Between March and December of 1979, an outbreak of infections due to multiply antibiotic resistant Serratia marcescens took place in a 50-bed neonatal intensive care unit. Fifteen neonates suffered major infections (sepsis, meningitis and pneumonia) with one death, and 20 suffered minor infections (conjunctivitis, cystitis, wound infections). Epidemiologic investigation failed to reveal a common source; S. marcescens, however, was isolated from an employee's hand, emollient skin cleanser, suction tubing, and three in-use manual infant resuscitation bags. The skin cleanser and equipment-cleaning agents were ineffective against S. marcescens. Asymptomatic, colonized infants were the major reservoir of S. marcescens. These infants were identified by daily cultures of the nose, umbilicus and rectum. The rectal swab most commonly (76%) yielded first-positive cultures in previously uncolonized infants, and was ultimately positive in 92% of colonized infants. A control program was begun by: 1) removing all inanimate sources of S. marcescens; and 2) cohorting patients and staff into a S. marcescens-exposed group and a new patient group. The new patient group of infants was surveyed by daily triple-site cultures for colonization and subsequent transfer to the S. marcescens-exposed group. After four months, the epidemic was controlled and the organism eradicated from the neonatal intensive care unit.

An infection control program was instituted at The Victoria General Hospital, an 800-bed acute care hospital, in July 1977. Serratia marcescens had infected or colonized (I/C) 225 to 232 patients yearly for each of the three previous years. Since this organism is usually acquired nosocomially, we decided to use Serratia I/C as a marker for our infection control program. During the years 1977 to 1980, we identified and eliminated several reservoirs of Serratia (contaminated urine measuring containers, urometers, diabetic urine testing equipment and in-use contamination of 2% Hibitane). Readmission of previously I/C patients proved to be an increasingly important reservoir. During 1980, only 120 patients were I/C, and gentamicin-resistant isolates of S. marcescens had dropped from 44% in 1977 to 4.4% in 1980. Use of Serratia as a marker enabled us to monitor the efficacy of our infection control program and allowed us to prove to our health care workers the usefulness of many of the measures we introduced.

The average charge per patient due to nosocomial infection for 215 nosocomial infections in 183 study patients was $693. These costs, however, were concentrated in very few patients; 5% of patients accounted for nearly one-third of total charges. The 10% of patients with highest nosocomial infection costs were patients on Medical or Surgical services; these services were utilized in 71% of patients with nosocomial infection and accounted for 86% of the attributable charges. Among the 22 most costly infections, 17 occurred in surgical wounds and lower respiratory tract. Although these sites accounted for 46% of the infections, they resulted in 77% of the total nosocomial infection charges. Patients with a primary diagnosis of injury had particularly costly infections. Combined analysis of these variables revealed two groups for whom nosocomial infections were especially costly: surgical patients who acquired wound infections after injuries, and medical patients with lower respiratory infections.

Gram-negative bacilli frequently cause epidemics in high-risk newborn intensive care units. Recently, an epidemic caused by a multiply-resistant K. pneumoniae, serotype 21, occurred in the Vanderbilt University intensive care nursery. The background of this outbreak included an increasing endemic nosocomial sepsis rate, operation of the facility in excess of rated capacity, and increasingly inadequate nurse-to-patient staffing ratios. The epidemic lasted 11 weeks; 26 (12%) of the 232 infants at risk in the unit became colonized. Five infants developed systemic illness and one died. Cohorting, reinforcement of strict handwashing and isolation procedures, and closure of the unit to outborn admissions resulted in rapid termination of the outbreak. Followup studies performed on infants colonized with the epidemic bacterium demonstrated persistent fecal shedding up to 13 months following discharge from the hospital. This epidemic had a detrimental influence on high-risk newborn and obstetric health care delivery in an area encompassing portions of three states. Under a system of progressively more sophisticated referral units, nosocomial infections occurring at a tertiary center can have an impact on other hospitals within the network.