Skip to main content

Big Data 4 Better Hearts - Sofia ref.: 116074

Periodic Reporting for period 5 - BigData Heart (Big Data 4 Better Hearts - Sofia ref.: 116074)

Reporting period: 2021-03-01 to 2022-02-28

Launched in March 2017, BigData@Heart is a five-year project consisting of patient networks, learned societies, SMEs, pharmaceutical companies and academia. The project is dealing with a changing healthcare landscape, where the sustainability and quality of healthcare provision in Europe are being challenged. Demographic change and rapid innovation are leading to inconsistent medical care across Europe. Despite remarkable progress in the management of the most common cardiovascular diseases (CVDs) in Europe today, namely, acute coronary syndrome (ACS), atrial fibrillation (AF), and heart failure (HF), their disease burden remains high. Optimal management of these conditions is complicated by their complex etiologies, poor definition at the molecular level, and the added burden of co-/multi-morbidities. This leads to unpredictable and large variations in interindividual therapeutic response, heterogeneous prognoses, and treatment guidelines that are based on conventional risk factors and clinical markers of end-organ damage. These barriers pose major problems in the development and delivery of targeted CVD treatments. The aim of BigData@Heart is to apply big data approaches to ACS, AF and HF in the hope to improve patient outcomes. Bringing together key stakeholders in the field of CVDs, BigData@Heart’s ultimate goal is to address the challenges outlined above by developing a big data-driven translational research platform of unparalleled scale and phenotypic resolution. The research platform will deliver clinically relevant disease phenotypes, scalable insights from real-world evidence driving drug development and personalized medicine through advanced analytics.
WP1: Project management
The project governance structure is fully operational and progress and risks are being monitored with relevant tools. An Editorial Board was set up to coordinate the papers planned by the consortium to strengthen collaborations between (public and industry) partners and ultimately plan communication activities for these publications in collaboration with WP6. Currently, 61 peer-reviewed articles have been published.
WP2: Disease understanding and outcomes definition
WP2 is working on research on data-driven phenotypes and further approval for the use of the larger CALIBER and ABUCASIS datasets has been granted. As a result of two stakeholder meetings, a checklist for EHR studies has been suggested and will be published soon. A globally agreed AF Standard Set of Outcomes and electronic health record codes for all variables have been incorporated. It will enable measurement and comparison of important outcomes in a consistent manner with other countries. As highlight of last year a paper on "Identification and Mapping Real-World Data Sources for Heart Failure, Acute Coronary Syndrome, and Atrial Fibrillation" has been published (https://doi.org/10.1159/000520674)
WP3: Data sources (Mapping, selecting and curating existing data)
WP3 has completed the search strategy and literature review of contemporary data sources. Findings will be published in a research paper. The BD@H community on the EMIF platform has been created as an open access catalogue of the consortium. The fingerprint template has been completed for multiple datasets and published for CALIBER. The real-world data sources for HF, ACS, and AF were added to the EMIF catalogue: emif-catalogue.eu. Multiple mapping workshops with data owners have taken place. CALIBER and ABUCASIS datasets have been mapped to the OMOP data model. The RADAR-base platform is used to study patients with AF. The Uni of Birmingham is conducting a year-long study as part of the follow-up phase of the Rate control Therapy Evaluation in permanent AF (RATE-AF) trial (radar-base.org). The EHR phenotypes are curated in the open-access CALIBER Portal and have been used by 40 international research groups in 61 publications (doi.org/10.1093/jamia/ocz105).
WP4: Enrichment with Omics
WP4 has been working on the systematic evaluation of principles and opportunities of data enrichment strategies. WP4 has initiated a consortium-wide HF phenotype group to deliver validated definitions of disease outcomes and covariates relevant to HF across the consortium. An enrichment project has been conducted in collaboration with the industry partner SomaLogic and the third party Erasmus University of Rotterdam. A paper has been published and several have been submitted arising from this collaboration.
WP5: Data analysis
Working groups have been established between different partners within WP5, as well as close links with other WPs and case studies to maximize synergy and resources and ensure coherent output at the end of the project. Industry data sets have been shared with the consortium and have been used for analysis in a case study. Several work package overarching case studies are in progress, some with shared data from the industry. A prospective data set from Novartis is currently being negotiated to be shared with the academic partners of the consortium.
WP6: Communications of results and guidance documents (dissemination and exploitation)
WP6 has been working on raising further awareness for the project and increasing the dissemination of outputs. The project results were disseminated at several conferences, publishing external newsletters, social media presence on Twitter and interviews with key scientific work package leads in the consortium. Webinars were organised to communicate information and the results of the projects.
WP7: Ethics legal and data privacy (Governance, ethical and legal aspects)
An agreement has been reached by the consortium on the final decision-making procedure regarding data access and sharing and via a webinar the related report and outcomes have been disseminated.
It is the first time that consented cohorts, electronic health records in population settings, disease quality improvement registries, trial data, and clinically recorded imaging data will be studied together to identify mismatches and deliver novel disease vocabulary and outcome definitions in the cardiovascular realm in Europe. This new vocabulary should assist the development of new medications, interventions, and targeted management recommendations that improve patient outcomes. BigData@Heart’s ambition is to translate these new findings into universal definitions for ACS, AF, and HF. This will impact clinical trial design and contribute in the transition of economically feasible personalized medicine. More specific, the expected impact of BigData@Heart on science, industry, policies and the patient population includes: 1) definitions of disease and outcome that are universal, computable, and relevant for patients, clinicians, industry and regulators, 2) informatics platforms that link, visualise and harmonise data sources of varying types, completeness and structure, 3) data science techniques to develop new definitions of disease, identify new phenotypes, and construct personalized predictive models, and 4) guidelines that allow for cross-border usage of big data sources acknowledging ethical and legal constraints as well as data security. Achieving this impact is driven by the structured approach that safeguards the involvement of all key stakeholders and the wide dissemination and swift implementation of project results.
bd-h-new-logo.png