Unit-level staffing, workload, and adverse events in Army Acute Care Hospitals: 2003-2006

Abstract

Background: A sequence of reports from the Institute of Medicine revealed quality and patient safety issues in hospitals. The crucial role of nurses was recognized, although much of the research has addressed staffing and, more recently, the work environment. Moreover, most of this research has been at the hospital level where attribution of care processes to adverse events is difficult. The Army, Air Force and Navy's hospitals have faced the same economic and quality challenges. However, military medical institutions also have the increased demands of supporting a nation at war. Purpose: This study examined unit level impact of nursing staffing and workload on medication errors and patient falls in Army hospitals between 2003 and 2006. Methods: A descriptive correlational longitudinal design was used to conduct a secondary analysis of 23 Army inpatient units from the Military Nursing Outcomes Database (MilNOD). Relationships among staffing, workload, and quality of care were examined from 2003 to 2006. A cross sectional design using only 2006 data examined the influence of practice environment on outcomes. Generalized Linear modeling (GZLM) was used to accommodate nested data. Results: Large turbulence was expected in the Army inpatient units from 2003-2006. Although some years were significantly higher, turbulence was far less than expected. Staff complement (measure of percent mix of military, civilian and contractor staff) was a significant predictor of medication errors. Patient census was a significant predictor of falls. The professional nursing Practice Environment Scale (PES) was a partial mediator of medication errors in all types of units. Although tested as both, the practice environment did not significantly mediate or moderate falls. Conclusion: This study supports the growing literature on nurse staffing and the influence on patient outcomes. This study was unique in that data was collected at the unit/shift level, outcomes assessed are considered nurse-sensitive, and statistical techniques accounted for the nested data. Despite these advantages, limitations of the measurement of nursing-related acuity and patient turnover are acknowledged. Well-defined and consistent measures are necessary to enhance interpretability and application across nursing settings.