张江科学大数据创新实验室推进数据治理并探索实际研究应用   2026-09-22

 

The Innovation Laboratory for Big Scientific Data of Zhangjiang showcased how it uses historical medical data to support practical medical research and launched its first data products at the 2026 Zhangjiang Pharma Valley Conference and the BioShanghai Week 2026.

Basic structured cleansing alone isn't enough to make public medical data truly useful for industry, what's needed is business-level governance built around real-world use, said Zhang Jingnan, general manager of Shanghai Digital Industry Group, the Zhangjiang Group subsidiary that operates the lab.

Her team found the gaps first-hand: some fields had been truncated during data collection, leaving medical histories and treatment records incomplete.Problems like that rarely surface from behind a desk, Zhang told Yicai. Only by working against companies' actual business needs and running real analytical workflows could her team spot missing fields, inconsistent data standards, and other shortfalls -- a cycle the lab calls "using applications to drive governance."

After flagging the gaps to the district data bureau and health commission, which coordinated with medical institutions to re-upload the missing records, the lab ran three rounds of data correction following its Aug. 29 launch last year, before its first project cleared security and ethics review the following February.

Established in August last year, the lab has consolidated healthcare and related public data resources from Shanghai's Pudong New Area. On Sept. 18, it launched Dia Insight, a disease-specific dataset for real-world research, and PDCorpus, a corpus for training large language models.

In April, the lab signed a cooperation agreement with Novo Nordisk Shanghai Pharmaceutical Trading to conduct health economics research in metabolic diseases, including diabetes and obesity. On Sept. 18, it inked deals with institutions including the China Innovation Center of Roche and the Institute of Orthopedics at the Chinese People's Liberation Army General Hospital.

The lab has received more than 200 companies since its establishment, with some having completed data exploration, an assessment of whether existing data can meet specific R&D requirements before formal cooperation begins, Zhang noted.

On data security and ethics review, the lab manages data access through three zones -- green, yellow, and red -- with progressively stricter controls, Zhang said. The yellow zone uses simulated samples generated based on the structure of real data.

Projects requiring access to original data must pass data security reviews by the district data bureau and medical ethics review before entering the red zone, which only the lab's own designated staff may access, processing data in pairs and unable to remove it, she added. External personnel involved in the research are limited to the yellow zone.

The lab’s current data collaborations are limited to scientific research, Zhang pointed out.

The lab has access to about four billion healthcare data records covering 67 medical institutions and about six million permanent residents in Pudong New Area, according to Zhang. Although its data volume is not the largest nationwide, data authorities from other provinces and cities have already approached it to discuss potential cooperation.

Zhang believes that public data platforms need to balance functionality and security. “We need to ensure security, but we also need to make data more beneficial to companies,” she said. “If people are too afraid to use it because of security concerns, wouldn't that be a step backward?”

Source: Yicai Global

 

---