Capitalising on the explosion in the amount of data being collected today by healthcare professionals, a US researcher has developed a software program which automates the detection of risk factors.

E-commerce technique inspires new approach to sifting medical data

Healthcare professionals are increasingly being encouraged to computerise their patients’ medical files and to store online the data flows they gather. The Health Information Technology for Economic and Clinical Health (HITECH) Act, an integral part of the American Recovery & Reinvestment  Act (ARRA) which came into force under the Obama administration, contains provisions to encourage US physicians and hospitals to introduce Electronic Medical Record (EMR) systems by 2014. Now a researcher at the University of Notre Dame has developed an interface whose purpose is to exploit to the full these volumes of medical data by borrowing the approach of Big Data analytics in order to extend the field of diagnostics.

An interface offering personalized diagnostics

The Notre Dame system aims to provide a diagnostic approach which draws on similar cases, using sophisticated algorithms to extract meaning from the masses of data being collected on an ongoing basis. It identifies health risk factors by comparing information from one patient with other patient profiles in the database showing similar characteristics. The technique used by the platform is known as‘collaborative filtering’, an approach which is already being used by Amazon and Netflix as a basis for their recommendation systems. The CARE (Collaborative Assessment and Recommendation Engine) program thus enables personalized risk factor diagnostics by taking a hybrid approach based on a comparison of individuals coupled with data from the wider population.

Towards a patient-centric healthcare system

By using medical files which are all available on the same database, CARE has developed a preventive diagnostics system with many advantages. Relying on a more complete analysis than the standard population profiling approach, making comparisons with other similar cases, the doctor will be able to intervene earlier to encourage changes in the lifestyle of his/her patients and start up a dialogue which is likely to result in lower hospital re-admission rates and in better patient health in general. To facilitate ongoing patient monitoring, systems such as CARE could also be used in tandem with self-measurement devices, as is the case at several hospitals in Boston.


By Thomas Meyer
Journalist, Business Analyst