To assess the prevalence of tuberculosis (TB) or a positive skin test in healthcare workers (HCWs) providing services to human immunodeficiency virus (HIV)-infected individuals and to determine prospectively the incidence of new infections in this population.

This prospective cohort study enrolled 1,014 HCWs working with HIV-infected populations from 10 metropolitan areas. Purified protein derivative (PPD) tuberculin skin tests were placed at baseline and every 6 months afterwards on those without a history of TB or a positive PPD. Demographic, occupational, and TB exposure data also were collected.

Outpatient clinics, hospitals, private practice offices, and drug treatment programs providing HIV-related healthcare and research programs.

A voluntary sample of staff and volunteers from 16 Community Programs for Clinical Research on AIDS units.

Factors related to prior TB or a positive skin test at baseline included being foreign-born, increased length of time in health care, living in New York City, or previous bacille Calmette-Guerín vaccination. The rate of PPD conversion was 1.8 per 100 person years of follow-up. No independent relation was found between the amount or type of contact with HIV-infected populations and the risk of TB infection.

These data provide some reassurance that caring for HIV-infected patients is not related to an increased rate of TB infection among HCWs in these settings.

To use DNA fingerprinting to characterize nosocomial spread of Mycobacterium tuberculosis following hospitalization of a patient with acquired immunodeficiency syndrome and active pulmonary tuberculosis, for whom respiratory isolation was not initiated promptly.

Epidemiological investigation.

A tertiary-care medical center in Tennessee.

Patients and healthcare workers potentially exposed to the infectious patient in 1992.

Of 172 healthcare workers exposed, 35 (20%) were judged to have acquired tuberculous infection. Risk of acquisition was greatest for nurses and medical receptionists. Active tuberculosis later developed in one healthcare worker and one hospitalized patient. Nosocomial transmission was supported by epidemiological evidence and DNA fingerprinting. The outbreak strain of Mycobacterium tuberculosis differed from other isolates at this hospital, but its DNA hybridization pattern was highly similar to that of the multidrug-resistant outbreak strain W that has been prevalent in New York City, suggesting a common strain ancestry. However, the Tennessee isolates were susceptible to all first-line antituberculous agents.

This report suggests the possibility that a molecular characteristic(s) shared by these successful outbreak strains is associated with increased transmissibility or pathogenicity and emphasizes the need for continued vigilance for tuberculosis in the nosocomial setting.

To follow over time surgical-site infection (SSI) rates in patients admitted to the hospital on the same day as their elective surgery (group I), and to compare these rates with those of other surgical patients who also required postoperative hospitalization but either were hospitalized for 1 or more days immediately prior to surgery or underwent surgery on a nonelective basis (group II).

Observation of overall SSI rates over time for group I and II patients and study of risk factors associated with increased SSI rate among group I neurosurgical patients.

A 900-bed, tertiary-care community hospital with >10,000 surgical procedures performed each year on patients requiring postoperative hospital stay.

A total of 48,464 surgical procedures were performed on consecutive patients from 1990 to 1994, with 18,794 (39%) involving group I patients. In addition, 511 consecutive procedures performed on neurosurgical patients requiring postoperative hospitalization from July 1994 to May 1995 were analyzed.

In 1990, the overall SSI rate of group I patients was significantly lower than that of group II patients (0.4% vs 1.3%, relative risk [RR], 3.6; 95% confidence interval [CI95], 2-6.4; P<.0001), but, by 1994, there was no longer any significant difference between the SSI rates of group I and II patients (1.8% and 1.6%, respectively; P=.38). In 1994, the SSI rate in group I neurosurgical patients was significantly greater than that of group II patients (3.4% vs 0.4%; RR, 8.3; CI95, 1.05-66; P=.02). During the period April through September 1994 (warm months), group I neurosurgical patients were associated with a significantly lower American Society of Anesthesiologists score and SSI risk index (based on National Nosocomial Infection Surveillance System data) and a higher likelihood of having “clean-class” wounds, but significantly greater-than-expected SSI rate when adjusted for patient risk index and type of procedure (4.7% vs 1.4%; RR, 3.3; CI95, 1.3-8.6; P=.02). During the same period, the observed and expected SSI rates were not significantly different for group II neurosurgical patients.

As a whole, the SSI rates among patients undergoing elective surgery on the same day of their hospital admission is no longer significantly lower than that of other patients who also require postoperative hospitalization. For certain procedures, such as those commonly performed on the neurosurgical service, elimination of preoperative hospital stay may be associated with greater-than-expected risk of SSI.

To assess the effect on staff- and patient-related complications of a needleless intermittent intravenous access system with a reflux valve for peripheral infusions.

A 6-month cross-over clinical trial (phase I, 13 weeks; phase II, 12 weeks) of a needleless intermittent intravenous access system (NL; study device) compared to a conventional heparin-lock system (CHL, control device) was performed during 1991 on 16 medical and surgical units. A random selection of patients was assessed for local intravenous-site complications; all patients were assessed for the development of nosocomial bacteremia and device-related complications. Staff were assessed for percutaneous injuries and participated in completion of product evaluations. A cost analysis of the study compared to the control device was performed.

A 1,100-bed, teaching, referral medical center.

594 patients during 602 patient admissions, comprising a random sample of all patients with a study or control device inserted within a previous 24-hour period on study and control units, were assessed for local complications. The 16 units included adult inpatient general medicine, surgical, and subspecialty units. Pediatrics, obstetrics-gynecology, and intensive-care units were excluded. All patients on study and control units were assessed for development of nosocomial bacteremia and device-related complications. All staff who utilized, manipulated, or may have been exposed to sharps on study and control units were assessed for percutaneous injuries. Nursing staff completed product evaluations.

The study device, a needleless intermittent intravenous access system with a reflux valve, was compared to the control device, a conventional heparin lock, for peripheral infusions.

During the study, 35 percutaneous injuries were reported. Eight injuries were CHL-related; no NL-related injuries were reported (P=.007). An evaluation of 602 patient admissions, 1,134 intermittent access devices, and 2,268 observed indwelling device days demonstrated more pain at the insertion site for CHL than NL; however, no differences in objective signs of phlebitis were noted. Of 773 episodes of positive blood cultures on study and control units, 6 (0.8%) were device-related (assessed by blinded investigator), with no difference between NL and CHL. Complications, including difficulty with infusion (P<.001) and disconnection of intravenous tubing from device (P<.001), were reported more frequently with CHL than with NL. Of nursing staff responding to a product evaluation survey, 95.2% preferred the study over control device. The projected annual incremental cost to our institution for hospitalwide implementation of NL for intermittent access for peripheral infusions was estimated at $82,845, or $230 per 1,000 patient days.

A needleless intermittent intravenous access system with a reflux valve for peripheral infusions is effective in reducing percutaneous injuries to staff and is not associated with an increase in either insertion-site complications or nosocomial bacteremia. Institutions should consider these data, available institutional resources, and institution-specific data regarding the frequency and risk of intermittent access-device-related injuries and other types of sharps injuries in their staff when selecting the above or other safety devices.