Beijing Tsinghua Changgeng Hospital and Beijing Electronic Digital Intelligence Technology Co., Ltd. (BDEI) announced a strategic cooperation on October 16, aiming to jointly develop the first specialized large model in the field of pharmacy in China. The project aims to optimize pharmacy workflows through AI technology and improve the efficiency and accuracy of medication safety assessments for special populations such as the elderly, children, and pregnant women.
From an industry perspective, the rapid update of drug information and continuous introduction of new drugs require pharmacists to spend a lot of time assessing risks due to individual differences and complex drug interactions. Traditional pharmacy service models relying on manual experience are no longer sufficient to meet modern clinical needs for medication safety, which has become a critical issue that healthcare institutions need to address.
Image source note: The image is AI-generated, and the image licensing service provider is Midjourney.
On the technical level, this pharmacy large model will be built on BDEI's "Spark·Medical Base" platform, combined with the clinical practice data and research resources of Tsinghua Changgeng Hospital. As an institution approved by the National Medical Products Administration for drug clinical trials, Tsinghua Changgeng Hospital has 30 professional clinical trial qualifications, covering key disease areas such as oncology, cardiovascular diseases, and neurology, and is equipped with experienced clinical research teams. These resources provide high-quality clinical data for model training.
From an application perspective, the pharmacy large model will be first implemented at Tsinghua Changgeng Hospital, establishing a feedback loop between technological iteration and clinical medication. Through verification and optimization in real clinical scenarios, the model will gradually enhance its understanding and decision support capabilities for complex medication situations, helping pharmacists quickly identify potential medication risks and provide more accurate medication plans for patients.
The collaboration between the two parties also involves medical data governance and infrastructure construction. Plans include promoting the construction of a trusted data space in the field of pharmacy, ensuring the high-quality supply and compliant circulation of medical data. Additionally, they will explore adaptation solutions with domestic AI chips and study flexible computing power deployment models suitable for the medical industry, which has practical significance for the autonomous control of medical AI applications.
From an industrial development perspective, both parties also plan to jointly participate in the formulation of industry standards and write related white papers, providing replicable technical solutions and implementation paths for the medical industry. This is of reference value for the standardized development and large-scale promotion of pharmacy AI applications.
It should be noted that the development of the pharmacy large model involves a complex medical knowledge system, clinical evidence, and regulatory requirements. The model needs to accurately understand the mechanism of drug action, principles of drug interactions, and physiological characteristics of special populations, while also complying with clinical medication guidelines and regulations. The effectiveness and safety of such vertical domain large models need to be strictly clinically validated and tested through long-term practice.
From a market competition perspective, there are already many companies that have entered the field of pharmacy AI applications, including intelligent prescription review systems and medication recommendation tools. The advantage of the cooperation between Tsinghua Changgeng Hospital and BDEI lies in the hospital's clinical resources and data accumulation. However, whether the large model can significantly improve the work efficiency of pharmacists and reduce medication errors in actual applications still needs to be quantitatively evaluated through clinical comparative studies. Moreover, how to balance AI-assisted decision-making with the professional judgment of pharmacists is also a problem that needs to be solved for the successful implementation of such systems.