WP4 handles the analysis of data from the different digital devices and technologies used in IDEA-FAST. It involves quality checking of the collected device-specific data, assessment of natural variations in the measures, preprocessing, feature extraction and candidate biomarker generating and eventually correlating these with targets such as indicators of fatigue, sleep disturbance, ADL/HRQoL etc. Critical will be the interaction with WP3 and WP2, to provide clinical and device-use information in the shaping of the analytical solutions. In fact, the characterization of the type of signals, their statistical properties, the definition of specific context of data collection and their clinical relevance will all be necessary information to define the analytical problem to be solved by WP4. As first activity, a deep dive into the technical and algorithmic analytical solutions available from literature and form the direct experiences of the academic, SME and EFPIA members will be performed. Clinical relevance and links to the “gold standard” will be given by WP2, while Practical usability findings and limitations provided by WP3 will be taken into account. Where possible, suitable datasets to these initial hypothesis will be selected from the existing data sources and data from EFPIA partners and pre-processed for the analysis (e.g. selection of gold standards, selection of patients, synchronization, handling of missing and outlier values). Using these data and those obtained in the FS, WP4 will attempt to develop efficient algorithms to generate potential new digital endpoints from the devices that will be implemented in the CVS. To get to this point a series algorithms will be proposed and applied to profile the performance of the devices used in the FS, so to provide support to device selection for the CVS, as well as final reporting of device-specific results in Part B of the project. Eventually, in conjunction with WP7 and WP9, algorithms of the best device-specific digital endpoints will be integrated in a data analytics package for review by EMA/HTA/FDA. The activity of WP4 will rely on the support of WP5 data management system.
Objectives
O4.1: Identify candidate digital endpoints for fatigue and sleep disturbances
O4.2: Identify other candidate digital endpoints for other ADL/disability/HRQoL in IMID and NDD
O4.3: Develop automated detection/computation methods of digital endpoints & underlying patterns and metrics
O4.4: Evaluate the performance of the candidate device-related digital endpoints
O4.5: Prepare a final package of data analytics methods for review by EMA/HTA/FDA
Deliverables
D4.1: Definition of assessment protocol for device-specific digital endpoints
D4.2: Specifications for candidate digital endpoints for fatigue, sleep disturbances and other ADL and HRQoL measures in NDD and IMID
D4.3: Performance assessment of candidate device-specific digital endpoints data – part A
D4.4: Requirements specification data analytics software package
D4.5: Performance assessment of candidate device-specific digital endpoints data – part B D4.6: Software package digital endpoints data analytics
