By Topic

Cardiovascular risk stratification in decision support systems: A probabilistic approach. application to pHealth

Sign In

Cookies must be enabled to login.After enabling cookies , please use refresh or reload or ctrl+f5 on the browser for the login options.

Formats Non-Member Member
$31 $13
Learn how you can qualify for the best price for this item!
Become an IEEE Member or Subscribe to
IEEE Xplore for exclusive pricing!
close button

puzzle piece

IEEE membership options for an individual and IEEE Xplore subscriptions for an organization offer the most affordable access to essential journal articles, conference papers, standards, eBooks, and eLearning courses.

Learn more about:

IEEE membership

IEEE Xplore subscriptions

4 Author(s)
Atoui, H. ; Methodologies of Inf. Process. in Cardiology, INSERM, Lyon ; Fayn, J. ; Gueyffier, F. ; Rubel, P.

There is a growing demand for developing personalized and non-hospital based care systems to improve the management of cardiac care. The EPI- MEDICS project has designed a personal ECG monitor (PEM) capable of recording a simplified 4-electrode, professional quality 3-lead ECG, detecting arrhythmias and ischemia by means of committees of Artificial Neural Networks (ANN), and alerting the relevant health care professionals. Our objective is to improve the patient risk stratification and to reduce the number of false positive and false negative alarms by taking into account the demographic and clinical data featured by the user's electronic health record (EHR) stored in the PEM device. To design and assess such a new type of system, we adopted a decision making solution based on Bayesian networks (BN) that we trained to predict the risk of a cardiovascular event (infarction, stroke, or cardiovascular death) based on a set of demographic and clinical data (age, BMI, etc.) as provided by the INDANA database, from which we randomly extracted a training set of 15013 subjects and a testing set of 5004 subjects. The BN is then compared to an ANN committee (N=50) and a logistic regression (LR) model in terms of sensitivity, specificity, and area under the ROC curve (AUC). AUC = 0.80 for the BN, 0.75 for the ANN committee and 0.74 for the LR model. The Bayesian network approach achieved a high overall accuracy over both the neural network and logistic regression models on the testing set, and therefore can be useful in pHealth systems such as the PEM.

Published in:

Computers in Cardiology, 2006

Date of Conference:

17-20 Sept. 2006