A retrospective investigation was made to compare the occupational risk of tuberculosis in personnel assisting human immunodeficiency virus (HIV)-infected and uninfected subjects with active tuberculosis.

We retrospectively reviewed 6 years of hospital activity in 3 units where HIV-infected patients with tuberculosis are hospitalized and in 2 units where non-HIV-infected tuberculosis patients are hospitalized. The risk of occupational tuberculosis in healthcare workers who assisted HIV-infected and non-HIV-infected patients with tuberculosis was investigated.

The risk of occupational tuberculosis in healthcare workers was studied by considering the numbers of potential source cases (hospitalized patients with tuberculosis) in the two conditions investigated (HIV-positive and HIV-negative). Both potential source cases and cases of tuberculosis in healthcare workers had to be microbiologically proven in order to be considered..

Seven cases of tuberculosis occurred in persons who cared for 85 HIV-infected subjects with tuberculosis, while only 2 cases occurred in staff members who took care of 1,079 HIV-negative tuberculosis patients over the same period (relative risk=44.4; 95% confidence interval = 8.5-438).

Tuberculosis seems no longer to be a neglectable risk in healthcare workers assisting patients with HIV infection. Further study is urgently needed to see whether such unexpectedly high dissemination of tuberculosis also is demonstrable in the community.

To describe the distribution of nosocomial infections among surgical patients by site of infection for different types of operations, and to show how the risk of certain adverse outcomes associated with nosocomial infection varied by site, type of operation, and exposure to specific medical devices.

Surveillance of surgical patients during January 1986-June 1992 using standard definitions and protocols for both comprehensive (all sites, all operations) and targeted (all sites, selected operations) infection detection.

Acute care US hospitals participating in the National Nosocomial Infection Surveillance (NNIS) System: 42,509 patients with 52,388 infections from 95 hospitals using comprehensive surveillance protocols and an additional 5,659 patients with 6,963 infections from 11 more hospitals using a targeted protocol.

Surgical site infection was the most common nosocomial infection site (37%) when data were reported by hospitals using the comprehensive protocols. When infections reported from both types of protocols were stratified by type of operation, other sites were most frequent following certain operations (e.g., urinary tract infection after joint prosthesis surgery [52%]). Among the infected surgical patients who died, the probability that an infection was related to the patient's death varied significantly with the site of infection, from 22% for urinary tract infection to 89% for organ/space surgical site infection, but was independent of the type of operation performed. The probability of developing a secondary bloodstream infection also varied significantly with the primary site of infection, from 3.1% for incisional surgical site infection to 9.5% for organ/space surgical site infection (p<.001). For all infections except pneumonia, the risk of developing a secondary bloodstream infection also varied significantly with the type of operation performed (p<.00l) and was generally highest for cardiac surgery and lowest for abdominal hysterectomy. Surgical patients who developed ventilator-associated pneumonia were more than twice as likely to develop a secondary bloodstream infection as nonventilated pneumonia patients (8.1% versus 3.3%, p<.001).

For surgical patients with nosocomial infection, the distribution of nosocomial infections by site varies by type of operation, the frequency with which nosocomial infections contribute to patient mortality varies by site of infection but not by type of operation, and the risk of developing a secondary bloodstream infection varies by type of primary infection and, except for pneumonia, by type of operation.

To describe 4 healthcare workers who developed measles despite pre-existing antimeasles antibody levels.

Hospital employees working in patient care areas from July through November 1990 were screened for measles antibody levels using a commercially available enzyme immunoassay (EIA). The clinical course and laboratory evaluation of the 4 healthcare workers who developed measles were reviewed.

An academic tertiary care children's hospital.

A convenience sample of resident physicians, nurses, ward clerks, Child Life workers, physical and occupational therapists, radiology technicians, and housekeeping staff were screened regardless of age, immunization status, or history of measles infection.

Of 1,311 employees working in patient care areas, 900 (68.6%) had sera tested for measles antibody. Fourteen (1.5%) were negative, 338 (37.6%) had low-positive antibody levels, 372 (41.3%) were mid-positive, and 171 (19%) were high-positive; 5 (0.6%) showed equivocal results. Four healthcare workers vaccinated in the past developed measles. All had positive pre-illness measles antibody levels and all had a significant rise in measles-specific IgG following infection. Three of them had received at least 2 live measles vaccinations prior to caring for patients with measles.

These cases raise concerns regarding detection of adequate protective measles immunity. We recommend that all healthcare workers observe respiratory precautions in caring for patients with measles.

To determine risk factors for and modes of transmission of an outbreak of acute nonbacterial gastroenteritis among residents and staff in a nursing home.

Cohort study of residents and questionnaire survey of employees.

One hundred twenty-bed nursing home in Alabama.

From July 11,1991, through July 25,1991,77 of 120 residents (attack rate = 64%) and at least 14 of 49 employees (minimum attack rate = 29%) developed acute gastroenteritis characterized by vomiting and diarrhea; few residents developed fever > 100°F. Nine residents required intravenous rehydration, and 2 residents died.

The risk of developing illness was greater for female residents (64/92 versus 13/28; relative risk [RR] = 1.5; 95% confidence interval [CI95 ]=1.0-2.3) and for employees who reported handling residents' soiled linen, stools, or vomitus more frequently (>5 times a shift versus = 5 times a shift: 7/13 versus 7/31; RR=2.4; CI95 = 1.1-5.4). Direct transmission of infection, probably via person-to-person spread, sustained the outbreak. Temporal clustering analysis demonstrated that the risk of becoming ill 1 or 2 days after a roommate became ill was significantly greater than that of becoming ill at other times during the outbreak (RR= 2.2; CI95 = 1.3-3.8). No Salmonella or Shigella species, ova, or parasites were identified from 12 fecal specimens obtained from ill residents.

Although stool and serum specimens were not available for viral studies, the clinical symptoms and incubation period were consistent with illness due to Norwalk-like viral agents. This outbreak emphasizes the severity of acute nonbacterial gastroenteritis among elderly and debilitated residents of nursing homes and the need for prompt use of enteric precautions in controlling outbreaks of gastroenteritis in these facilities.

To discuss the close affinity between the continuous quality improvement (CQI) concept of monitoring a process for the introduction of special causes and epidemic surveillance.

A case study of a CQI tool for infection control epidemic surveillance.

A 668-bed acute care hospital with 5 intensive care units supporting heart, liver transplant, and trauma teams. The infection control department consists of 2.5 full-time employee nurse epidemiologists and a medical director who is an infectious disease specialist.

The outbreak investigation illustrates principles of CQI and their relationship to epidemic surveillance through the use of applied statistical methods.

The expanding role of hospital epidemiology is enhanced by applying CQI concepts to improving health care. The application of epidemiologic tools and principles to the problems of nosocomial infections is strongly connected to the CQI concept of using dependable data to improve processes.