A recent analysis showed that effects of anti-depressants like cipramil are clinically irrelevant. When their effects are studied, you find a small effect that is 'statistically significant; but it is that small, that it is below the size that clinicians consider relevant. In adolescents there are conflicting observations that these pills may increase suicide.
If there is any use for anti-depressants, it is in very severe, clinical depressions; only because in these patients the effects of placebo's are less.
So, the average person should not stay on anti-depressants at all. Psychotherapy is a more meaningful approach. I favor existential, experiential and interpersonal appoaches.
http://medicine.plosjournals.org/archive/1549-1676/5/2/pdf/10.1371_journal.pmed.0050045-S.pdf
Unfortunately, your personal overview does not stand up against the reference that I included: in most depressed patients the effects of anti-depressant are statistically significant but clinically irrelevant. As a doctor we learn to treat patients, not numbers. Treatment should be clinically relevant, not statistically. These drugs, most often, are not.
Your overview unfortunately carries many papers with several biases, one of them being that studies with positive outcomes are more likely to be published than negative ones. Farmaceutically paid studies contribute to this.
#2 - While I do agree that drug therapy should be the last option it does help people quickly unlike your method of treatment. Getting a patient out of the woods quickly is sometimes very important and could be life saving. You said as a doctor you treat the patient not numbers, I don't remember mentioning numbers at all. I am only talking about getting people better quickly and safely. I also think my opinion does stand up. Your 2 references out of maybe 10's of thousands is not something I woul