Dr. Jaffe presented an update on the risk of HIV infection to health care workers. He noted that information has come from three sources: 1) analysis of national AIDS surveillance data: 2) prospective studies of injuries and exposures of health care workers: and 3) anecdotal reports.

The national AIDS surveillance data indicate that 5% of reported AIDS cases occur in health care workers; health care workers represent 5.7% of the US workforce. By itself, this would suggest that employment as a health care worker is not associated with a disproportionately large risk of AIDS. The health care workers with AIDS have an age and sex distribution that is similar to that of all AIDS patients; the health care workers with AIDS are more likely to be white, however. The health care workers with AIDS are more likely to he homosexual/bisexual men (73%) than are all AIDS patients (62%); conversely, they are less likely to admit to intravenous drug use.

A case-control study was conducted on an epidemic of bacteremia and meningitis caused by Serratia marcescens in the neonatal intensive care unit and special care nursery of a general hospital in Mexico City, Mexico. A 19.9% incidence of bacteremia and meningitis was recorded in contrast to 1.4% and 3.7% during pre-epidemic and postepidemic periods; a 69% mortality rate was observed. Peripheral IV catheters and the use of mixed IV fluids prepared in the wards were the major risk factors (P < 0.001). Rectal and nasopharyngeal cultures were positive in 68% of asymptomatic neonates and hand cultures were positive in 16.7% of personnel. Strains were resistant to all aminoglycosides and broad-spectrum penicillins, and belonged to the A5/8 bio-group. Containment of this outbreak was difficult because of failure to identify colonized infants early in the epidemic and because of persistent carriage of S marcescens by personnel. Comparisons between this hospital and tertiary care centers in Mexico suggest that in developing countries nosocomial infections could be of greater magnitude in secondary than in tertiary level centers.

Twenty-four outbreaks of nosocomial bloodstream infection (BSI) were investigated by the Centers for Disease Control from Jan 1, 1977 to Dec 31, 1987. Intravascular pressure monitoring devices (transducers) were the most commonly identified source of bacterial and fungal BSI outbreaks and were implicated as the source of infection in eight (33%) outbreaks. These included outbreaks caused by Candida parapsilosis (2), Serratia marcescens (2), Klebsiella oxytoca (1), Pseudomonas cepacia (1), Acinetobacter calcoaceticus (1), and one polymicrobial bacteremia outbreak due to Acinetobacter, Pseudomonas, Citrobacter, and Enterobacter species. In all eight outbreaks, reusable transducers improperly disinfected or fitted with domes that had been improperly sterilized served as reservoirs for the organism. Compared with nosocomial BSI outbreaks not related to transducers, those in which transducers were implicated as a reservoir involved a larger mean number of patients (24 v 9; P = 0.007), and were significantly more likely to involve intensive care unit patients (23/24 v 3/9; P = 0.025) and to have a longer mean duration (11 v 3 months; P= 0.007). These findings show that the characteristics of transducer- and non-transducer-related BSI outbreaks differ, and that centers using intravascular pressuremonitoring devices must be aware of and implement recommended infection control strategies for care and maintenance of these devices.

A cost-benefit analysis was used to assess four strategies to prevent catheter-associated urinary tract infections in an acute care setting. Routine use of catheters with sealed collection junctions, selective use of these catheters, and oral antibiotic prophylaxis all result in fewer deaths, fewer infections, and lower overall costs than not using any of these. Routine use of sealed junction catheters results in fewer infections and deaths than does selective use. When the cost of a nosocomial urinary tract infection is $500, routine use of sealed junction catheters is also less expensive than selective use in many circumstances. Oral antibiotic prophylaxis would result in the lowest net cost and the fewest deaths and infections, if it were as effective as parenteral prophylaxis, if more than 72% of patients received it, and if important negative factors such as selection of antimicrobial resistance and adverse drug reactions are not considered. When there is no extra cost of sealed junction catheters, their use is less expensive than the oral prophylaxis strategy if the total cost of oral prophylaxis, including the cost of adverse reactions, is greater than $15. If the extra cost of sealed junction catheters is $4 per unit, their use is less expensive than oral prophylaxis when its cost exceeds $35. Prevention of catheter-associated urinary tract infection reduces the overall cost of patient care, even when the prevention itself incurs costs. This analysis supports the routine use of sealed junction catheters in most acute care situations that require indwelling catheter drainage. Currently, we do not recommend routine antibiotic prophylaxis.