A novel software platform to facilitate analyses of linked health and administrative data Mark D ATKINSON a,1 , Daniel S THAYER a , Jonathan I KENNEDY a , Rebecca A HILL a , Simon THOMPSON a , Ronan A LYONS a , David V FORD a and Sinead T BROPHY a a College of Medicine, Swansea University, Swansea, UNITED KINGDOM Abstract. Due to the high complexity of routinely collected data, it can be difficult to realise its full potential for assisting with public health and epidemiological studies. A framework for a novel software platform is described herein which addresses issues in data complexity in a large warehouse of linked health and administrative data. The framework consists of three main parts relating to; data preparation, data automation and customization, and collaborative analyses. This has the potential to advance the field of data linkage research and aid in international research centre collaboration. Keywords. Data linkage, data cleaning, data interpretation Introduction The Secure Anonymised Information Linkage (SAIL) Databank in Wales collects health and other administrative data, from a wide variety of sources at a national and regional level to address research questions in public health and epidemiology [1]. There are numerous datasets within SAIL, however the majority will have comparable data issues. These errors and inconsistencies in the data include duplicate records, multiple codes to express the same condition, and measures being recorded in different units . To be able to undertake successful analysis and ensure quality output, these data problems need to be overcome utilising in-depth knowledge of each dataset. Herein, we present a novel framework for a software platform to facilitate analyses of linked health and administrative data which allows some aspects of data preparation and analysis to be automated and others to be customised where appropriate. 1. Methods The framework will consist of three parts; 1. Data layer – Cleaned, analysis-ready versions of all core datasets. 1 Corresponding Author. M.atkinson@swansea.ac.uk