MiADE at HealTaC conference 2024, Lancaster, UK
![miade-presentation-at-healtac-conference](https://www.ucl.ac.uk/health-informatics/sites/health_informatics/files/styles/small_image/public/healtac2024.jpg?itok=-PeziUKf)
The presentation shared the team’s experience and insights into integrating new software with a live health record system, preliminary findings from MiADE’s current feasibility trial, and future research plans supported by our EPSRC grant.
The talk sparked lively discussions on implementation at UCLH, challenges in widespread adoption, and the potential for new user interfaces for NLP-assisted record entry beyond existing EHR limitations.
The MiADE team also exchanged valuable knowledge with other researchers on using Large Language Models (LLMs) to extract complex, context-driven information from clinical records.
MiADE at HealTaC conference 2023, Manchester, UK
![core_team_at_healtac_conference](https://www.ucl.ac.uk/health-informatics/sites/health_informatics/files/styles/medium_image/public/miade_teamhealtac.png?itok=qx4oPoGI)
The demo session began with a brief introduction to the MiADE system and its integration at UCLH, followed by a live demonstration and code tutorial of the algorithm to extract problems and medications from clinical text, filter out irrelevant information, and structure metadata (Tutorial can be found at https://colab.research.google.com/github/uclh-criu/miade-tutorials/blob/master/notebooks/miade_tutorial.ipynb).
The team is currently planning a live implementation and evaluation study of MiADE integrated with the Epic electronic health record system at UCLH, and will be open-sourcing its packages and making them available for general use soon.
Presenting the MiADE journey at SNOMED CT Expo 2023, Atlanta (USA)
![l-zhu-presenting-at-snomed-event](https://www.ucl.ac.uk/health-informatics/sites/health_informatics/files/styles/medium_image/public/snomed-ct_atlanta2023.jpg?itok=HiVbunTB)
SNOMED CT is a common language for clinical terms, also serving as a clinical data standard both globally and within the NHS. UCLH was the first English-speaking site in the world to implement SNOMED-CT for diagnosis recording in Epic, a leading EHR system; as such it has always been considered as an exemplary implementation site by both SNOMED CT International and NHS England.
Leilei’s talk examined the implementation challenges within Epic encountered by the team and the demanding need to enhance the structured data input, while acknowledging that natural communication happens through free text. Leilei also reviewed the data standards landscape in Epic, highlighting how SNOMED CT serves as one of the building blocks to enable the success of the tool.
The talk drew great interest from a global audience, with the session recording a very high number of attendees.
NHS England stated, “If you want to hear about the true realities of implementing SNOMED CT you can't do much better than listening to Leilei Zhu sharing her UCLH experiences with SNOMED CT and EPIC.”
SNOMED International’s take on the project was, “Real-world implementation of SNOMED CT, with real-world problems and practical solutions that support clinicians, based on the experience from UCLH. Amazing overview, including the use of NLP to reduce the clinician burden of data entry. Exciting developments, with lots of progress in the project, with plenty still to do.”
Discussions with delegates at the event confirmed that an EHR system may not always achieve what the original vision set out to do. We are currently tackling a global digital health implementation issue. Data standards such as SNOMED CT, as an enabler, can be leveraged in many ways to address problems. Implementors should think innovatively about the potential different implementation approaches and not restrict themselves to one method in the wider landscape.