The German findings on antipsychotic prescription in naturalistic conditions provide a good illustration of the question raised in our editorial regarding the risk of widespread use of antipsychotics in primary care (Reference Verdoux and BegaudVerdoux & Bégaud, 2004). Contrary to the statement of the author, there is no contradiction between her point of view and ours regarding the fact that antipsychotics are widely prescribed in primary care. The exact quotation of our text is ‘drugs assumed to be mainly prescribed by specialists, such as antipsychotics’, suggesting that there may be differences between assumptions and facts. The German findings are strikingly similar to those obtained using the social security insurance database in Aquitaine (south-west France), showing that general practitioners (GPs) were the most numerous prescribers (64%) of risperidone, followed by psychiatrists (34%) (Reference Martin, Bégaud and VerdouxMartin et al, 2004). More recently, we have carried out a survey questionnaire mailed to all GPs (n = 3829) practising in Aquitaine to explore management of early psychosis in primary care, with a response rate of 23% (further details available from the authors on request). As part of this survey, we collected data on prescription behaviour. We found that over the preceding month, one out of three GPs (34%) initiated at least one new treatment with one of the antipsychotic drugs marketed in France at this time (amisulpride, olanzapine, risperidone). Over the same 1-month period, nearly half of GPs (42%) reported a visit from at least one pharmaceutical company representative promoting these antipsychotic drugs. Hence, we have further reason to feel concerned by the widespread use of these drugs in primary care, probably, at least in part, as a result of drug promotion. As proposed by Thiels, it is thus of great interest to explore which specific subgroups are at greater risk of being exposed to increased use of antipsychotics. With respect to the hypothesis that there may be a trend favouring antipsychotic instead of benzodiazepine prescription in elderly patients, it is however difficult to conclude that such a strategy is more risky than beneficial for this population, considering the risks potentially associated with benzodiazepine prescription in the elderly (Reference Verdoux, Lagnaoul and BégaudVerdoux et al, 2004).