Prevalence surveys in different countries have shown that about 1 in 10 hospitalized patients at any one time have acquired an infection and the incidence of infection is usually about 3% to 5%. Can this infection rate be significantly reduced and is there an irreducible minimum below which further reduction cannot be obtained? Cross-infection could probably be almost eliminated by the use of plastic isolators or life islands, and endogenous infections reduced by the wider and more rational prophylactic use of topical and systemic antimicrobial agents, and possibly a range of vaccines. However, the use of these techniques on all patients would clearly be impractical and too expensive for use in the foreseeable future. Excessive isolation techniques would have a detrimental psychological effect on many patients. In addition, failure to eradicate the organisms of the normal flora might be associated with more infections caused by antibiotic-resistant bacteria, fungi and viruses which would be more difficult to treat. Even if organisms could be eradicated, morbidity and mortality might not be correspondingly reduced. Priorities must therefore be set on the most effective use of staff and finances. These not only involve a consideration of infection, but all areas of clinical practice, eg, costs of cardiac transplants and provision of geriatric facilities. It therefore seems likely that little additional finance will be available for control of infection, other than possibly for transmissible infections with high mortality such as AIDS.