Legislation is being considered that would require more-detailed reporting on staffing.

Hospital staffing data A lack of uniform, reliable, and readily available data on hospital staffing similarly is widely cited as preventing better understanding of nurse staffing (Kovner et al., 2000; Needleman et al., 2001; Sochalski, 2001; Unruh, 2002). Researchers who want to examine national patterns of staffing frequently use data collected by AHA’s Annual Survey of Hospitals. However, this survey does not ask for staffing data by hospital unit; it collects aggregate staffing data at the level of the hospital, combining all different types of inpatient units (e.g., ICUs, labor and delivery, pediatrics), outpatient units, and any hospital-based long-term care units (Kovner et al., 2000; Spetz et al., 2000). It further collects data on all nurses and does not distinguish nurses providing direct patient care from those in purely administrative or managerial positions (Kovner et al., 2002). Moreover, while the survey asks hospitals to report full-time and part-time licensed nurses, it does not define “full-time” RNs and assumes that part-time licensed nurses work 20 hours per week on average. This assumption is inconsistent with data from the National Sample Survey of RNs, which indicate that part-time nurses work closer to 30 hours per week. The AHA staffing numbers are thereby likely to underestimate the hours worked by nursing personnel (Spetz et al., 2000). Furthermore, when hospitals do not respond to the survey, the AHA “imputes” a response. Therefore, some of the data are estimates rather than true self-reports, which may substantially reduce their accuracy (Spetz et al., 2000). Because of these limitations, some researchers use staffing data obtained from states (Lichtig et al., 1999; Needleman et al., 2002). However, not all states collect these data, and those that do often receive data that are incomplete and unreliable (Lichtig et al., 1999).

Specific ways of improving data on hospital nurse staffing include (1) counting all nursing staff (RNs, LVNs/LPNs, and NAs) in nurse staffing reports, (2) developing universal definitions of nurse categories and procedures for calculating full-time and part-time nursing staff, and (3) separately reporting staffing for inpatient and outpatient care and for specific nursing units (Needleman et al., 2001; Sovie and Jawad, 2001). Based on the importance of nurse staffing levels to patient safety, the role of the health care marketplace in promoting patient safety, and the current poor quality of nurse staffing data, the committee makes the following recommendation:

Recommendation 5-4. DHHS should implement a nationwide, publicly accessible system for collecting and managing valid and reli-



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