Large Models Are Reshaping Hospital Outpatient Services: Wuhan Union Hospital Partners with Baidu Health to Launch AI-Powered Smart Clinic

As China’s healthcare reforms continue to advance, the efficiency and quality of medical services have improved significantly; nevertheless, the strain on medical resources remains severe. Particularly in Grade A tertiary hospitals, persistent issues such as “long waiting times with brief consultations” and “instant sell-outs of appointment registrations” cast a shadow over patients’ healthcare experiences.

As service providers, doctors are also overwhelmed by the massive demand. Taking Wuhan Union Hospital as an example, this well-established Grade 3A hospital with a 158-year history receives nearly 7 million outpatient visits annually. To ensure patients receive satisfactory medical care while making the service process as “heartwarming” as possible, a large number of medical staff endure significant work pressure over extended periods.

To address the challenges faced by both patients and healthcare providers, Wuhan Union Hospital has partnered with Baidu Health to initiate a paradigm shift driven by intelligent technologies through the use of large language models. Recently, the two parties formally signed a strategic cooperation agreement to jointly build an “AI Smart Outpatient Clinic,” aiming to explore “new quality productive forces” in the medical field by introducing artificial intelligence technology.

Currently, the “AI Smart Outpatient Clinic” has been successfully piloted at Wuhan Union Hospital for over a month, with large language models deployed across more than ten clinical departments. Supported by diverse service modules such as intelligent triage and guidance, automated appointment slot addition, and smart waiting rooms, the “AI Smart Outpatient Clinic” not only resolves various patient pain points during registration, waiting, and consultation, but also frees physicians from inefficient doctor-patient communication, enabling them to maximize their professional value.

The significant achievements made in the early stages have laid a solid foundation for the next phase of AI advancement. Amidst the deep integration of technology and healthcare, innovative forces capable of transforming the medical system are emerging here.

Compared with traditional artificial intelligence technologies, the advantages of large language models in multimodal data processing enable them not only to integrate vast amounts of medical knowledge and provide robust support for complex clinical decision-making, but also to simulate near-human language understanding and communication capabilities, thereby assisting patients in addressing medical inquiries.

Leveraging these unique advantages, Wuhan Union Hospital and Baidu Health have targeted three high-volume scenarios in this collaboration—intelligent triage, intelligent appointment slot allocation, and intelligent waiting rooms—which require frequent doctor-patient communication or consume significant physician resources, thereby seizing the opportunity to reimagine the “Smart Outpatient Clinic.”

First,Intelligent Triage and ConsultationTargeting the common triage and referral needs of patients. In recent years, China has been striving to implement a tiered diagnosis and treatment system, encouraging patients to seek care at primary healthcare institutions and be referred stepwise to large tertiary Grade A hospitals. However, in practice, many patients are reluctant to follow this pathway; they often bypass primary care and go directly to large hospitals for medical attention due to minor abnormalities in simple test results.

Large tertiary hospitals, already burdened with patient overcrowding, are clearly reluctant to see this trend continue. After all, the scale of such behavior not only exacerbates the diagnostic and treatment pressures on major hospitals but also creates difficulties for patients seeking medical care. Large general hospitals have a highly specialized departmental structure; for instance, Wuhan Union Hospital has more than 40 specialties, each further divided into over 10 subspecialties. Relying solely on their limited health knowledge, ordinary patients find it challenging to quickly select the appropriate department and identify the right physician.

However, hospitals cannot refuse patients’ medical consultation requests that are “not very accurate.” The introduction of the intelligent triage application in this collaboration is precisely aimed at addressing the aforementioned issues.

Powered by large language models, AI can simulate the pre-consultation process, guiding patients to accurately describe their symptoms. Leveraging its reasoning capabilities to synthesize and summarize this information, the system precisely matches patients with the appropriate clinical departments and specialists suited to their conditions. This approach enables hospitals to maximize the utilization of effective medical resources, ensuring that every physician delivers their full professional value.

Furthermore, intelligent triage and referral systems can proactively recommend that patients self-order various diagnostic tests based on their clinical needs. Previously, these testing steps were embedded within the consultation workflow, requiring patients to make multiple trips for examinations and payments before receiving comprehensive diagnostic and treatment recommendations. With large language model (LLM)-enabled pre-consultation testing, patients can complete all necessary examinations during the waiting period, significantly enhancing both the consultation experience and efficiency.

Compared to intelligent triage,Smart PlusIts value lies in filling the "vacuum zones" within hospitals' past service systems.

Chen Shu, Director of the Outpatient Office at Wuhan Union Hospital, stated at the meeting, “Expert appointments at Wuhan Union Hospital can be booked online up to two weeks in advance. However, for high-demand departments such as thyroid and breast surgery, available slots are snapped up instantly as soon as the booking system opens. Many patients are unfamiliar with the online system and fail to secure on-site appointments, which may cause them to miss the optimal window for treatment.”

