Back

Maternal and Infant Research Electronic Data Analysis (MIREDA): A protocol for creating a common data model for federated analysis of UK birth cohorts and the life course.

Seaborne, M.; Jones, H.; Durbaba, S.; Sanchez-Soriano, C.; Orton, C.; Ford, D.; Hough, A.; Mason, D.; Mendez-Villalon, A.; Giles, T.; Santorelli, G.; Nirantharakumar, K.; Quinlan, P.; Brophy, S.; Poston, L.; Reynolds, R.; Cockburn, N.; Gonzalez-Izquierdo, A.

2024-04-08 public and global health
10.1101/2024.04.08.24305489 medRxiv
Show abstract

IntroductionBirth cohorts are valuable resources for studying early life, the determinants of health, disease, and development. They are essential for studying life course. Electronic cohorts are live, dynamic longitudinal cohorts using anonymised, routinely collected data. There is no selection bias through direct recruitment, but they are limited to health and administrative system data and may lack contextual information. The MIREDA (Maternal and Infant Research Electronic Data Analysis) partnership creates a UK-wide birth cohort by aligning existing electronic birth cohorts to have the same structure, content, and vocabularies, enabling UK-wide federated analyses. ObjectivesO_LICreate a core dynamic, live UK-wide electronic birth cohort with approximately 100,000 new births per year using a common data model (CDM). C_LIO_LIProvide data linkage and automation for long-term follow up of births from MuM-PreDiCT and the Born in initiatives of Bradford, Wales, Scotland, and South London for comparable analyses. C_LI MethodsWe will establish core data content and collate linkable data. Use a suite of extraction, transformation, and load (ETL) tools will be used to transform the data for each birth cohort into the CDM. Transformed datasets will remain within each cohorts trusted research environment (TRE). Metadata will be uploaded for the public to the Health Data Research (HDRUK) Innovation Gateway. We will develop a single online data access request for researchers. A cohort profile will be developed for researchers to reference the resource. EthicsEach cohort has approval from their TRE through compliance with their project application processes and information governance. DisseminationWe will engage with researchers in the field to promote our resource through partnership networking, publication, research collaborations, conferences, social media, and marketing communications strategies.

Matching journals

The top 1 journal accounts for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.