This document conveys the results of the On-Site Customer Analytics and Reporting (OSCAR) project. This nine-month project started on January 2014 and was conducted at Philips Research in the Chronic Disease Management group as part of the H2H Analytics Project. Philips has access to telehealth data from their Philips Motiva tele-monitoring and other services. Previous projects within Philips Re-search provided a data warehouse for Motiva data and a proof-of-concept (DACTyL) solution that demonstrated the linking of hospital and Motiva data and subsequent reporting. Severe limitations with the DACTyL solution resulted in the initiation of OSCAR. A very important one was the unwillingness of hospitals to share personal patient data outside their premises due to stringent privacy policies, while at the same time patient personal data is required in order to link the hospital data with the Motiva data. Equally important is the fact that DACTyL considered the use of only Motiva as a telehealth source and only a single input interface for the hospitals. OSCAR was initiated to propose a suitable architecture and develop a prototype solution, in contrast to the proof-of-concept DACTyL, with the twofold aim to overcome the limitations of DACTyL in order to be deployed in a real-life hospital environment and to expand the scope to an extensible solution that can be used in the future for multiple telehealth services and multiple hospital environments. In the course of the project, a software solution was designed and consequently deployed in the form of a virtual machine. The solution implements a data warehouse that links and hosts the collected hospital and telehealth data. Hospital data are collected with the use of a modular service oriented data collection component by exposing web services described in WSDL that accept configurable XML data messages. ETL processes propagate the data, link, and load it on the OS-CAR data warehouse. Automated reporting is achieved using dash-boards that provide insight into the data stored in the data warehouse. Furthermore, the linked data is available for export to Philips Re-search in de-identified format.
|Award date||1 Oct 2014|
|Place of Publication||Eindhoven|
|Publication status||Published - 1 Oct 2014|