ABSTRACT THE EFFECTS OF CHANGES IN THE HOSPITAL NURSING WORKFORCE AND PRACTICE ENVIRONMENT ON THE OUTCOMES OF SURGICAL ONCOLOGY PATIENTS: A TWO-STAGE PANEL STUDY Jill M. Vanak Eileen T. Lake Prior research has documented that a better educated nursing workforce, higher nurse staffing levels, and better nurse practice environments are significantly associated with improved quality of care and lower patient mortality in multiple patient populations. Most prior research has used cross-sectional data to analyze associations between variables at a single time point. Little research has addressed whether changes in hospital nursing characteristics over time are associated with changes in outcomes. This two-stage panel study, in which cross-sectional samples of patients and nurses in acute care hospitals in Pennsylvania were compared at two time points, provides evidence of the relationship between changes in hospital nursing characteristics and patient outcomes. The objective of this study was to examine the effects of changes in hospital-level proportion of baccalaureate-prepared nurses, nurse staffing, and the nurse practice environment on changes in rates of failure-to-rescue and 30-day mortality in a surgical oncologic patient population between two points in time. The study was a two-stage panel designed secondary analysis that examined the effect of changes between 1999 and 2006 in nursing characteristics in 135 hospitals on changes in risk-adjusted mortality and failure-to-rescue of 29, 356 adult oncology patients admitted for primary surgical intervention for the purposes of disease management. The study combined information about nursing characteristics from nurse surveys with patient characteristics and outcomes derived from a state cancer registry and hospital discharge abstracts and hospital characteristics drawn from administrative databases. Multivariate regression modeling was employed to jointly assess the effect of changes in the organization of nursing within an institution on outcomes, controlling for both patient and hospital characteristics. The overall mean percentage of nurses with a baccalaureate degree across hospitals did not change significantly between 1999 and 2006. The mean number of patients per nurse across all hospitals was 5.81 in 1999 and 5.76 in 2006, a non-significant change, with the vast majority of hospitals decreasing or increasing the average number of patients per nurse by less than one patient. Nurse-reported practice environment scores increased significantly during the study period. A number of hospitals had increases in level of nurse education, nurse staffing, and rating of the practice environment over the period, while many others had decreases. Some of the changes in both directions were sizable. Improvement in nurse staffing was associated with reductions in failure-to-rescue and mortality rates. The addition of one patient to the nurse's average workload resulted in an average increase of 4.34 deaths for every 1,000 patients. For the subset of patients with complications, the addition of one patient to the nurse's average workload resulted in an average increase of 13.47 deaths for every 1,000 patients. When controlling for patient characteristics, with every 10% increase in the proportion of nurses with a baccalaureate degree, hospitals had an average reduction of 5.07 deaths for every 1,000 patients. This association was not significant in models that controlled for hospital characteristics. Investments in hospital nursing features including increasing the proportion of baccalaureate-prepared nurses and lowering patient-to-nurse ratios within hospitals may contribute to improvement in outcomes of surgical oncology patients.