Wie formt KI das medizinische Ökosystem neu? Transnationale Kräfte und lokale Weisheit schaffen zusammen eine herausragende Antwort.

Due to the seriousness and complexity of the industry, healthcare has long been regarded as the toughest nut to crack in the application of emerging technologies. However, the rise of AI is now completely rewriting this situation. With a high enough "ceiling," healthcare is becoming one of the most promising core tracks for AI.

In the past few years, there have been continuous technological iterations in the field of "AI + healthcare," including AI-driven drug discovery, medical imaging diagnosis, and intelligent auxiliary decision-making for specific clinical diseases. Domestic medical institutions have successively integrated large models, and technology giants are also actively deploying in the AI healthcare track. The successive introduction of national policies on "artificial intelligence + healthcare" is also driving the intelligent transformation of the healthcare ecosystem.

In the transformation driven by both policies and the market, Roche, the globally leading Swiss pharmaceutical giant, is becoming an important promoter of the transformation of the healthcare ecosystem. In November 2025, at the annual China International Import Expo in Shanghai, in addition to launching more than 40 products in its full product matrix and diverse innovative solutions, Roche Pharmaceuticals China also demonstrated another side of itself - an active embracer of AI.

"In today's era of the AI technology wave and the growing needs of patients, we deeply understand that medical innovation is no longer a one-way technology output. Instead, it is a process where researchers, technology providers, and industry players jointly create value," said Dr. Li Bin, Vice President of Medical Affairs at Roche China.

The first step towards future healthcare starts with "liberating doctors."

AI is accelerating the reconstruction of the operating mode of the healthcare industry. Perhaps the best entry point for this transformation is to free up doctors' time.

In 2021, a survey report on physicians by the School of Social Sciences at Tsinghua University showed that physicians spend an average of about 7.77 hours on outpatient consultations and about 1.47 hours on scientific research per day. Meanwhile, nearly 70% of doctors believe that "the pressure of scientific research and publishing papers is high." Under the dual pressure of work and research, the professional burnout of doctors as "frontline healthcare workers" has remained at a relatively high level for a long time.

How to liberate doctors from heavy workloads has become a key proposition for improving the efficiency of the healthcare system.

The problem in scientific research often lies not in the lack of ideas, but in the lack of time and tools. The AI scientific research tool "Xiaoluo Zhiduoxing" launched at this press conference was born based on this insight.

On the one hand, with technological progress, the number of literatures and cases that doctors and medical students have to face each year is doubling. On the other hand, when dealing with difficult and complicated diseases or exploring cutting - edge theories, doctors also need to quickly access literature materials to support clinical decision - making. Based on generative technology and Roche's years of practice and methodological precipitation in the field of medical research, "Xiaoluo Zhiduoxing" can provide a one - stop solution for the scientific research workflow in areas such as literature interpretation, solution generation, and statistical charting. The ultimate goal is to liberate doctors from cumbersome information processing and achieve the ideal state of "both saving lives and making academic achievements."

From the current results, "Xiaoluo Zhiduoxing" has covered 90 cities and 700 top - tier hospitals across the country, helping more than 4,000 doctors implement more than 600 scientific research projects and saving more than 20,000 hours of research work in total.

"Artificial intelligence cannot replace doctors, but doctors who can use artificial intelligence will replace those who cannot," said Zhang Shaoting, CEO of SenseTime Healthcare. As a close partner of Roche, SenseTime Healthcare also participated in the development of "Xiaoluo Zhiduoxing." When time can be used more effectively, doctors' ways of thinking and interacting with patients will also be affected and changed.

This is also the future healthcare path mentioned by American scholar Eric Topol in his book Deep Medicine, that is, a more human - centered medicine supported by machines. In this process, the greatest opportunity for artificial intelligence is not to avoid mistakes, but to rebuild the long - standing and valuable connection and trust between doctors and patients. In the view of Professor Fan Jia, an academician of the Chinese Academy of Sciences and from Zhongshan Hospital Affiliated to Fudan University, this is the key reason why AI cannot replace human doctors. Only when doctors can be liberated from complicated tasks with the help of AI tools can they give patients more listening, understanding, and companionship, and truly return healthcare to a "people - centered" approach.

From cure to "intelligent" cure: How does AI reshape the underlying efficiency of the healthcare system?

Data barriers are often regarded as one of the biggest challenges for the implementation of smart healthcare. For example, among public hospitals, different hospitals' data collection frameworks and information management systems often have inconsistent standards and formats, resulting in obstacles to inter - hospital connectivity. In fact, on the patient side, especially for patients with difficult and complicated diseases, due to the complex diagnosis and treatment process, they often need to visit multiple medical institutions. The resulting medical records are like "information islands" scattered in different places, characterized by multiple sources, multiple modalities, and a low degree of structuring.

