Objective: The aims of this study were to analyse (1) if it is possible to classify general practitioners (GPs) or groups of practitioners (practice units) as overall narrow or wide prescribers and (2) to what extent the size of practice drug formularies is associated with general prescribing behaviour, practice activity and demography. Methods: Data were retrieved from the Odense University Pharmacoepidemiologic Database (OPED) and the prescribing patterns of 177 practice units were analysed in a 4-year period including 2000 and 2003. We characterised the drug repertoire for new treatments using a Formulary Diversity Scale based on two formulary measures, the DU 90% and the Formulary Selectivity Index. Multiple regressions were used to analyse the association between formulary size and practice characteristics. Results: The final Formulary Diversity Scale comprised 15 therapeutic drug groups and 151 different analogues. The average score based on the DU 90% segment was 46 drugs (range: 28-65) corresponding to 30% of drugs available. Using the Formulary Selectivity Index the average diversity score was 0.70. Practices with a high number of patients and a high percentage of elderly patients were more likely to use a wide range of drugs. Conclusion: The Formulary Diversity Scale revealed wide variation between practice units, indicating the existence of both overall narrow and wide prescribers among GPs. The number of patients attending the practice and the percentage of elderly patients were directly associated with the use of more analogues.