North Thames GMC

Great Ormond Street Hospital (North Thames GMC), Jon Reed, Programme Manager, jon.reed@gosh.nhs.uk

Start: 01/09/2017, End: 01/09/2019

Context: 

The North Thames NHS Genomic Medicine Centre (NTGMC) has been set up to help gain a better understanding of the genetic causes of cancer and rare disease. It is lead by Great Ormond Street Hospital and includes Barts, London North West, Moorfields, RNOH, UCLH and the Royal Free.

Challenge:
NTGMC were tasked with integrating complex data across multiple sites, and drawn from 100’s of different systems into a common model. As part of the project they wanted to create a lasting legacy and shared repository for laboratory test results and pathology information. However, the data sources were difficult to integrate and interrogate because as they did not use a shared standard and the metadata about the local testing formats was not easy to access. Every single hospital coded their clinical tests differently e.g. Albumin was called 3398 in one site, CO310 in the other and ALB etc. in another.

How we helped:
We worked with local teams in each hospital to document local pathology formats and exposrts and loaded the reference data into a machine and human readable format in the metadata exchange. We then ran our matching algorithms to produce “suggest mappings” against the machine readable reference data between the different codes and metadata associated with those codes i.e. 3398=CO310=ALB etc. These mapping were then given to clinical teams to validate through our central repository and via a number of reports. Once the mappings were validated we allowed the ETL to consume the mappings within their data processing pipeline and feed the data into an Open EHR repository. The mappings and the reference data is now available for future use and the over a million laboratory results have been integrated into a common format.

Transferable Lessons:

North Thames did not have a shared repository for metadata. They used a number of different spreadsheets to document the information but they weren’t accessible centrally and they were not stored in a machine-readable format which was a problem for system integrators. By storing the data specifications and reference data in a central repository, it could be leveraged to automate and increase the efficiency of mapping different reference data across sites, allowed quicker clinical review and provided a central source of truth for multiple organisation to work to.