Nursing studies have shown that nursing care delivery changes affect staff and organizational outcomes, but the effects on client outcomes have not been studied sufficiently.
To describe, at the level of the nursing care unit, the relationships among total hours of nursing care, registered nurse (RN) skill mix, and adverse patient outcomes.
The adverse outcomes included unit rates of medication errors, patient falls, skin breakdown, patient and family complaints, infections, and deaths. The correlations among staffing variables and outcome variables were determined, and multivariate analyses, controlling for patient acuity, were completed.
Units with higher average patient acuity had lower rates of medication errors and patient falls but higher rates of the other adverse outcomes. With average patient acuity on the unit controlled, the proportion of hours of care delivered by RNs was inversely related to the unit rates of medication errors, decubiti, and patient complaints. Total hours of care from all nursing personnel were associated directly with the rates of decubiti, complaints, and mortality. An unexpected finding was that the relationship between RN proportion of care was curvilinear; as the RN proportion increased, rates of adverse outcomes decreased up to 87.5%. Above the level, as RN proportion increased, the adverse outcome rates also increased.
The higher the RN skill mix, the lower the incidence of adverse occurrences on inpatient care units.
Mary A. Blegen, RN, PhD, is an associate professor and associate dean, College of Nursing, The University of Iowa, Iowa City, IA.
Colleen J. Goode, PhD, CNAA, FAAN, is an associate vice president, Patient Services, University of Colorado Hospital, Denver, CO.
Laura Reed, MSN, RN, is a Department Personnel Specialist, University of Iowa Hospitals and Clinics, Iowa City, IA.
Accepted April 30, 1997.
Supported in part by a grant from MAIN/MNRS.
The authors wish to thank the staff members of the quality assurance office at University of Iowa Hospitals and Clinics for their assistance in compiling and checking the data used in this study.