摘要
Background: The vasoactive-ventilation-renal (VVR) score includes pulmonary and renal dysfunctions not previously addressed by the vasoactive inotrope score (VIS) and may be a better predictor of cardiac care unit (CCU) length of stay (LOS) in patients undergoing re-entry sternotomy (defined as no earlier than 30 days after previous sternotomy) for congenital heart disease (CHD). Methods: Patients undergoing re-entry sternotomy for CHD from August 1, 2009 to June 30, 2016 were studied retrospectively. A total of 96 patients undergoing 133 re-entry procedures were identified. VVR scores were calculated on CCU admission post-procedure (at 0 hour), 24-hour, and 48-hour after admission to the CCU. The response variable was CCU LOS.? Recursive partition analysis identified variables predicting LOS. Results: 133 re-entry sternotomies in 96 patients made up the samples of the database;11 samples were removed due to incomplete data or placement on ECMO. Of the initial 25 features, 5 were removed for near zero variance and 3 categorical features were removed for non-information. Covariance analysis did not demonstrate any significant correlation amongst the remaining features. Initial recursive tree regression using ANOVA, cross validation and conditional predictive p-value (cp) = 0.01 produced 3 trees. The tree with lowest cross validation error was selected. The resulting 2 split trees with ventilator days less than 20 days and VVR score at 48 hours greater than 23 identified three CCU LOS groups with mean CCU LOS of 77.6, 55.1, and 9.5 days. Conclusions: Recursive partition analysis identified ventilator days greater than 20 days and the sub-population VVR at 48 hours as predictive of CCU LOS in patients undergoing re-entry sternotomy for CHD.
Background: The vasoactive-ventilation-renal (VVR) score includes pulmonary and renal dysfunctions not previously addressed by the vasoactive inotrope score (VIS) and may be a better predictor of cardiac care unit (CCU) length of stay (LOS) in patients undergoing re-entry sternotomy (defined as no earlier than 30 days after previous sternotomy) for congenital heart disease (CHD). Methods: Patients undergoing re-entry sternotomy for CHD from August 1, 2009 to June 30, 2016 were studied retrospectively. A total of 96 patients undergoing 133 re-entry procedures were identified. VVR scores were calculated on CCU admission post-procedure (at 0 hour), 24-hour, and 48-hour after admission to the CCU. The response variable was CCU LOS.? Recursive partition analysis identified variables predicting LOS. Results: 133 re-entry sternotomies in 96 patients made up the samples of the database;11 samples were removed due to incomplete data or placement on ECMO. Of the initial 25 features, 5 were removed for near zero variance and 3 categorical features were removed for non-information. Covariance analysis did not demonstrate any significant correlation amongst the remaining features. Initial recursive tree regression using ANOVA, cross validation and conditional predictive p-value (cp) = 0.01 produced 3 trees. The tree with lowest cross validation error was selected. The resulting 2 split trees with ventilator days less than 20 days and VVR score at 48 hours greater than 23 identified three CCU LOS groups with mean CCU LOS of 77.6, 55.1, and 9.5 days. Conclusions: Recursive partition analysis identified ventilator days greater than 20 days and the sub-population VVR at 48 hours as predictive of CCU LOS in patients undergoing re-entry sternotomy for CHD.
