Blogs & News
Aridhia and Sydney Children’s Hospital Network – Data Transformation to Improve Paediatric Care
Aridhia is delighted to have worked closely with the Sydney Children’s Hospital Network (SCHN) in a proof-of-concept (POC) effort. This POC was part of the Sydney Kids Learning Health Initiative (LHI), a partnership program between SCHN and the University of Sydney that aims to improve the care of children throughout the hospital system.
Who are SCHN?
SCHN in Australia is the largest provider of specialised health services for children and adolescents. Comprising two world-renowned hospitals, it is the largest paediatric healthcare service in Australia, made up of over 8,000 clinical and non-clinical staff across its network of hospitals and health services. SCHN offers a comprehensive range of services spanning from general paediatrics to highly specialised treatments. With a commitment to excellence in patient care, innovative research, and education, SCHN strives to improve the health outcomes and well-being of young patients across New South Wales and beyond. Through its multidisciplinary approach and state-of-the-art facilities, SCHN remains at the forefront of paediatric healthcare, ensuring that every child receives the highest quality of medical attention and support.
The POC
This POC aimed to develop a solution that can source data from the SCHN electronic medical record (EMR) system via an existing data source and develop the capability and services to transform and store this data in a trusted research environment with analytics capacity.
This POC was Phase 1 of the LHI, where SCHN and Aridhia built a small-scale infrastructure system (utilising the Aridhia DRE) to support the needs of a demonstration project that used patient data to improve the management of lower respiratory tract infections. It demonstrated that data could be extracted into a cloud-based repository, in a form suitable for analysis, while meeting all governance requirements, and provide real-world value to end users.
Data scientists and clinicians use the platform to explore the data, perform analysis, and generate reports and clinical dashboards. The current data infrastructure system at SCHN includes a mirrored eMR server and SQL data warehouse; however, the system still requires a high level of manual effort for extraction and dataset assembly that addresses the complexity of healthcare data. The LHI pipeline has built on this existing capacity by developing an automated process for extracting, transforming, and loading (ETL) the data from the existing SCHN database (which uses the Cerner Data dictionary) into the Observational Medical Outcomes Partnership (OMOP) common data model (CDM).
The OMOP CDM will allow for the systematic analysis of disparate observational databases. We have covered OMOP in a previous blog, where we describe working with the CDM, and what makes the Aridhia DRE the ideal environment to overcome some of the challenges that arise during transformation projects. Aridhia previously supported this transformation at Great Ormond Street Children’s Hospital.
Transformation of the SCHN data into the OMOP model was achieved via the Healthcare Landing Zone (HLZ). The HLZ is a customisable section of the DRE which provides a secure template for hosting healthcare data in the cloud. Controlled access is provided through a virtual private network (VPN), ensuring security and deployment management consistency. Data from the on-premises data lake is de-identified and pushed into the HLZ storage account daily. From here, it is moved into a suitable database before various OMOP tools are used (WhiteRabbit, Rabbit-In-a-Hat, etc.) to map and then create an ETL pipeline to output the data into an OMOP database.
Looking forward
The successful POC for this data pipeline involved the production of a clinical dashboard that monitors the management of admitted children with a lower respiratory tract infection. This involved transformation to the OMOP CDM which will be limited to the data elements required for the POC project.
Through the implementation of this automated OMOP transformation pipeline, the SCHN workspaces enabled researchers and data scientists to work with approved SCHN OMOP data, through managed inbound data that undergoes security scanning and requires approval from the workspace administrator before entry. Users can access the workspace remotely through a web browser with multi-factor authentication and can use a DRE workspace to query (in real-time) the data sitting in the remote ‘OMOP transformed’ database in the HLZ, to undertake their research and curate novel datasets for clinical dashboards.