Duke 2008 | ASHE | Equity and Efficiency in Health and Healthcare

There is emerging evidence that higher nurse staffing levels are associated with better patient outcomes in acute care hospitals. Data availability has limited this previous research: studies using secondary data for large samples of hospitals have used annual data that combined all units (wards) in the hospital into a single observation. Some studies have used micro data at the unit level, but these studies have been limited to small samples. Only one previous study has used longitudinal data. Our project is extending this work by using four years of monthly, unit-level data for a large panel of hospitals to obtain more precise estimates of the effects of nurse staffing on patient outcomes. It also extends previous work by considering the effects of human capital factors such as nursing education, experience, and turnover.

Data were obtained for all Veterans Affairs (VA) acute inpatient care for Fiscal Years 2003-2006. Nursing sensitive patient outcomes were derived using the Agency for Healthcare Research and Quality Patient Safety Indicators (PSI) software. Monthly nurse staffing for each unit were obtained from the VA Decision Support System. The human capital variables were extracted from VA payroll data (PAID). All variables were aggregated by month for each inpatient unit.

Fixed effects regressions were used to examine the effects of nurse staffing and nursing human capital on patient outcomes, controlling for the number of patient discharges, patient age and case-mix, and hospital characteristics. A non-linear functional form was used for nurse staffing levels. Because of the differences between them, separate models were estimated for intensive care units (ICUs) and acute care units.

Initial results have been run for 401 acute and 173 ICU units. Consistent with previous studies, higher levels of nurse staffing were significantly associated with reduced PSI rates for some, but not all, of the nursing-sensitive PSIs. We also found that higher levels of firm-specific nursing human capital were associated with reduced rates for some of the PSIs. Further, the estimated effects of nursing staffing on patient outcomes tended to be smaller than what has previously been reported. We are investigating if these differences in results are due to our use of more detailed data (unit-level monthly data vs. hospital-level annual data) or to our inclusion of additional human capital factors.

This study extends the literature on the relationship between nurse staffing and patient outcomes by using monthly data from each inpatient unit of the hospital, and by considering human capital factors that have not been considered in most previous studies. In contrast, most previous studies have used annual data for all inpatients, with no controls for human capital factors. Given the size of our sample and the detailed information that we have available about nurse staffing, this study significantly increases the strength of the evidence that nurse staffing affects patient outcomes. The fact that other characteristics of the nurse staff besides the number of nurses affect patient outcomes implies that minimum nurse staffing requirements may need to be revisited.