Paul Offit seems to be irrationally biased. In his longer guardian piece on the same topic he is extremely selective in his referencing of the scientific literature. This is his summary of the findings on one popular herbal treatment for depression:

"Depression is a serious illness; to treat it, scientists have developed medicines that alter brain chemicals such as serotonin. [...] Alternative medicine practitioners, however, have a better idea – a more natural, safer way to treat depression: St John's wort. Between November 1998 and January 2000, 11 academic medical centres randomly assigned 200 outpatients to receive St John's wort or a placebo, finding no difference in any measure of depression."

He omits the fact the same study also did NOT detect any effect of Sertraline/Zoloft (one of the very medicines he refers to so approvingly):

"Neither hypericum nor sertraline could be differentiated from placebo on the primary efficacy measures."

This points to a lack of statistical power in the study (which the authors of the study acknowledge).

He also omits to mention the many studies showing st john's wort to be as effective as SSRIs, with fewer side effects (eg see the cochrane summary here http://summaries.cochrane.org/CD000448/st.-johns-wort-for-tr...).

I don't trust this author.