the ICU nurse-to-patient ratio had no effect on any of the outcomes analyzed, giving weight to the suspicion that variables other than nurse staffing per se affect patient outcomes in the hospital context.
In a study examining Medicare mortality rates, Aiken and colleagues (1994) found that magnet hospitals (that is, hospitals with low RN turnover and vacancy rates and high levels of RN satisfaction) have lower patient mortality than control hospitals. In a presentation to the IOM committee (October, 1994), Aiken summarized the study findings. These findings indicate that lower Medicare mortality rates, as well as improved work-related well-being for RNs, are linked to hospital organization characteristics that result in RNs having: (1) more autonomy to provide care in their professional roles and within their areas of expertise; (2) greater control over what other care givers do in the patient care environment and over resources; and (3) well-documented and well-developed professional relationships with physicians.
Aiken and colleagues also conducted a study of specialized AIDS care units. The combined results of these two studies provide interesting information concerning the organization of nursing care. The magnet hospital study indicates that the preferred organizational structure is one in which the hospital management sees its primary responsibility as delivering patient care, and therefore both places a high value on the quality of nursing services and actively supports the professional role of nursing services. The research on specialized AIDS care units demonstrated how, in the absence of the preferred hospital-wide organization of nursing services, unit-level organization of care can help create environments where the RN autonomy and control that promote lower mortality rates can develop. Specifically, Aiken and colleagues found that AIDS care units foster RN autonomy and control through RN specialization (which promotes autonomy and interaction with physicians based on mutual expertise) and the correlation between patients' high care needs and RNs' areas of specialization. Furthermore, these two studies confirm that the same factors that lead hospitals to be identified as effective from the standpoint of the organization of nursing care are associated with lower mortality among Medicare patients.
Aiken and colleagues concluded that although RN-rich staffing ratios are sometimes associated with improved outcomes, the results of their research indicate that such staffing ratios are essentially a proxy measure for other organizational attributes of hospitals that grant nurses autonomy over their own practice and control of the resources necessary to deliver patient care and create good relationships with physicians. The committee concurs with the findings of Aiken and colleagues (1994, p. 783), "that the mortality effect derives from the greater status, autonomy and control afforded nurse in the magnet hospitals, and their resulting impact on nurses' [RNs'] behaviors on behalf of patients—i.e., this is not simply an issue of the number of nurses, or their mix of credentials."
Clearly, one of the research challenges in determining the relationship between staffing and quality of care has been the difficulty of isolating the factors