Hair analysis has been presented by some authors as a possible tool of investigation for estimating patients' compliance to long-term therapies. This paper summarises the different publications that have been devoted to this topic and highlights the available statistically data presented to support this proposition. Qualitative results of such determinations may be of some interest but due to the enormous interindividual variations of quantitative data, the idea of using hair analysis to ascertain whether a patient has taken his treatment exactly as prescribed, clearly appears to be inapplicable.