Significant reduction of antibiotic use in the community after a nationwide campaign in France, 2002-2007

Abstract

Background: Overuse of antibiotics is the main force driving the emergence and dissemination of bacterial resistance in the community. France consumes more antibiotics and has the highest rate of beta-lactam resistance in Streptococcus pneumoniae than any other European country. In 2001, the government initiated "Keep Antibiotics Working"; the program's main component was a campaign entitled "Les antibiotiques c'est pas automatique" ("Antibiotics are not automatic") launched in 2002. We report the evaluation of this campaign by analyzing the evolution of outpatient antibiotic use in France 2000-2007, according to therapeutic class and geographic and age-group patterns.

Methods and findings: This evaluation is based on 2000-2007 data, including 453,407,458 individual reimbursement data records and incidence of flu-like syndromes (FLSs). Data were obtained from the computerized French National Health Insurance database and provided by the French Sentinel Network. As compared to the preintervention period (2000-2002), the total number of antibiotic prescriptions per 100 inhabitants, adjusted for FLS frequency during the winter season, changed by -26.5% (95% confidence interval [CI] -33.5% to -19.6%) over 5 years. The decline occurred in all 22 regions of France and affected all antibiotic therapeutic classes except quinolones. The greatest decrease, -35.8% (95% CI -48.3% to -23.2%), was observed among young children aged 6-15 years. A significant change of -45% in the relationship between the incidence of flu-like syndromes and antibiotic prescriptions was observed.

Conclusions: The French national campaign was associated with a marked reduction of unnecessary antibiotic prescriptions, particularly in children. This study provides a useful method for assessing public-health strategies designed to reduce antibiotic use.

Figure 1. Antibiotic use and flu-like syndromes in France, from July 2000 to March 2007.

Weekly totals of antibiotic prescriptions and FLS cases per 100 inhabitants plotted against time.

Figure 2. Winter antibiotic prescriptions in France by region, from October 2000 to March 2007.

The number of October–March prescriptions is divided by the number of regional inhabitants for the respective year in each of 22 France's regions: Al (Alsace), Aq (Aquitaine), Auv (Auvergne), BN (Basse Normandie), Bou (Bourgogne), Br (Bretagne), CA (Champagne-Ardenne), Ce (Centre), Co (Corse), HN (Haute Normandie), Li (Limousin), Lo (Lorraine), LR (Languedoc-Roussillon), IDF (Ile de France), FC (Franche-Conté), MP (Midi-Pyrénées), NPDC (Nord-Pas de Calais), PACA (Provence-Alpes-Cote d'Azur), PDL (Pays de Loire), PC (Poitou-Charente), Pi (Picardie), RA (Rhones Alpes).

Figure 3. Antibiotic prescriptions in France per 100 inhabitants by age group, from July 2000 to March 2007.

The number of October–March prescriptions for each age group for each year is divided by the number of inhabitants of each age group.

Figure 4. One step ahead forecasts of the interrupted ARMA model and observed antibiotic prescriptions data.

October–March horizontal lines indicate the estimated average level by the interrupted ARMA model (in red) and the observed average level without any campaign effect (in black) each winter. The percentages listed above the peaks denote the ratio of change.