There is a growing tendency to separate out responders and non-responders in studies and report on the success of responders. This stinks like fish…
- Recent pain reports including many Fibromyaglia drug studies like to remove non-responders and concentrate on findings of responders.
- Some of those are going to be placebo effect cases – you many be able to tell because their response may be too early for the medication to take effect. In some depression studies, those who remit in the first 1-2 weeks are eliminated due to anticipated placebo effect – in pain studies it might be wise to do the same but this would make their results much less pretty.
- non-responders are part of the herd and they need to be included in analysis so people get a clear impression of what to expect.
- There is a legitimate reason to eliminate cases – the are called outliers and there are strict statistical requirements for evoking that process – Com’on guys be legit!!
any comments?