That was exactly my first thought. We could definitely throw the SSRI efficacy into doubt though by sampling CS fluid or something for actual serotonin content in cohorts who are subjectively feeling improved by SSRIs.
And, even if they are better than just an active placebo, the mechanism is probably not the effects on serotonin levels. (Discussed previously here: https://news.ycombinator.com/item?id=32160703)