Seoul to hook digital health startups up with hospital data
The city government of Seoul in South Korea has partnered with the Ministry of Health and Welfare to assist health and medical startups in harnessing medical and health data from hospitals.
Based on MOHW’s announcement, they will choose seven startups to help connect with 43 hospitals currently providing access to de-identified, anonymised data for R&D of new digital technologies, including AI-driven solutions.
Up to 50 million won (around $35,000) will also be provided as business funding alongside expert consultation in medical data analysis and research conduct.
Hong Kong’s eHealth app adds pathology report access
Pathology reports are now available for viewing on Hong Kong’s eHealth mobile health app.
According to the Health Bureau, users can now view these reports on the Investigations function 14 days after their release by the Hospital Authority, Department of Health, and private healthcare providers.
As of February, all public health settings and 40 private providers in nearly 100 sites are uploading pathology reports to Hong Kong citizens’ eHealth accounts.
Since it was launched in 2021, eHealth now provides access to nine EHR types, with more coming, including radiology images, Chinese medicine prescription records, and dental check-up records and conditions.
The eHealth app expansion is part of actions to deliver the five-year eHealth+ plan, which began in 2023 and sought to integrate healthcare data sharing, service delivery, and care journey management into the mobile health app.
National Cancer Center formalises pathology data using NLP
South Korea’s National Cancer Center has demonstrated the use of natural language processing models in extracting key information from free-entry, semi-structured breast pathology reports.
In a media release, the centre said its research team fine-tuned existing NLP models developed by Google (BERT-basic, BioBERT, ClinicalBERT) using 1,215 breast cancer pathology reports. The models later demonstrated a high accuracy of 96% and above in formalising these reports.
According to the researchers, hospitals were challenged by the tedious work involved in converting volumes of unstructured pathology reports into a structured format, which is essential for cancer research. This was partly due to previous government restrictions on medical data format conversion. It was only recently that some leniency was allowed; free-entry medical data can now be converted into a structured format using NLP.
link