The daily challenges that trouble clinicians, such as searching through extensive databases for literature, spending significant effort finding guidelines, and tediously organizing patient case information, now have a more efficient AI solution. Recently, JD Health's evidence-based medicine AI tool designed specifically for clinical doctors, JD Zhiyi, has received a significant comprehensive upgrade. The capabilities of JD Zhiyi have been fully integrated into the "JD Doctor" app, achieving all-around improvements in evidence-based professionalism, medical safety assurance, and practical feature usability. With more accessible, easier-to-use, and more reliable AI capabilities, it tangibly reduces the burden on doctors, making clinical diagnosis, treatment, and research work more effortless and efficient. (JD Zhiyi is fully integrated into the JD Doctor app; it can be experienced for free by downloading the app.)
This upgrade is built upon the technical foundation of the "Jing Yi Qian Xun" large medical model and has been meticulously refined by closely focusing on real-world work scenarios for doctors, including clinical diagnosis, case analysis, research learning, and patient management. It precisely addresses frequent pain points in a doctor's workflow, such as cumbersome literature searches, difficulty finding evidence, and the complexity of integrating patient information. The newly upgraded features are now officially available for all doctors to use; they can be experienced for free by downloading the JD Doctor app. JD Health will continue to use professional and reliable AI technology to enhance the quality and efficiency of clinical practice and medical research.
**Evidence-Based Capabilities Continuously Strengthened; Safety Evaluation System Sets New Industry Benchmark**
The core of this JD Zhiyi upgrade is to make evidence-based professionalism and medical safety more robust and trustworthy, ensuring that every suggestion provided by the AI is based on evidence, rigorous, and safe to use. In terms of underlying technology optimization, JD Zhiyi has undergone a comprehensive iteration of its self-developed evidence-enhanced large medical model. Building upon the original stable architecture, it further optimizes the AI's generation logic through multiple rounds of parameter fine-tuning, training on vast amounts of high-quality clinical data, and comparative learning across multiple model versions, fundamentally improving the accuracy and professionalism of AI responses.
Concurrently, the product has established a full-process safety protection system, forming a complete closed loop from data compliance review and multi-round validation of AI content to intelligent early warnings for high-risk scenarios, thereby minimizing the risks associated with medical AI applications. Incorporating feedback from frontline clinicians, JD Zhiyi has also customized more practical decision-making logic for different medical scenarios. For high-risk situations like complex consultations and medication plans, it strengthens cross-verification of multiple evidence sources and strictly adheres to expression boundaries. For high-frequency scenarios such as daily diagnosis and literature interpretation, it balances professional rigor with practical applicability, ensuring that AI outputs comply with medical standards while being directly suitable for clinical work.
To uphold the safety底线 of medical AI, JD Health's expert committee has developed a comprehensive four-dimensional Medscope evidence-based evaluation system for JD Zhiyi. This is accompanied by an authoritative evaluation set covering 46 major clinical departments, conducting thorough verification across four key dimensions: evidence quality, content accuracy, expression standardization, and overall practicality, ensuring that AI outputs are traceable, authoritative, and reliable. Additionally, a dedicated MedSafety evaluation system has been established, covering 26 types of high-risk medical situations, such as contraindicated medication use and complication warnings. If any safety red line is breached, an immediate one-vote veto is implemented, always prioritizing clinical medical safety.
Furthermore, this upgrade includes a dynamic update of globally authoritative medical databases, fully incorporating the latest editions of diagnosis and treatment guidelines, top-tier international research findings, and official drug information, ensuring that AI outputs consistently align with cutting-edge medical standards. Through rigorous multi-dimensional validation, JD Zhiyi's model hallucination rate continues to decrease; among vertical medical models, it ranks first in both evidence-based medical evaluation and specialized safety assessments.
**All-Scenario Functional Innovation and Upgrade, Deeply Adapted to the Entire Clinical Workflow**
Closely aligned with the real needs of doctors' daily work, this JD Zhiyi upgrade introduces several practical new features, making operations simpler and scenario coverage more comprehensive, thereby truly integrating the AI tool into the doctor's daily workflow.
Regarding data processing, JD Zhiyi has newly enhanced powerful multimodal capabilities, fully supporting uploads in various formats like photos, images, and PDFs. Various clinical materials such as lab reports, prescriptions, test sheets, medication boxes, and skin images can be quickly analyzed. Doctors no longer need to manually input data or page through documents; they can simply take a photo to quickly obtain AI analysis results, significantly simplifying the information organization process. The newly launched "Dynamic Evidence Localization" feature can automatically and accurately locate the original text statements supporting clinical viewpoints, making AI suggestions fully traceable and further enhancing evidence-based reliability.
In terms of assisting with precise patient diagnosis and treatment, JD Zhiyi has integrated access to tens of millions of patient medical records accumulated on the platform. Doctors can conduct evidence-based analysis directly on their own accumulated patients from the platform or independently upload data from offline outpatient visits. The AI will conduct a comprehensive analysis by combining the patient's full spectrum of information, including historical consultations, prescriptions, medical records, and doctor's notes, moving beyond single-question answers to provide personalized, precise recommendations that better fit the individual patient's situation, making diagnosis and treatment more targeted.
Addressing doctors' needs for research learning and knowledge updates, this upgrade also adds a personalized subscription and push notification function. Doctors can subscribe to the latest diagnosis and treatment guidelines and cutting-edge domestic and international medical information based on their specific practice department and research interests, receiving professional content promptly and easily staying abreast of the most advanced knowledge.
Since its initial launch, JD Zhiyi has consistently prioritized the needs of clinical doctors, evolving from basic feature construction to specialized optimization of evidence-based capabilities, and now to this comprehensive major iteration, continuously refining the product's professionalism and practicality. In the future, JD Health will continue to leverage the "Jing Yi Qian Xun" large medical model and its full-scenario AI medical matrix, constantly iterating on JD Zhiyi's technology and features, deepening adaptation to clinical scenarios, using cutting-edge AI technology to unleash doctors' professional value, and assisting the internet healthcare industry in developing towards a smarter, more professional, and more inclusive direction with high quality.
Comments