The need to know about drug prescribing practices, in order to assess their relevance, has led us to perform a drug prescription study in two medical departments of Cardiology at 'Neuro-Cardiologique Hospital' in Lyon (France). In moving from theoretical project design to its practical achievement, important methodological problems, initially unsuspected, were encountered. For each main point of the study (choices regarding medical department, the person holding the inquiry, data collection, etc.) and any subsidiary operations (data key-boarding, codification and analysis), the main problems encountered are described after considering advantages and limits of potential solutions. We propose useful and relevant questions to be asked, before performing a drug prescription study, even a modest one, or interpreting the results of such a study.