The problem I have with this kind of research are the intrinsic ambiguities regarding "depression" and its treatment. Depression is notably heterogeneous in it

The problem I have with this kind of research are the intrinsic ambiguities regarding "depression" and its treatment. Depression is notably heterogeneous in its characteristics and origins. Even when careful diagnosis declares two people to have the same illness, it can't be assumed their conditions are actually identical. Furthermore, we know drug treatments can affect recipients unpredictably, depending on physiological factors. Similarly, psychotherapies, which are harder to administer in a consistent way, are subject to even greater variability in delivery and effect vs. medications. The bottom line is estimating the real-world effectiveness of any therapy is at best hazardous, and the results of research (or meta-analysis of studies) are likely only tangentially applicable to individual treatment situations. From many years on the front lines, I'm convinced depression and other disorders can be effectively treated and suffering reduced. What's important to remember is treatment is as much art as science, it can't be bottled or reduced to formulas. Outcome mileage varies, but it's largely dependent on the quality of the treatment team, the power of the partnership between patient and practitioner can't be overemphasized.