A co-ordinated survey of 3899 medical patients in 169 wards, performed simultaneously in eight countries, showed a point-prevalence of urinary-tract infection (UTI) and bacteraemia of 12.6 and 1.6 per cent, respectively. One-half of the infections were acquired after the patients' admission. The bacteriological patterns of hospital- vs community-acquired infections were different, but showed no unexpected features. Antibiotic treatment was recorded in 22.3 per cent of the patients in this study, urinary-tract disinfectants, sulphonamides or penicillins being used in 95 per cent of those treated for UTI. The overall prevalence of urinary-tract drainage was 11.0 per cent with no significant difference between the two sexes. At ward level the rate of catheterized patients varied from below 5 per cent to more than 25 per cent, indicating--besides variations in the ward populations--differences in policies. The association between nosocomial UTI and the presence of an indwelling catheter and/or female sex was confirmed, while high age appeared to be a secondary risk factor among catheterized patients. The prevalence of nosocomial bacteraemia in patients with UTI was five times higher than in those without urinary-tract involvement, and a significant part of the nosocomial cases of both UTI and bacteraemia was clearly device-related. Guidelines for the use of indwelling catheters should be restrictive and provide for prompt removal. When introduced and followed they will effectively reduce nosocomial UTI and bacteraemia.