Approaching the quality of antibiotic prescriptions in primary care using reimbursement data.

Our objectives for this investigation were: (i) to design quality measures of outpatient antibiotic use that could be calculated at the individual general practitioner (GP) level using reimbursement data only; and (ii) to analyse the variability in antibiotic prescriptions between GPs regarding these measures in south-eastern France. Based on the literature and international therapeutic guidelines, we designed a set of quality measures in an exploratory attempt to assess the quality of antibiotic prescriptions. We performed a cross-sectional study of antibiotic prescriptions in adults in south-eastern France in 2009, using data from the outpatient reimbursement database of the French National Health Insurance (NHI). We carried out a cluster analysis to group GPs according to their antibiotic prescribing behaviour. Six quality measures were calculated at the GP level, with wide variations in practice regarding all these measures. A six-cluster solution was identified, with one cluster grouping 56 % of the sample and made up of GPs having the most homogeneous pattern of prescription for all six quality measures, probably reflecting better antibiotic prescribing. Total pharmaceutical expenses (per patient), penicillin combinations use, quinolone use and seasonal variation of quinolone use were all positively associated with a more heterogeneous and possibly less appropriate use of antibiotics in a multivariate analysis. These quality measures could be useful to assess GPs' antibiotic prescribing behaviour in countries where no information system provides easy access to data linking drug use to a clinical condition.