New deliverable available: ‘Initial prototypes of specific components for all BigMedilytics pilots’

Big Data-driven strategies are the smartest way to unravel knowledge needed to improve the sustainability of current health systems across the European Union and achieve better patient outcomes. What are the most suitable core technologies and platforms to address the issues with the highest impact on the sector? What are the main challenges?

The goal of work package 1, led by DFKI, is to oversee the transfer of mature Big Data technologies into the 12 BigMedilytics pilots that cover the areas with the main impact on the sector. This transfer will happen in three cycles: initial prototypes based on pilot requirements, updated prototypes based on pilot internal validation, and final implementations based on pilot external validation.

The new deliverable, under the title Initial prototypes of specific components for all BigMedilytics pilots, presents a first overview of the different software prototypes and shows the different components, data, and challenges to overcome. The goal of this document is a first comparison between the different pilot developments.

The overview provides a first analysis of the pilots according to the application of natural language processing (NLP) and image analytics (IA) components, the integration of state-of-the-art machine learning (ML) algorithms for prediction scenarios, the integration of event detection components of healthcare-related real-time data streams, the processing of large structured and unstructured data sources, the multi-velocity processing of heterogeneous data streams, and the security and privacy of data access and processing.

Defining a Blueprint to describe the best practices

The final outcome of the work package 1 is the Big Data Healthcare Analytics Blueprint known as the BigMedilyticsBigMatrix, a mapping between the requirements and the technical components. This Blueprint will ensure that concepts presented in the pilots can be easily scaled and replicated in Europe. It will describe the best practices that need to be followed from both a technology and health policy perspective when rolling out big data solutions for the healthcare sector in Europe.

The matrix will be multidimensional, taking into account aspects of technologies, pilots/businesses, and communities. This matrix will also make a distinction between data sources that have different velocities, as for example real-time location data and electronic medical records for hospital workflow focused pilots on the Industrialization of Healthcare theme.