Key result
Adding six variables to a core model fails to significantly improve short-term CABG mortality prediction.
Why the study?
It was unknown whether more comprehensive risk-adjustment models significantly impact hospital risk-adjusted mortality rates after CABG.
Do more comprehensive risk-adjustment models significantly impact hospital risk-adjusted mortality rates after CABG compared to simplified models?
Population
5,517 patients undergoing isolated CABG in Ontario in 1993
Comparison
12 increasingly comprehensive risk-adjustment models with 6 core vs core plus 6 level 1 variables
Design
Retrospective observational study across nine cardiac surgery hospitals
Authors
Loading...
Core variables stabilize CABG hospital mortality rankings; supports simplified risk-adjustment models in registry data.
Observational (n=5,517)
Yes
Do more comprehensive risk-adjustment models significantly impact hospital risk-adjusted mortality rates after CABG compared to simplified models?
Absolute Event Rate: 0.79% vs 0.77%
p-value: p=0.063
A small number of core variables are sufficient for fairly comparing risk-adjusted mortality rates after CABG across hospitals, allowing for simplified and efficient interprovider comparisons.
Tu et al. (1997) conducted an observational in Coronary artery bypass graft surgery (CABG) (n=5,517). Comprehensive risk-adjustment model (core + level 1 variables) vs. Core variables only model was evaluated on Area under the receiver operating characteristic (ROC) curve for short-term mortality (p=0.063). Adding six level 1 variables to a risk-adjustment model of six core variables only slightly improved the ROC curve area for predicting short-term mortality after CABG from 0.77 to 0.79 (p=0.063).