• Users of CMU medical system 25-hospital data

• Researchers using 5.37M medical records

• Researchers using medical imaging (X-ray/CT/MRI)

• Biomedical signal/gene/physiology data analysts

• AI/smart medicine researchers

• Patients benefiting from privacy protection

China Medical University Hospital, led by Vice Superintendent Chun-Liang Chen and Manager Han-Mi Chen, integrates the IT Office (data source), Big Data Center (clinical data de-identification), and AI Medical Diagnosis Center (medical imaging de-identification) into the "Hidden but Not Obscured" smart medical data management mechanism. The CMU medical system encompasses 25 hospitals with 5.37 million accumulated records covering text records, medical imaging (X-ray/CT/MRI), biomedical signals (ECG/EEG), bioinformatics (genes), and personal physiological/environmental data, spanning medical centers, regional, district hospitals, and clinics. Clinical data de-identification uses pseudonymization (ID hash as primary key), merger transformation (ordered encryption of visit/form numbers), fuzzification (dates to ages, addresses to towns), text report processing, and consistency transformation (offset by same value to preserve research value). Medical imaging de-identification handles DICOM tags by classification and masks/restores personal info in image margins. Implemented in 2020 conforming to ISO/IEC 29100, ISO/IEC 29191, CNS 29100-2—Taiwan-first and only medical institution to pass CNS 29100-2 accreditation. Established a one-way hashing mechanism for de-identified yet linkable data; published 40+ papers, obtained 12 patents (4 pending review), 1 US FDA approval (K212624), 4 FDA/TFDA medical software licenses.

• Hidden yet not obscured: de-id and linkable

• Only TW medical institution with CNS 29100-2 certification

• 25 hospitals, 5.37M records

• One-way hashing preserves linking

• Covers clinical data + medical imaging

• Fuzzification/consistency/merger transformations

• 25 hospitals in medical system

• 5.37M medical records

• 40+ journal papers

• 12 patents obtained, 4 under review

• 1 US FDA, 4 FDA/TFDA software licenses

• HIS/PACS/genetics/pathology/cancer database integration

• ISO/IEC 29100/29191/CNS 29100-2 certifications

• Taiwan-first & only CNS 29100-2 medical institution

• US FDA software certification K212624

• SOPs SOP-A80-B0501/SOP-JZ0-A0001

• Medical data application review procedure

• Patient-signed consent for special-purpose use

• Hash function protection measures

• Text content with PII processed by case

• DICOM tag classification

• Image margin PII masking

• Strict strong/weak/no-PII classification

• Adheres to ISO/IEC international PII standards

• User-friendly medical data

• Online application/dashboard convenience

• Patient PII protection recognized

• Trust enhanced by ISO certification

• 25 hospitals collaborate

• Transparent personal consent forms

• 40+ journal papers

• 12 patents (4 under review)

• 1 US FDA approval (K212624)

• 4 FDA/TFDA software licenses

• ISO/IEC 29100, 29191 intl certifications

• Taiwan-first & only medical institution with CNS 29100-2

• Balances personal privacy and medical research

• Leads future smart medicine development

• Builds linkable medical Digital Twin

• Integrates 25-hospital data

• Benchmark for domestic institutions

• Enhances medical data reuse value