Acute Care Hospitals

Because of the substantial changes that have occurred in the environment of acute care hospitals (see Chapter 1), studies based on older data are not the most useful for understanding staffing effects. Rather, the strongest evidence comes from studies published in the last 15 years (Aiken et al., 1999, 2002; Amaravadi et al., 2000; Blegen and Vaughn, 1998; Blegen et al., 1998; Bolton et al., 2001; Bond et al., 1999; Dimick et al., 2001; Flood and Diers, 1988; Hartz et al., 1989; Hunt and Hagen, 1998; Kovner and Gergen, 1998; Kovner et al., 2002; Lichtig et al., 1999; Needleman et al., 2002; Pronovost et al., 2001; Shortell et al., 1994). All of these are cross-sectional studies that explored correlations between measures of nurse staffing levels and rates of adverse occurrences. They examined in-hospital deaths and nonfatal adverse outcomes, including various types of nosocomial infections, decubitus ulcers, and falls. A variety of acute care hospital settings were examined, including intensive care units (ICUs), general medical–surgical units, and various specialty units. In some studies, process errors were measured, including medication errors.

The amount of nursing service (staffing level) in a given unit or hospital typically is expressed administratively as nursing hours per patient per day (hppd). It is also expressed as a nurse-to-patient ratio, or the average number of patients for each nurse; for example, 1:4 or 1:6 represents one nurse for every four or six patients, respectively. Higher levels of hppd indicate higher nurse-to-patient ratios.1

An important methodological issue in studies of hospital staffing is the unit of analysis. Sometimes staffing-level data are obtained for individual nursing units within hospitals; at other times, staffing data are aggregated across the entire hospital. Measures of outcomes similarly are aggregated across individual patients to the unit or hospital level to produce an incidence rate of adverse events. A problem with hospital-level aggregation is that heterogeneous nursing units, such as pediatric units, labor and delivery units, adult medical and surgical units, and ICUs, are combined. As a result, data on hospital-wide staffing levels may not well represent the staffing levels experienced by patients in a given nursing unit or of interest to poten-


Discussions of nurse-to-patient ratios can often be confusing. A nurse-to-patient ratio is expressed as a numerical relation; e.g., one nurse for each six patients is a nurse-to-patient ratio of 1:6. Because this figure often resembles a fraction (e.g., 1/6), a “higher” nurse-to-patient ratio is one in which the ratio of nurses to patients, expressed as a fraction, comes closest to the whole number 1. That is, a 1:2 ratio (one nurse for every two patients) is a higher nurse-to-patient ratio than one nurse for every six patients (1:6). In this chapter, we attempt to avoid this confusion by using the expressions “more nurses” or “fewer nurses” per patient.

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