Disparities between two common data sources on hospital nurse staffing

Purpose: To compare nurse staffing measures derived from two widely used data sources: the American Hospital Association (AHA) Annual Survey of Hospitals and the California Office for Statewide Health Planning and Development (OSHPD). Design: Descriptive cross-sectional study with measures of nurse staffing level and skill mix constructed from each database for 372 nonfederal, acute care hospitals in California. Methods: Discrepancies in nurse staffing estimates between the two databases were examined. Relationships of nurse staffing with risk-adjusted patient outcomes (decubitus ulcer, failure to rescue, and mortality) were assessed through multivariate analyses and compared for nursing measures derived from the two databases. Findings: For small, rural, or nonteaching hospitals, AHA reported substantially higher registered nurse (RN) hours per patient day than did OSHPD. RN proportion among licensed nurses matched most closely in the two databases. RN hours per patient day derived from both databases showed significant inverse relationships with decubitus ulcer and mortality, and the association was stronger for the measure based on the OSHPD data. RN proportion derived from the OSHPD data was significantly associated with all three patient outcomes, but the AHA measure had a significant relationship only with decubitus ulcer. Conclusions: Compared with the AHA survey, the OSHPD data on hospital nurse staffing appear to be more complete and also were more closely associated with patient outcomes. Efforts to refine the AHA survey as a national database for nurse staffing will significantly enhance the capacity for monitoring nurse workforce and its effect on quality of care.