LinkDoc Launches HUBBLE Medical Big Data Decision Support System to Empower Clinical Research and Precision Diagnosis
On April 26, at the “Medical Big Data Development and Application Conference” themed “Sharing, Interconnectivity, and Building a Community for Health and Medical Big Data,” Mr. Luo Ligang, CTO of LinkDoc Technology; President Wang Ping of Tianjin Medical University Cancer Institute and Hospital; and President Xu Ruiping of Anyang Cancer Hospital jointly launched the HUBBLE Medical Big Data Clinical Decision Support System.
China’s medical big data faces challenges such as a large volume of cases, unstructured data, lack of unified industry standards, and difficulties in patient follow-up, resulting in insufficient application in the healthcare sector; currently, only a small amount of data can be utilized for clinical research. Hospital administrators in China hope to leverage data to support management decision-making, experts and scholars seek data-driven support and motivation for scientific research, and clinicians are even more eager for data guidance when formulating diagnostic and treatment protocols and recommendations, thereby enhancing their confidence.
The HUBBLE system features three core services:
I. Supporting Management Decision-Making: HUBBLE intelligently “diagnoses” potential issues in hospital quality management for partner hospitals and departments through the Dean’s Dashboard and business reports. It visually presents findings via six major modules, including patient analysis, medical quality analysis, and operational efficiency analysis, thereby providing a data-driven basis for hospital management decisions;
II. Research Project Management: The HUBBLE research tool is fully aligned with clinical academic research design, incorporating methodologies and tools grounded in medical statistical thinking. It facilitates efficient research project design, study population definition, variable configuration, and statistical visualization based on structured data. Additionally, it includes built-in services for common analyses such as descriptive statistics, intergroup comparisons, and survival analysis.
III. AI-Assisted Diagnosis and Intelligent Imaging Diagnosis: LinkDoc aims to leverage vast amounts of clinical medical record data and imaging data, combined with precise sample annotations by medical experts, to enable machines to effectively learn expert knowledge through artificial intelligence technologies. This approach delivers intelligent assisted diagnosis and imaging diagnostic services, helping primary care physicians detect and confirm diseases while improving diagnostic and treatment efficiency.
Luo Ligang, Chief Technology Officer of LinkDoc, stated that the core operational mechanism of the HUBBLE Medical Big Data Clinical Decision Support System is built upon vast amounts of medical big data, while also incorporating the expertise of specialists across various disciplines. Technical teams leverage advanced IT technologies and deep learning algorithms to develop specialized solutions tailored to the field of oncology, thereby providing physicians with visualized, scenario-based, and intelligent system solutions. Furthermore, feedback from physicians during clinical use continuously optimizes the system, enhancing its accuracy.
Luo Ligang mentioned that LinkDoc has been building real-world oncology databases for hospitals over the past two years. This database encompasses LinkDoc’s two core services: one is the structuring of medical record data, and the other is telephone follow-up services. In terms of medical record data structuring, LinkDoc employs its DRESS engine, which enables efficient manual structuring and achieves a processing efficiency seven times higher than that of traditional ETC structuring systems.
On the other hand, current patient medical records merely document the diagnostic and treatment processes, lacking feedback on therapeutic outcomes. Therefore, the second service offered by LinkDoc’s Real-World Oncology Database is telephone follow-up for hospitals, facilitating the management of post-discharge patients. Leveraging these two core services, LinkDoc can integrate post-discharge rehabilitation data with clinical data, thereby generating highly valuable structured data.
From left to right: Chang Tao, Product Director at LinkDoc; Luo Ligang, Chief Technology Officer at LinkDoc; and Wang Xiaozhe, Chief Architect at LinkDoc.
Wang Xiaozhe, Chief Architect at LinkDoc, pointed out that clinical big data currently faces two major challenges. First, data sources are highly heterogeneous. Unlike traditional internet data, clinical data encompasses not only routine laboratory tests, diagnostic examinations, and medical records, but also genetic testing results and health information from physical examinations. To unlock the value of these diverse data types, they must be aggregated and structured in an analyzable format. However, the diversity of data sources makes it impossible to address all issues with a single technical approach. Consequently, a comprehensive automated structuring engine must be constructed by integrating multiple methods—including manual structuring, template-matching-based automated structuring, and statistical models such as statistical learning and deep learning—to meet the standards required for structured systems in clinical practice.
Second, clinical practice imposes stringent requirements on computer-aided diagnosis. In medical data and diagnostic support scenarios, the interpretability of conclusions—specifically, the causal inference chain—is subject to rigorous scrutiny. Consequently, applications and product designs commonly employed in big data analytics that rely solely on correlation-based findings are entirely unsuitable for the specialized domain of medicine. Therefore, it is essential to adopt a standardized approach and leverage deep learning to construct an auxiliary diagnostic model. Such a model aims to minimize physicians’ workload while providing objective, impartial, third-party insights. This helps prevent misdiagnoses and missed diagnoses that may result from physician fatigue due to prolonged work hours or insufficient experience.
Chang Tao, Product Director at LinkDoc, stated that as a healthcare big data company, LinkDoc primarily provides clients with services for data structuring and data application layers.
In terms of data structuring, LinkDoc’s hospital services have undergone three generations of iteration. The current data structuring service system has reduced the time required for entering an entire medical record from the initial two hours, to 20 minutes, and now to seconds for database ingestion.
At the application level, LinkDoc provides hospitals with the HUBBLE Clinical Decision Support Platform, a service built upon LinkDoc’s structured data. This platform offers data analytics to assist administrators such as hospital presidents and department heads in making management decisions, enabling them to review analytical results related to medical quality and operational efficiency.
In the realm of clinical research, traditional workflows require physicians to manually select medical records from archives, organize and enter data into Excel spreadsheets, and then import them into statistical software to generate corresponding charts and graphs. Currently, HUBBLE enables the direct import of structured data into newly created research projects. Based on the researcher’s definition of study and control groups within a project, variables and calculation methods can be selected via simple clicks. During the statistical analysis phase, researchers no longer need to perform complex operations in separate statistical software; the HUBBLE system significantly enhances the efficiency of the entire data processing workflow for physicians.
In addition, LinkDoc is also exploring clinical decision support and imaging diagnosis with some partners. In the future, it hopes to implement these two directions in practical clinical applications, helping doctors reduce misdiagnosis and missed diagnosis, reduce work intensity, improve efficiency, and allow them to have more time to provide more precise services for patients.