studies trumped them with longitudinal measurement. Usually, the Kirsch studies are brought up however they can be dismissed as well as researchers went back and looked at his work and could not even replicate his results. How is that for publication bias? Dear Vital you seem to think that only drug companies use publication bias, well that is not the case. In fact publication bias is seen mostly in CAM. For example, almost every study showing a positive outcome for acupuncture was affected by publication bias. There are thousands of them. What makes you think the Kirsch and Fournier studies are not biased?