The problem with any studies on biases or placebo is who to use and their own experiences and methodology.
You will find some patients that have chronic long term pain who use methodology for pain management and identifying which medications to assist, will probably have different results to the standard placebo test patients, this is further compounded by how long said placebo effect lasts both in terms of between "doses" and how long placebo treatment can be used to trick the person.
Regarding imitating heroin, it is worth noting that some studies have shown addicts do have a rush-joy before actually injecting, replicating in several ways the behaviour of the actual heroin itself.
Another separate but potentially similar model relationship identified is that of natural mOFC increased activity boosting enjoyment-satisfaction showing correlation with expected quality of wine based upon price.
Again though what would be the long term results, and also if using those who test-taste wine most days as a wine producer.
So two aspects identified for sure I know of (appreciate some may be able to add more but just regarding recent discussions); pleasure-satisfaction-reduced/increased pain and chemical/brain activity region response within the brain.
If you really want to test placebo generally and say it works (and see limitations); see what happens to a long term highly physically dependent drug addict (no longer highs but taken for the body) on Heroin and replace this with a placebo and tell them it is the same drug.....
I need to dig out a few studies I have, one is interesting as it investigated the pain effect of placebo and its limitations-behaviour.
Unfortunately placebo term and how it is associated is too general due to multiple mechanisms actually at work and each with different traits.
Cheers
Orb