The TRANSFoRm Project
The Learning Healthcare Project
Please watch Professor Brendan Delaney discussing the TRANSFoRm Project and the challenges it has encountered by clicking on the thumbnail.
The TRANSFoRm project aims to develop the technology to enable a rapid learning healthcare system that can improve patient care by speeding up translational research and enabling more cost effective participant recruitment and follow up in randomized control trials.
The TRANSFoRm project brings together a multidisciplinary consortium of 21 partner organisations from 10 EU member states. These include experts in ontology, integration, distributed systems, security, data mining, user-facing design, evaluation and clinical research. It is a 5-year project funded by the European Commission.
The technical challenge of achieving a rapid learning healthcare system is being explored through three clinical use cases in general practice. The project will develop the methods to support research on data held within electronic health records. Electronic health records are not currently designed for the easy collection of clinical data required to enable decision support tools.
The first use case focuses on single nucleotide polymorphisms and the associated responses to drugs used to treat type 2 diabetes mellitus using a combination of primary care databases and genomic databases. This use case is used to assess how routine healthcare records from multiple primary care data repositories can be used to address epidemiological (and genotype-phenotype) research questions.
The second use case examines symptom relief and quality of life associated with the treatment of Gastro-Oesophageal Reflux Disease with proton pump inhibitors using a randomised controlled trial design. Data was collected through the electronic health record and web based questionnaires to include patient related outcome measures.
The third use case was a diagnostic decision support system that was developed and integrated into existing commercial electronic health records.
This work has helped to identify key challenges in developing learning health systems as well as methods to address them. In particular:
• The information that is recorded in an EHR does not necessarily reflect what took place in the clinical episode
• Different use of codes and terminologies between systems do not map directly to each other
• Absence of disease, signs or symptoms are hardly ever recorded.
Some of these problems can be addressed through the use of ontologies and common data models that maintain meaning across systems. Archetypes, (which are used to link ‘meaning’) can also be used to overcome this problem.
Please visit the TRANSFoRm Project website for more information