Outsourcing staff leads to in-house infections

Dire implications in era of managed care

Synopsis: Patients who acquired bloodstream infections were more likely to be hospitalized during periods when the nursing staff of the surgical ICU included more float and agency nurses.

Source: Robert J, et al. Infect Control Hosp Epidemiol 2000; 21:12-17.

Abstract: In order to determine risk factors for nosocomial primary bloodstream infections (BSI), Robert and colleagues compared all patients admitted during a 13-month period to a surgical intensive care unit who acquired primary BSI (case-patients, n = 28) with 99 randomly selected patients (controls) hospitalized in the same unit. Primary BSIs were defined according to Centers for Disease Control and Prevention criteria. The study took place in a 1,000-bed university-affiliated teaching hospital.

A case patient was defined as any patient hospitalized in the ICU for more than three days in whom a BSI was identified. Case patients were identified from a review of medical records and confirmed by infection control staff. A control patient was defined as any patient hospitalized in the same unit for more than three days who did not acquire a BSI. Control patients were randomly selected from a chronologically ordered list of all patients in that unit who were admitted during the survey period. Regular staff were defined as nurses permanently assigned to the unit. Pool staff were part-time or agency nurses who were assigned to the unit for varying periods to ensure adequate staffing.

BSI rate 4.6 per 1,000 patient days

During the study period, there were 5,601 patient days in the surgical ICU and a mean occupancy rate of 71%. The overall primary BSI rate in the unit was 4.6 per 1,000 patient days. Case and control patients were similar in age, APACHE II score, and type of central venous catheter (CVC). Case patients were more likely than controls to be hospitalized during the five-month portion of the study period that had a lower regular-nurse-to-patient and higher pool-nurse-to-patient ratios (period two) than during the prior eight months (period one; P < 0.001). Case patients also were more likely to be in the surgical ICU for a longer period (P < 0.001), to be mechanically ventilated longer (P < 0.001), to receive more antimicrobial agents (P < 0.001) and total parental nutrition (P < 0.001), to have more CVC days (P < 0.001), and to die during the hospitalization (P < 0.001).

Case patients had significantly lower regular-nurse-to-patient and higher pool-nurse-to-patient ratios for the three days before BSI than did controls. In a logistic regression model, only admission during period two (odds ratio [OR] 3.2; 95% CI, 1.2-8.2) and length of ICU stay (OR 1.3; 95% CI, 1.2-1.4) remained significantly associated with an increased risk for BSI. In second and third models that included covariates of length of ICU stay (CVC-, TPN-, and ventilator days), admission during period two remained independently associated with developing a BSI (ORs, 3.8 and 3.1, respectively). Several other variables (overall nurse-patient ratio, prior surgery, and antimicrobial treatment) were introduced as covariates but were not associated with an increased risk of BSI.

Commentary by Leslie A. Hoffman, PhD, RN, medical-surgical nursing chairwoman in the department of acute/tertiary care, the University of Pittsburgh School of Nursing.

The major finding of this study was that, in addition to other well-known risk factors, admission to a surgical ICU during a period when there is a change in composition of the nursing staff (fewer regular staff, more pool staff) may be a risk factor for acquiring a BSI. An initial comparison of case and control patients revealed no significant between-group differences in factors likely to increase risk of BSI. However, there were differences in the composition of the nursing staff.

Consequently, the 13-month survey period was divided into an eight-month period when pool staff were less frequently used and a subsequent five-month period when pool staff provided more care. The breakpoint was chosen as the month in which increased use of pool nursing staff started. The mean total nurse-to-patient ratio was slightly higher when more pool staff were present. However, BSI occurred more frequently (7.6 vs. 2.8 BSI/1,000 patient days). In addition, for the three days immediately preceding the BSI (or index date for controls), case patients had a lower regular-nurse-to-patient ratio and significantly higher pool-nurse-to-patient ratio than controls.

Staff composition is key

Prior studies have associated understaffing with increased risk of nosocomial infection in the ICU or the hospital as a whole. However, this study is the first to report that the composition of the ICU nursing staff, and not just the overall number of nurses, may be an important factor in increasing infection risk. It is possible that factors not measured in this study confounded the association between nursing staff composition and risk for BSI. However, with the exception of changes in house staff, Robert et al. noted no changes in patient care standards during the two comparison periods.

Those findings have implications for policy in this era of health care reform. Funds for staff nurse education have been severely reduced, and positions for clinical nurse specialists eliminated. The ICU is a high-cost center. Consequently, efforts to reduce health care costs may result in changes in ICU staffing, including substitution of more highly trained for less well-trained personnel, and outsourcing, such as using agency nurses rather than hospital personnel. At the same time, patient acuity (average severity of illness among patients in the ICU) is increasing. The findings of this study suggest that changes in nursing staff composition should be monitored closely, with attention given to potential direct and indirect costs associated with such changes.