The big grain of salt for all depression treatments is that depression is a "wastebasket diagnosis". The only real distinction available between numerous pathol

The big grain of salt for all depression treatments is that depression is a "wastebasket diagnosis". The only real distinction available between numerous pathologies is what medication seems to help them. This means any group of patients diagnosed and not treated yet will have a grab bag of different actual pathologies. Whatever thing you try might help a few, make a few others worse, and have no effect on the rest. This is why we keep seeing reports that antidepressants "don't work". If the patients in this study were already on bupropion, that has narrowed the field and can explain the difference from others. It doesn't mean DXM is much better than others, it just means it has been better targeted. You can think of a regular RCT of an antidepressant as if it were checking whether splinting the lower legs is good for trauma victims who come in with all kinds of broken bones -- arms, legs, clavicles. In the RCT, everybody gets their lower legs splinted. Of course only the patients who broke a shinbone benefit. So if this trial pre-selects people on bupropion, it would be like splinting shins only of people with leg injuries. You still miss on the femur, knee, and ankle patients, but the odds of a hit are better. Validity of RCTs depends utterly on reliable and specific diagnosis. It is very easy for an RCT to go wrong. Treating them as automatically right is like treating any other experiment as automatically right. Calling RCTs "the Gold Standard" blinds us to its failures.