Presentation on Personalized Health Care: a data mining challenge in clinical decision support

On Friday, Jan 20th, Saskia van Loon will give a talk on Personalized Health Care.  Saskia performs a PhD project in the TU/e Information Systems group (supervisors: Anna Wilbik, Uzay Kaymak), in collaboration with Catharina Ziekenhuis Eindhoven (supervisors: Arjen-Kars Boer, Volkher Scharnhorst).

Where: TU/e, Paviljoen, K16

When: 12:30-13:30
Agenda Item: see http://is.ieis.tue.nl/?ai1ec_event=1457-2

Abstract: 

The focus of this PhD project (as part of Impuls 2) is on improving decision making by healthcare professionals through providing decision support based on fusing data from multiple sources, including patient self-monitoring information. The health information gathered in daily life circumstances complements information from professional diagnostic tools and can thereby contribute to clinical decision making. Philips Research has developed a prototype eWatch that is able to collect different hemodynamic parameters based on photoplethysmography. Within the PhD project, special attention will be given to laboratory data from the Catharina Hospital as many high- quality data are generated here and the data are also well documented. One part of the project is focused on cardiology, i.e. categorizing and treating heart failure as well as on improving the outcome of cardiac resynchronization therapy (CRT). Heart failure is now categorized in four stages based on subjective data. The categorization of heart failure is used in clinical decision making, where cardiac treatment eligibility is partially based on the stage of heart failure. Combining diagnostic patient data and monitoring cardiac functioning in daily life can objectify this classification of heart failure and subsequently improve outcome of different cardiac therapies. CRT is performed by implanting a device, i.e. an implantable cardioverter-defibrillator (ICD) or pacemaker, equipped with an additional lead stimulating the left ventricle. In 70% of the patients this will resynchronize the cardiac ventricular contractions and improve cardiac output. However, until now, 30% of patients does not respond for unknown reasons. Combining available patient data and monitoring of these patients in daily life circumstances provides information that can improve the success rate of this procedure. The second part of the project is focused on bariatric surgery. The number of patients with obesity and morbid obesity is increasing steadily. Sometimes a bariatric surgery such as gastric bypass or sleeve, is the only treatment option. Patients eligible for a bariatric surgery are screened by a multidisciplinary team of the obesity clinic of the Catharina Hospital. Here, decision support is needed as patient data is coming from multiple sources and success of the treatment is hard to predict beforehand. Predictive models are built based on the multiple source data, with special attention to laboratory data, modelling weight loss, improvement of comorbidities and quality of life.