Sir: Sander (Psychiatric Bulletin, January 2001, 25, 33) is correct in pointing out some of the limitations of urinary detection of olanzapine as a proxy for compliance, as previously described by myself (Coates, Reference Coates1999, Reference Coates2000). Currently, only a negative result shows non-compliance, whereas a positive result is open to various interpretations. I am presently studying two ways of potentially addressing these shortcomings, which may prove helpful.

First, I am investigating the quantification of the urinary levels of olanzapine, rather than just using a qualitative test. This should provide more of an indication of the actual compliance when levels are ascertained. Second, the measurement of urinary metabolites, either quantitatively or qualitatively, may lead to a more sophisticated approach in the future. In particular, 10-N-glucuronide is the most abundant metabolite but 4′-N-desmethy-lolanzapine is correlated to clearance (Reference Callaghan, Bergstrom and PtakCallaghan et al, 1999) and this may give a better indication of a person's recent compliance.

Currently, however, non-detection of urinary olanzapine remains the best objective test of non-compliance and with these further developments it may prove to be even more valuable in clinical practice.