Its an interesting question Shanker - sorry for a late reply, I've just happened on this thread.
I think that ultimately it depends on the research question and study design. RCTs show comparative outcome between groups, so its effectiveness of one treatment - but only in direct comparison to the other. You're quite right that efficacy is a completely separate issue - hopefully one that's been addressed prior to running a RCT.
In terms of the evidence hierarchy RCTs trump most other study designs - if well conducted - which is an important point as there are plenty shoddy trials out there.
Its perfectly reasonable though to assess effectiveness of an intervention in a cohort study design - especially if there are no clear comparator treatments/therapies. Perhaps its best to do this first and then power an appropriate RCT based on the results.
It is a very interesting question. In intervention research there is often a distinguishment between explanatory and pragmatic studies. An explanatory study involves the specific effect of the presumed active ingredient of the relevant intervention. To this end, a comparison with a placebo intervention is necessary (placebo controlled design). However, a pragmatic research focuses on whether a new intervention is superior to already established treatments for that indication. Within each of the compared treatment strategies the placebo effect needs to be optimized . For a stand-alone placebo intervention is in a pragmatic study no place. In the pragmatic context of a study, it is not necessary in order to be able to separate the specific and placebo effects. Many studies in physiotherapy and manual therapy are characterized as pragmatic. It is a full-fledged research design and it is necessary to control the comparability of the effect parameter in the treatment groups by means of randomization and blinding. In physicotherapy, it is almost impossible to perform a placebo-controlled study. Therefore, the pragmatic RCT is the best one.