Background
Despite the small but promising body of evidence for cardiac recovery in patients that have received ventricular assist device (VAD) support, the criteria for identifying and selecting candidates who might be weaned from VAD support have not been established.
Methods
A clinical decision support system (CDSS) was developed based on a Bayesian Belief Network that combined expert knowledge with multivariate analysis. Expert knowledge was derived from interviews of 11 members of the Artificial Heart Program at the University of Pittsburgh Medical Center. This was supplemented by retrospective clinical data from all VAD patients considered for weaning between 1996 and 2004 (n=19). Artificial Neural Networks and Natural Language Processing (NLP) were employed to mine these data and extract 28 most sensitive variables.
Results
Three decision support models were compared. The model, exclusively based on expert-derived knowledge, was the least accurate and most conservative. It under-estimated the incidence of heart recovery: incorrectly identifying 4 of the successfully weaned patients as transplant candidates. The model derived exclusively from clinical data performed better but mis-identified two patients: one who was successfully weaned, and one who ultimately needed a cardiac transplant. An expert-data hybrid model performed best, with 94.74% accuracy and 75.37% ~ 99.07% confidence interval, misidentifying only one patient who was weaned from support.
Conclusions
A CDSS may both facilitate and improve the identification of VAD patients who are candidates for cardiac recovery, and may benefit from device removal. It could be potentially used to translate success of active centers to those less established and thereby expand utilization of VAD therapy.