These fragmented data not only increase the burden on doctors in medical record tracing and data sorting but also hinder the improvement of the efficiency of precision medicine. Therefore, the intelligentization of clinical documents is often regarded as one of the most promising areas for breakthrough in AI healthcare. A recent research report on the current situation of AI healthcare released by Silicon Valley venture capital firm Menlo Ventures clearly pointed out that clinical documentation is one of the two "high - ROI" scenarios for AI investment in the healthcare system. In the United States alone, its market size has reached as high as $600 million.

In China, this demand is even more urgent. With a large number of patients and a complex healthcare system, from consultations to subsequent treatment plans, ward rounds, surgeries and other aspects, the complexity of medical record information is even greater, and the pressure on clinical doctors for data sorting and analysis is also heavier. To address this pain point, Roche's ecological partner, Zhongshan Hospital Affiliated to Fudan University, jointly developed an MDT (Multidisciplinary Team) intelligent agent with the Shanghai Artificial Intelligence Laboratory. This is another important AI tool that Roche shared at the Import Expo in addition to "Xiaoluo Zhiduoxing."

In the past, it took doctors' teams several hours to manually compile a complete patient medical history. But now, the MDT intelligent agent can not only intuitively generate an "overview of the patient journey" and trend charts of core indicators but also significantly shorten the medical record sorting process to within 20 minutes. Moreover, it can effectively avoid the omissions and deviations that are inevitable in manual sorting, providing a reliable information basis for the precise diagnosis and treatment of multidisciplinary expert teams.

"It's true that artificial intelligence is an assistant to doctors, but if it is only positioned as an 'assistant,' this role is too small," said Academician Fan Jia at the Import Expo site. He pointed out that AI tools should not only provide single - point assistance but should become a force that empowers the entire system and platform. "We need to continuously accumulate more complex and comprehensive data to enable artificial intelligence to continuously evolve through learning and analysis, forming the ability and tools to truly solve difficult and complex problems. This is where its value lies in the healthcare field."

It is worth noting that the MDT intelligent agent is not only for specific treatment scenarios. With continuous data accumulation and feedback optimization, the model can also continuously evolve and ultimately feed back into clinical and scientific research innovation, driving the healthcare system towards intelligence from the bottom up.

Starting from clinical practice, jointly building a new ecosystem of "AI healthcare + scientific research."

In a healthcare industry report at the beginning of the year, Cathie Wood, the founder of ARK Invest, boldly asserted that healthcare is currently the most undervalued area for AI application. In the industry, there are also many views that "the end - game of AI is healthcare."

Although a future where everyone can enjoy precise medical services is beautiful, the reality is far more complex than the vision. The insufficient supply and uneven distribution of high - quality medical resources are long - standing structural problems in the healthcare system. Data from the National Health Commission shows that in 2023, the number of tertiary hospitals in China accounted for only 10% of the total number of hospitals in the country, but they undertook 62% of the country's hospital outpatient visits. Meanwhile, the individualized, diversified, and multi - level medical needs are growing rapidly, and the dilemma of "difficult and expensive access to medical services" still exists.

To truly achieve healthcare equality, AI must be deeply integrated into front - line clinical application scenarios to understand the real pain points of doctors and the actual needs of patients. For Roche, which has participated in the Import Expo for the eighth consecutive year, the "spill - over effect" of the Import Expo not only lies in introducing new technologies and products to a broader market but also in integrating the advantages of all parties. Starting from clinical practice, Roche collaborates with hospitals, doctors, and technology partners to deeply integrate cutting - edge AI technologies with medical research insights.

After more than 30 years of in - depth development in the Chinese market, Roche has accumulated rich experience and resource integration capabilities in the field of smart healthcare. In addition to the MDT intelligent agent and the "Xiaoluo Zhiduoxing" AI scientific research solution unveiled at this year's Import Expo, Roche is also continuously building an open and co - creative medical innovation ecosystem through a series of enabling projects, such as the scientific research collaborative innovation platform "Zhixing Research Institute" and the "Digital Healthcare Incubator 2.0" that connects clinical insights with product R & D.

The essence of healthcare is the exploration of life. When AI is combined with healthcare, this exploration will become faster and more precise. As Academician Fan Jia said, "Human understanding of ourselves - including understanding of diseases and the human physiological structure - is currently less than 5%. The emergence of artificial intelligence will greatly accelerate the process of human understanding of the human body, diseases, and biology. Moreover, this speed is no longer measured in centuries or millennia but in months and days."

With AI as the wings and scientific research as the pen, Roche Pharmaceuticals China, in the role of an "ecosystem connector," is mobilizing the forces of multiple parties in industry, academia, research, and enterprises to promote medical innovation, continuously expanding the boundaries of life cognition, and trying to make the achievements of medical innovation benefit every patient.