Objective
To examine nursing’s impact on risk adjusted mortality for pediatric cardiac surgical patients across 38 free standing children’s hospitals in the USA.
Methods
A patient level analysis was conducted using secondary data from the PHIS and NACHRI datasets for the two year period of 2005-2006 and responses from Chief Nursing Officers to examine nursing indicators. Cases <18 years of age undergoing surgical repair of a congenital heart defect were identified using ICD-9-CM codes. Risk adjustment was performed using the RACHS-1 method. Analyses were performed using generalized estimating equations models to account for the intra-institutional correlation among cases from the same hospital.
Results
21,885 cases of congenital heart surgery were identified. There is variation in ICU nursing hours per patient day (15-32). Variation exists in ICU nursing skill mix (80-100%)and 20 hospitals had 100% registered nurse staffing in ICUs. The association between hours of nursing care and mortality (odds ratio =0.88, p=0.23)and between Magnet status and mortality(odds ratio=0.90, p=0.42)were not significant. ICU nursing worked hours tend to increase as hospital volume of congenital heart surgery increases. (Spearman rank correlation coefficient= 0.39 p=0.027). Hospital volume was significantly associated with risk-adjusted mortality. There was significant variation in mortality across hospitals (median unadjusted mortality rate 3.9%, range 1.6 to 15.8%). An odds ratio of 0.93 (p<0.001)was associated with each increase in volume of 100 cases.
Conclusion
Significant variation exists in nursing indicators and mortality across Children's Hospitals. Higher surgical volume centers have a reduced likelihood of death controlling for demographic and clinical variables.