Antibiotic resistance is now generally accepted as a major public health issue, but it is less clear which interventions should be used and which will work. Our principal objective is to minimise antibiotic resistance and increase the likelihood that standard treatments continue to be effective, for many decades, in patients who genuinely need them. We also wish to eliminate adverse reactions and waste in health funding due to unnecessaryuse.

Direct evidence of the positive effects of intervention on reducing resistance is scanty. We know that it is possible to reduce antibiotic consumption in communities and countries. For instance, in the five years to 1994 per capita usage of antibiotics actually dropped in Australia, but in six other Western countries consumption has increased steadily over a decade.1 Similar sharp drops in use of penicillin in Hungary followed the discovery of high rates of resistance in pneumococci.2 In neither case is there yet direct evidence that resistance rates have dropped as a result. Recently, investigators in Finland have shown a considerable drop in the prevalence of erythromycin resistant Streptococcus pyogenes after a national campaign to reduce use of macrolides.3 Other researchers, however, have shown significant variation in the prevalence of erythromycin resistance in S pyogenes in the absence of any intervention to change macrolide use,4 suggesting that natural fluctuations in the spread of strains may also influence prevalence of resistance. Thus we are left largely with empirical and commonsense strategies.

Several learned groups, such as the Fogarty International Centre, a task force for the American Society for Microbiology, and the World Health Organisation, have produced recommendations about appropriate action.5^–7 International experts have also offered sound …