1. Hospitals across China are beginning to monetize patient records, turning clinical data into a lucrative commodity driven by a government mandate to unlock economic value from data, which is feeding the medical AI industry while testing privacy laws and forcing new compliance standards [para. 1][para. 2].
2. In April 2024, two landmark transactions highlighted the shift: Beijing Tongren Hospital sold an ophthalmology dataset to Bayer Healthcare and Jiangsu Hengrui Pharmaceuticals for an undisclosed sum; the First Affiliated Hospital of Shandong First Medical University sold a clinical liver-disease dataset to Shandong Shanke Zhixin Technology for 30,000 yuan ($4,399) — the first such deal in Shandong province [para. 3].
3. This rush follows a 2024 three-year plan by the National Data Administration and 16 other agencies to generate economic value from data across 12 sectors, explicitly targeting healthcare [para. 4].
4. New compliance pathways are taking shape: hospitals process raw data into packaged products, submit them to state-backed data exchanges for legal review and asset registration, and list them for buyers, aiming to create a controlled environment where data is "usable but not visible" to mitigate leak risks [para. 5].
5. Xing Lumin, the architect of Shandong’s April transaction, began salvaging dormant data in 2023, stating "Medical data is all treasure" [para. 6][para. 7]. He found that even foot-traffic logs could be sold to security companies [para. 8]. His breakthrough came when Shanke Zhixin struggled to train an AI model for liver disease; open-source data lacked edge cases, leading to high false-positive rates, which Xing helped resolve by connecting the company to the hospital's data [para. 9][para. 10].
6. Other deals include Shenzhen People’s Hospital selling a geriatric dataset in October 2023, and a county-level hospital in Fujian selling neurology, cardiology, and geriatrics data valued at over 450,000 yuan in January 2025 [para. 11].
7. Transaction values range from tens of thousands to millions of yuan; Beijing Tsinghua Changgung Hospital finalized a deal for yttrium-90 liver cancer treatment data with Grand Pharma Group (price undisclosed), described by director Li Dong as reflecting "the market value of high-quality specialty data" [para. 12].
8. Listings at data exchanges are proliferating: Shenzhen Maternity and Child Healthcare Hospital has multiple products on the Shenzhen Data Exchange; private firm Xi’an Youjun Medical Information listed over 70 datasets on the Guiyang Big Data Exchange in April [para. 13]. Demand comes from medical AI developers, pharmaceutical companies, medical-device makers, contract research organizations, and academic institutions [para. 14].
9. The most common technical safeguard for compliance is anonymization, stripping records of identifying markers, but global standards remain fragmented [para. 15][para. 16]. Chen Shuang, an attorney, noted that data is considered anonymized if it cannot be linked with other data to identify a patient [para. 16].
10. Beijing regulators released trial specifications for health-data anonymization late last year, creating China’s first local industry standard, and several Beijing hospitals have since completed transactions using these guidelines [para. 17]. However, Chen warned these are local recommendations, not mandatory national laws, so future compliance risks remain [para. 18].
11. A more valuable but legally complex alternative is "de-identification" under China’s Personal Information Protection Law, where data is processed so a specific person can’t be identified without additional information, e.g., replacing an ID number with a random code while retaining a master key for traceability in drug trials [para. 19][para. 20].
12. Unlike anonymized data (treated as public asset), de-identified data remains classed as personal information requiring informed consent, raising questions about securing renewed authorization from past patients; China lacks a unified standard [para. 21]. Hospitals must clarify consent boundaries, conduct retrospective ethics reviews for historical data, and draft new consent forms for future patients [para. 22].
13. Li Dong emphasized that ethics committee reviews "must not be a formality" but a substantive legal barrier, and the lack of national guidance continues to cloud the market with uncertainty [para. 23].
AI generated, for reference only
Sign in with Google
Sign in with Facebook
Sign in with 财新