Web and Grid Services for Improving Ambient Intelligence Embedded in Pervasive, Personal ECG devices

There is an increasing demand for new eHealth solutions improving patients safety and reducing the overall cardiac mortality and morbidity by allowing the monitoring of the patient and of the citizen health status, wherever they could be, at home, at work, on travel. Decision support systems, cardiac event recorders and digital signal transmission systems were thus intensively investigated over the last decade. The EPI-MEDICS project has designed an intelligent, portable personal ECG monitor (PEM) that is able to record and to analyse a 3-lead SCP-ECG compliant electrocardiogram (ECG), and to synthesize the missing standard ECG leads by means of advanced neural network-based methods. In addition, the PEM device embeds decision-making methods taking into account the ECG analysis results, the patient risk factors and clinical data, and have the capability of generating different levels of alarms that are transmitted to the relevant health care providers by means of a standard Bluetooth-enabled, GSM/GPRS-compatible mobile phone. In this paper, we present strategies for the implementation and the deployment of patient/citizen specific neural network-based methods used for the synthesis and interpretation of the patients ECGs by means of an ad hoc infrastructure that allows neural learning on a large scale, thanks to Web services and grid computing technologies. The proposed solution is open, generic and adaptable to any standard SCP-ECG compliant portable electrocardiograph or PEM device from the EPI-MEDICS project. The aim is to facilitate the remote initialisation and adaptive update of the medical record and of the patient-specific configuration parameters of the intelligence embedded in the ECG device.