A series of studies of ICU outcomes conducted between 1994 and 1996 in all nonfederal, short-stay hospitals in Maryland found that 82 percent of these hospitals had day-shift ICU nurse staffing levels of one nurse for every one to two patients. Lower staffing levels—i.e., nurses caring for three or more ICU patients—were reported for 18 percent of hospitals. After adjusting for patient characteristics and for hospital and surgeon volume, patients who had abdominal aortic surgery in hospitals with fewer ICU nurses (i.e., each nurse caring for three or more patients) on the day shift were more likely to have postoperative complications, particularly pulmonary insufficiency and reintubation (Pronovost et al., 2001).

A second analysis of these data examined ICU direct-care nurse staffing on day and night shifts. Nurse staffing was coded as either low intensity (1:3 or greater nurse-to-patient ratio on the day and night shifts); medium intensity (1:3 or greater on either the day or night shifts, but not both), or high intensity (1:2 or lower on both day and night shifts). The majority of hospitals (63 percent) were staffed at a high-intensity level; 21 percent were staffed at a mixed-intensity level; and 16 percent had low-intensity staffing. After controlling for patient and organizational variables, the analysis showed that patients cared for on units with medium-intensity staffing were more likely to have cardiac and other complications than were patients cared for on high-intensity units. Patients cared for on units with low-intensity nurse staffing were more than twice as likely to have respiratory complications as patients on units with high-intensity staffing. Patients were more than five times as likely to develop pulmonary insufficiency and were more than twice as likely to be mechanically ventilated after 96 hours and reintubated when cared for on units with low-intensity staffing as compared with units with high-intensity staffing (Dang et al., 2002).

These sources and others (Amaravadi et al., 2000; Fridkin et al., 1996) indicate that nurse staffing levels of 1:2 or better not only are commonly used by large numbers of ICUs, but also have a protective effect on patients.

Medical–surgical units Information on medical–surgical staffing levels is available from two states and one multihospital, multistate data set. In California, CalNOC data show an average of 5.9 patients assigned to individual medical–surgical nurses across all shifts (Donaldson et al., 2001). California OSHPD data show similar average nurse-to-patient ratios of 1.0:5.2, with a median of 1.0:5.8 (Spetz et al., 2000). An examination of nurse staffing ratios within individual shifts from a convenience sample of representative medical–surgical units from 28 percent of California hospitals showed variation in staffing across shifts and by rural/urban status. This study estimated staffing levels using two methods: (1) computing a nurse staffing ratio based on the hospital-reported number of hours in a shift and the RN hours per patient for the shift, and (2) using staffing ratios



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement