Simpson et al (Psychiatric Bulletin, November 2005, 29, 410-412) conclude from their naturalistic audit of the National Institute for Clinical Excellence (NICE) criteria for the treatment of Alzheimer's disease with cholinesterase inhibitors that ‘the withdrawal of medication in line with NICE guidance is poor clinical practice and likely to have adverse outcomes in a large proportion of cases’.

Without a proper control group these conclusions are hard to support. It is possible that there may have been a greater deterioration in the group of 25 patients whose Mini-Mental State Examination score fell below 12 had cholinesterase inhibitors not been withdrawn. We cannot speculate that this group would have had better outcomes if they had continued on the medication.

A useful quote which aids the understanding of audits is ‘Research is concerned with discovering the right thing to do; audit with ensuring that it is done’ (Reference SmithSmith, 1992). Good published clinical audit projects can inform others on best methods and be used as comparators for other clinical audit work (Reference Talbot, Reynolds and StoneTalbot et al, 1997). The value of the naturalistic audit by Simpson et al lies in informing others about a possible discontinuation syndrome off cholinesterase inhibitors, but it would be interesting to compare their results with other similar audits involving withdrawal of medications according to NICE guidance. However, to discover the right thing to do and to validate their conclusions, only a well conducted randomised controlled withdrawal study would suffice.