110 Adopting reproducible analytical pipelines and open analytics at GOSH. (23rd February 2023)

• Record Type: Journal Article
• Authors: Bryant, William A; Booth, John; Bowyer, Stuart A; Briggs, Lydia; Key, Daniel; Zenonos, Alexandros; Spiridou, Anastassia; Sebire, Neil J
• Is Part Of: Archives of disease in childhood. Volume 108(2023)Supplement 1
• Page Start: A41
• Page End: A41
• Publication Date: 2023-02-23
• Subjects: Children -- Diseases -- Periodicals; Infants -- Diseases -- Periodicals; 618.920005
• DOI: 10.1136/archdischild-2023-gosh.110
• Languages: English
• ISSNs: 0003-9888

Abstract

Abstract : Here at GOSH we aspire to be world leading in data and digital innovation in healthcare. The Digital Research Environment Team (DRE), part of DRIVE, have worked to-date on research support and data engineering (clinical dataset availability) which has put us in a position to develop operational and deployable solutions in the Trust based on our research. In addition we are part of the wider community of NHS data professionals who gain through collaboration and open working across Trusts. We aim to create a formal advanced analytics function at GOSH adopting Reproducible Analytics Pipelines (RAP) and open working to facilitate the adoption of data-driven solutions for clinical, operational, planning and strategic purposes in the Trust. We will look to develop and share workflows, documentation, best practices, challenges and solutions with the rest of the NHS – and use work shared by others. We will adopt RAP and other recommendations in the Goldacre Review to accelerate efficient and reproducible working. We will work through governance/ICT requirements in collaboration with relevant teams in the Trust (working closely with the Innovation team in DRIVE) to establish a formal, governance-and security-compliant remit and systematic protocols for doing this work. We will continue to work across analytics, Information Services, ICT, IG and clinical/operational teams to work out best practices to support the full lifecycle of analytics projects. We have established a roadmap of objectives to move towards RAP and low friction cross-team working. We have developed and agreed principles of open coding, open documentation and dummy data generation from clinical datasets. We are running a number of pilot projects to test/develop each element of RAP to show the benefits of this approach within our data-driven Research Hospital.