Periodic Reporting for period 3 - MIDAS (Meaningful Integration of Data, Analytics and Services)
Reporting period: 2019-08-01 to 2020-02-29
WP1 provides the overarching administration, management, and coordination of the MIDAS consortium and project. It chairs and coordinates the MIDAS Policy Board. WP2 aims to deliver a solution addressing ethical and good practice challenges when working with large and varied datasets. WP2 will identify requirements, test systems to engage, inform, assure, and create a solution to maximise the benefit of data used for policy and the public. WP3 describes data sources prioritised by policy maker. Design and deploys the architecture and modules for integration of multi-site heterogeneous data sources. Data ingestion is developed to ensure data quality. WP4 will develop analytics platform for exploratory data analysis, health policy outcome simulation, and privacy-preserving machine learning models to support policy decision-makers in drafting and evaluating policy. WP4 also develops a framework for using systems approaches, thinking, and dynamics to supplement a data-driven approach to support political decision making. WP5 develops an interactive web-based tool to create decision support dashboards for understanding, finding explanations and making decisions from data. Three key tasks: design, implementation and user experience testing with user feedback gathered continuously inform each development iteration. WP6 leads an internal evaluation of the project, involving: identification of potential project impacts and the integration of the platform. This iterative evaluation process employs structured methods and processes. The interim evaluation results are fed back to partners to increase the likelihood of successful project outcomes. WP7 coordinates MIDAS dissemination, communication and exploitation activities, managing the MIDAS website and social media, organising and participating in events, and creating plans to ensure MIDAS results can be exploited by policymakers across Europe. WP7 focus on developing an integrated Exploitation Plan to allow the MIDAS Platform to evolve into a commercially sustainable solution in the longer term. WP8 ensures compliance with the ethics requirements of the project.
Work performed from the beginning of the project to the end of the period covered by the report and main results achieved so far
WP1 coordinate and manage the project, utilising the mechanisms from the first reporting period improved upon periodically. Issues and risks are monitored and resolved in conjunction with the Executive Board and wider consortium. WP1 organise and participate in external and internal events consisting of the MIDAS consortium and beyond. WP2 examined the context in which potential models meet the demands of legislative and good practice requirements for intended use of heterogeneous datasets within MIDAS. MIDAS MyData model has been evaluated, and a technical model established with a series of collaborative workshops with MIDAS policy board members and consortium stakeholders by acknowledging the four different MIDAS pilot aims. WP2 is exploring perceptions and testing an innovative model of engagement using technology, like chatbots and Honest Broker models as templates for Good Practice. WP3 researched, described and loaded data sources required for addressing each pilot’s requirements. MIDAS Architecture and Logic was deployed at all sites. Health data preparation GYDRA tool was developed with analysis-ready data delivery. MIDAS metadata model is core to iterative solution delivery and diverse dataset harmonisation. Synthetic data techniques were researched and tested for training models with good results. WP4 continued development of exploratory data analysis API and new crossfilter API analysis for several pilots. The MeSH classifier has been further developed. Data analysis has been applied and validated on two use cases of obesity and type two diabetes. WP5 develops the MIDAS Dashboard using feedback given from the initial prototype. The MIDAS UI was deployed at the pilot sites from where initial feedback was gathered. The second version of the MIDAS Dashboard UI was developed based on the feedback, and updates from version 1 to version 2 was initiated. WP6 has developed the final logic model and a protocol of realist evaluation. The Q method protocol was developed and the first round of the Q method data collection was completed. First round of interviews were undertaken, while second-round interviews are underway. WP7 set up the MIDAS website, dissemination templates, blog, newsletter, Twitter and YouTube Channel. The Policy Board was trained to develop Digital Stories to disseminate MIDAS, and seven Digital Stories have now been published by consortium partners. MIDAS has held several dissemination events and presented posters and papers in a wide and growing number of scientific conferences and journals. The Exploitation Plan V2 was developed to provide a basis for the sustainable MIDAS Open Source Foundation. WP8 via the established Ethics Board and independent ethics advisor continue to monitor the project and review and approve all activities that may have ethical concerns.
Progress beyond the state of the art and expected potential impact (including the socio-economic impact and the wider societal implications of the project so far)
WP1 work with the consortium towards progressing beyond state of the art by providing management structures and processes. WP2 is investigating the needs of the project within current legislative domains and creating an ethical, pragmatic and deliverable model of use and management. The fundamentals in access management and availability of personal data for policymaking with MyData approach needs a human-centred consent management where the purpose of the data use shall be openly presented for personal level, the concept communicated through the regional decision making, and in the policy board collaboration of activities that impacts EU level discussions (e.g. data donation & MyData.org workshops). The Honest Broker model, that utilises a central control and review system for healthcare-related data access for research purposes may offer transferable frameworks as is used in Northern Ireland and be adopted within MIDAS by the Basque region. WP3 has evolved Data preparation tool to Big Data and enabled together with a metadata catalogue the automatic delivery of analysis-ready data. Evaluated the utility of synthetic datasets for AI in the health context with diverse domain specific datasets. Defined a metadata-based data harmonisation approach. And published lessons learnt of QoD for Big Data. WP4 will develop scalable data analytics that incorporates privacy-preserving data mining tools and health policy outcome simulators exposed via an API interface used by the visualisation layer, and will be developed for analysing open data and social media. WP5 will develop dashboard generators that combine results from heterogeneous data sources and analytics from closed environments and the Internet. The system focuses on the visualisation of results from health data, news, MEDLINE abstracts and social media. WP6 has developed an innovative approach to evaluation, linking logic models and methods of framework analysis. A Q-methodology concourse was developed and deployed to succeed initial user interviews. WP7 continue leading dissemination, communication and exploitation efforts to maximise impact of MIDAS. An ambitious goal of WP7 is to create a partner-wide open-source foundation to support the ongoing exploitation and adoption of the MIDAS Platform as an innovative and integrated technical solution for healthcare providers across Europe. WP8 continues to function via the Ethics Board, overseeing and advising on all ethics issues throughout the project.