Reduction in outpatient antibiotic sales for pre-school children: interrupted time series analysis of weekly antibiotic sales data in Sweden 1992–2002

OBJECTIVES: The aim of this study was to use detailed weekly data on outpatient antibiotic sales for pre-school children in Sweden to test for the significance of trends during 1992-2002. We also report on the special features found in weekly antibiotic data, and how the interrupted time series (ITS) design can adjust for this. METHODS: Weekly data on the total number of dispensed outpatient antibiotic prescriptions to pre-school children were studied, as well as the individual subgroups commonly used to treat respiratory tract infections in children: narrow-spectrum penicillins, broad-spectrum penicillins and macrolides. In parallel, monthly data of paracetamol sales of paediatric dosages were analysed to reflect trends in symptomatic treatment. An ITS model controlling for seasonality and autocorrelation was used to examine the datasets for significant level and trend shifts. RESULTS: A significant increase in mean and change in level could be found in the total antibiotic data in 1997, also reflected in broad-spectrum penicillin data where a similar trend break occurred in 1996. For macrolides, a trend break with a decrease in mean was noted in 1996, but no trend breaks were found in narrow-spectrum penicillin data. In contrast to the general decreasing trends in antibiotic sales, the yearly over-the-counter sales of paracetamol in paediatric preparations increased during the same period, with no identified trend breaks. CONCLUSIONS: The overall decrease in antibiotic sales and increase in paediatric paracetamol sales might suggest that symptomatic treatment in the home has increased, as antibiotics are less commonly prescribed.