These are extensions of the more conventional randomised controlled trials which usually seek to find evidence that one intervention is statistically significantly better than another, non-inferiority trials just aim to say that one intervention is not statistically signficantly worse than the other. They are usually done comparing a new intervention with an existing one or comparing a cheaper one with the currently most used one.
Details #
Non-inferiority trials do occasionally get conducted in the psychosocial intervention realm but they are more common in the realm of pharmacological interventions and rightly so as, often but not always, drugs can be delivered “double blind” making the logic of the trial sound.
Probably there are three key things to know about non-inferiority trials. Firstly, that this is not the same as finding a “non-significant” outcome in a conventional trial; the evidence that is sought here is that the it is statistically significantly unlikely, given the trial results and numbers participating, that the one intervention is worse than the other by a pre-defined margin that would represent non-trivial inferiority. Secondly, they are, like any randomised controlled trial, only fully logical if the interventions can be delivered “double-blind”. Finally, that they are most useful if also analysed in terms of the costs as well as differences in change/outcome scores.
In our realm I think they are a huge waste of money unless the interventions really can be delivered double-blind. We should be spending the money we spend on trials of psycho-social interventions on analysing routine outcome data, practice-based evidence and practice-oriented research.
Try also #
- Blinding
- DBRCT = Double Blind, Randomised Controlled Trial
- Double blind
- Practice-based evidence (PBE)
- Practice-oriented research
- Randomised controlled trials
Chapters #
Not covered in the OMbook.
Online resources #
None linkely from me.
Dates #
First created 28.xii.25.
