Neural network cost estimates for heart bypass surgery

摘要: Abstract This paper reports on the results of using artificial neural network (ANN) technology to estimate treatment costs heart bypass patients based their diagnostic condition and clinical criteria. Our applications include: (1) predicting total episode cost data; (2) a method for providing rapid feedback assess change in within turbulent environment; (3) procedure identifying activity-based driver candidates that would normally not surface from an analysis accounting data. Clinical data were collected 250 at University Ottawa Heart Institute. The analysed support following conclusions: indicators obtained before surgery individual can be used surgery; average error decreases as we add information available during after surgical event; patient does require access records; (4) forecasting system describe may improve exception reporting by tracking criteria real-time basis throughout episode.