Feature - Health-e-Child

Anatomical heart model, with the colored patches representing heart segments, proposed by the American Heart Association. These divisions allow models to simulate localized abnormalities in diseased hearts, such as the focal thickening of the septum and local fiber disarray in hypertrophied cardiomyopathy. Things that look like ittle hairs represent the orientation of muscle fibers. Image courtesy Health-e-Child.

Last year, iSGTW reported on the initial deployment of Health-e-Child, a European project that fits grid technology to pediatric medicine. In this issue we provide an update.

New portal, new applications

Health-e-Child uses the Gateway, a grid-based platform developed by maat Gknowledge together with other European partners, to store, retrieve and manipulate medical records. One of the most successful applications running on this platform—winning a “Best Live Demo” award from the EGEE User Forum in Clermont-Ferrand, France—is the CaseReasoner application, which helps physicians diagnose patients with rare diseases.

“Some things are easy to diagnose, the common cold for example,” says Martin Huber of Siemens, technical coordinator for Health-e-Child. “Doctors see them often. Not so with the rare diseases Health-e-Child is concerned with.”

Such diseases, Huber explains, can be exceptionally difficult to diagnose, as they are so rarely seen that clinicians must rely solely on textbook knowledge. When a doctor has a patient with unusual or conflicting symptoms, CaseReasoner allows him or her to use the grid to search for similar conditions from anonymous patient records in any participating hospital. The doctor can compare recorded diagnoses, treatments and outcomes—helping prescribe treatment more confidently.

Another application creates individual-specific heart models for patients with enlarged right ventricles. Normally, cardiac modeling is a lengthy, manual process. Health-e-Child has made modeling near-automatic, saving time and giving more accurate measurements.

New sites

In the last year, eight European institutions joined Health-e Child, including three hospitals and five research facilities, universities and labs. The hospitals provide the project’s raw clinical data, while the other institutions develop the Health-e-Child applications and customize the gLite middleware to the project’s needs.

Their greatest challenge of the past year has been security. 

Areas of the brain most subject to variation in a population. Blue areas are the least likely to be variable, red the most. Understanding structural brain variation helps researchers and doctors detect and understand genetic or disease related ources of abnormality. Image courtesy of Health-e-Child

Patient security

“We are in 14 institutions and everyone handles security a little differently,” says Huber. “The idea of connecting the local networks to a grid makes the system administrators in hospitals particularly uneasy.”

Now that all partner institutions are connected and share the same standards, the project will soon start focusing on putting Health-e-Child to new and better uses. Researchers from the French research institute INRIA, for example, are developing neuro-oncological disease models. In time, their completed application will use information about a patient's brain tumor to give the physicians information about the tumor’s state and where it may be likely to spread.

“I am touched when I see the dedication of clinicians in the hospitals we work with,” says Huber. “If we can make their job a little easier, it will be the most rewarding thing I can think of. I know that if we can save them a little time, they will spend it with the patient.”

-Danielle Venton, iSGTW