Next steps

The work undertaken by the PaLM Mapping Project team has focused on mapping structured and semi-structured source coded data to standardised interoperable coded data to meet the pathology and laboratory medicine terminology mapping requirement. Essentially, this involves the replacement of legacy Read PBCL terminology with SNOMED CT PaLM. 

Next steps would be to expand the investigation into technologies that could standardise unstructured data and clinical narrative text as interoperable structured coded data. NHS England’s Data Science Team have already been investigating applying techniques such as named entity recognition to histopathology narratives to extract and tokenise key information in the patient test report. Their Advanced Analysis Components to Support SNOMED PaLM Mapping Project paper explains ‘extension beyond current requirements’ for the mapping next steps. Such techniques could potentially be integrated into the terminology mapping process to enhance the automated conversion of such data to corresponding SNOMED CT structured coding. 

Future work will involve investigating and applying new technologies emerging in data science and AI to national implementation of terminology mapping/ translation requirements, such as the pathology and laboratory medicine terminology mapping use-case. The goal is to further optimise automation of unstructured to structured data, and when coupled with governance and clinical assurance functionality, is predicted to further alleviate technical stress and release clinical capacity.

Further work to feed into the terminology mapping best practice and testing would be to include user research and interaction design expertise. This is flagged in the internal project brief as one of the limitations the PaLM Mapping Project team were unable to explore, and yet critical to the proposed user-centred design, interface design, and usability of a highly capable and functional terminology mapping tool.  

Please direct any enquiries about the pathology and laboratory medicine terminology mapping requirement to [email protected].