To further expand appointment slot availability and ensure that patients with urgent needs gain access to medical resources, Baidu Health has customized the “Smart Appointment Slot Addition” feature for Wuhan Union Hospital.

Zhang Yandong, General Manager of Medical Services at Baidu’s Healthcare Business Group, stated in an interview: “The value of ‘Smart Appointment Add-On’ lies in transforming what was once a unilateral request by patients into a mutual ‘agreement’ between doctors and patients.” Specifically, patients first interact with an AI model online and upload their test results. The model then extracts a summary of the medical history and key positive findings, helping physicians quickly determine whether a patient requires consultation with a specialist, thereby enabling precise appointment add-ons.

Through this approach, Wuhan Union Hospital has not only improved the efficiency of scheduling additional appointments, significantly reducing the time physicians spend reviewing records and communicating with patients, but also enabled more accurate identification of underlying causes, thereby enhancing the quality of diagnosis and treatment. For patients, those with genuine urgent needs for additional appointments can now have their needs met in a timelier manner.

Currently, the system is still in its pilot phase but has already shown initial effectiveness. According to Chen Shu, during the first three weeks of the pilot in the Department of Thyroid and Breast Surgery, the system received a total of 2,100 requests for additional appointment slots. After AI screening, the large language model retained approximately 280 applications, of which 268 were approved by expert review. As a result, a significant number of patients with malignant tumors requiring urgent surgery received timely treatment thanks to the introduction of the additional-appointment system.

FinalAI-Powered Smart WaitingThis represents a further optimization of consultation efficiency and the patient experience. Amidst overwhelming patient demand, many doctors compress consultation time to just five minutes to ensure that more patients can receive medical advice.

However, in practice, because physicians need to review medical records and analyze patients’ conditions during consultations, the actual consultation time that patients receive after long waits is far less than five minutes.

This is the key to Wuhan Union Hospital’s introduction of large language model (LLM)-based “digital humans” for AI-powered pre-consultation triage. With this tool, patients can interact with the digital human during their waiting time to provide their medical information and symptoms. The digital human then organizes this information to generate a text document resembling a final medical record and identifies key positive findings from the reports uploaded by the patient.

With AI support, doctors can quickly grasp a patient’s general condition by reviewing the organized medical record in just a few seconds during consultations. Ultimately, this not only saves time on inquiries and documentation but also makes doctor-patient communication more precise and efficient, leading to more accurate clinical decision-making.

According to Director Chen Shu, the feature has been piloted in more than ten departments at Wuhan Union Hospital, including Orthopedics, Gastroenterology, Thyroid and Breast Surgery, and Dermatology and Venereology, and has been widely recognized by physicians. In the future, Wuhan Union Hospital will work with Baidu Health to optimize the application, enabling its rapid deployment across additional departments.

While the development of smart services is undoubtedly important, in the ultimate vision of Hu Yu, President of Wuhan Union Hospital, fully leveraging the value of large language models requires their deep integration with the core elements of a smart hospital, thereby constructing a comprehensively intelligent healthcare system.

Therefore, building on the implementation of its AI-powered smart outpatient project, Wuhan Union Hospital has further established three core objectives, aiming to achieve a qualitative leap in medical services, patient experience, and hospital management.

First, Wuhan Union Hospital aims to further promote the deep integration of AI with healthcare scenarios, effectively advancing the development of smart healthcare. Specifically, by leveraging AI’s precise analytical and rapid processing capabilities, the hospital seeks to reduce inefficiencies and repetitive tasks in the consultation process, while assisting physicians in formulating more accurate diagnosis and treatment plans.

Secondly, Wuhan Union Hospital aims to optimize its smart services through AI, providing patients with a higher-quality experience. By implementing features such as intelligent triage, automated appointment scheduling, and smart waiting management, the hospital seeks to build an efficient and patient-centered medical service platform that enhances the healthcare experience and reduces waiting times.

Finally, in terms of smart management, Wuhan Union Hospital aims to gain a deeper understanding of patient needs and market dynamics through big data analysis and mining. By integrating these insights with the hospital’s actual operational status, it seeks to formulate more scientific management strategies, thereby comprehensively enhancing its management standards and operational efficiency.

For Baidu Health, its smart outpatient collaboration with Wuhan Union Hospital is also continuously enhancing the capabilities of its medical large language model. In an interview, Zhang Yandong stated, “Through our partnership with Wuhan Union Hospital, we have deepened our understanding of healthcare scenarios, and our vertical-specific models have been continually upgraded thanks to high-quality data. Looking ahead, we aim to further strengthen our cooperation with Wuhan Union Hospital and replicate this successful experience across more healthcare institutions, driving the medical industry toward greater intelligence and human-centric care. This will enable physicians to practice in a more relaxed environment and allow more patients to regain their health through more precise and compassionate medical services.”