Real-World Data Linkage Research Platform

Overview

This study aims to address the lack of intelligent governance tools in clinical data management to promote efficient governance and secure sharing of real-world health data. To achieve this, a self-adaptive, automated governance intelligent agent will be developed based on a High-Order Programming (HOP) architecture, integrating Large Language Models (LLMs) and deep learning techniques. The agent will continuously monitor and correct data quality issues in real time, improving data accuracy and usability. In parallel, the project will establish a trusted data-sharing framework by integrating AI Confidential Computing (AICC) with Trusted Data Matrix (TDM) technologies. This framework will enable secure, real-time cross-institutional data exchange and collaborative computation while protecting sensitive information. Overall, the study aims to transform fragmented clinical data into high-quality, standardized, and securely accessible resources, thereby facilitating the circulation of data value and advancing collaborative medical research.

Detailed description

This multicenter, observational cohort study aims to integrate longitudinal health data from China, including routine health examinations, electronic medical records, and disease registries. The platform is designed to address key data challenges in the medical domain, particularly in chronic diseases and suboptimal health status. It is driven by two primary objectives:

1. Intelligent and automated data governance To ensure high data quality, the platform will engineer a self-adaptive, automated governance intelligent agent. Integrating Large Language Models (LLMs) and High-Order Programming (HOP), this agent actively monitors and corrects real-world data issues, such as missing values, redundancies, and formatting inconsistencies. Through deep learning, the agent continuously optimizes its governance rules to adapt to complex medical data environments. 2. Trusted and secure data sharing To facilitate multicenter collaborative research, the study will establish a secure and trusted data-sharing framework. By integrating AI confidential computation (AICC) with Trusted Data Matrix (TDM) technologies, the platform provides hardware-level security guarantees. This ensures that real-time, cross-institutional data exchange and collaborative computation without exposing sensitive patient information.

Overall Objective The platform aims to transform heterogeneous clinical data into standardized, high-quality, and securely accessible resources, thereby enabling efficient data utilization and promoting the value circulation of medical data for real-world evidence research.

Interventions

• Other This is an observational study. No intervention will be applied.
  This is an observational study. No intervention will be applied.

Primary outcome measures

• Accuracy Rate of Automated Data Governance [Time frame: 2026.5.30 to 2028.12.31]
• Completeness Rate of Automated Data Governance [Time frame: 2026.5.30 to 2028.12.31]

Secondary outcome measures (3)

• Correction Accuracy of Automated Data Governance [Time frame: 2026.5.30 to 2028.12.31]
• Data Standardization Rate of Automated Data Governance [Time frame: 2026.5.30 to 2028.12.31]
• Cross-institutional Data Usability of Automated Data Governance [Time frame: 2026.5.30 to 2028.12.31]

Eligibility criteria

Inclusion criteria

• Participants will be eligible for inclusion if they meet all of the following criteria:

• Availability of any health-related data generated from routine clinical care, health examinations, or disease surveillance systems, regardless of disease type or health status.
• Presence of at least one type of usable data, including but not limited to diagnostic information (structured or unstructured), laboratory results, imaging data, or basic demographic information.
• Records contain sufficient information (appropriately anonymized) to allow data organization and, where feasible, linkage at the individual level across time points or data sources.

Exclusion criteria

• Participants or records meeting any of the following criteria will be excluded:

• Records lacking minimal essential information required to distinguish individual records or support basic analysis (e.g., completely missing identifiers or time information).
• Records confirmed to be invalid, including system-generated test data, corrupted entries, or records that do not represent real clinical or health-related events.
• Exact duplicate records that cannot be resolved through standard data processing (only one record will be retained when duplicates are identifiable).

Criteria are shown verbatim from the registry (in English). Final eligibility is always assessed by the study center.

Healthy volunteers: Yes

Study design

Observational model

Cohort

Study locations

China · 1 center

• Beijing Friendship Hospital, Capital Medical University.No. 95, Yongan Road, Xicheng Distr — Beijing

Publications

• D. Reddy, "Data Engineering Challenges in AI automation," 2023 International Conference on Computing, Electronics & Communications Engineering (iCCECE), Swansea, United Kingdom, 2023, pp. 107-112
• Penberthy LT, Rivera DR, Lund JL, Bruno MA, Meyer AM. An overview of real-world data sources for oncology and considerations for research. CA Cancer J Clin. 2022 May;72(3):287-300. doi: 10.3322/caac.21714. Epub 2021 Dec 29. PMID 34964981
• Kam K.H. Ng, Chun-Hsien Chen, C.K.M. Lee, Jianxin (Roger) Jiao, Zhi-Xin Yang; A systematic literature review on intelligent automation: Aligning concepts from theory, practice, and future perspectives; Advanced Engineering Informatics; 2021 January; Volume 47; 101246

Identifiers

NCT: NCT07635355 · Data-Link

Primary sources (government registries)

• ClinicalTrials.gov (US NIH)

View this study on ClinicalTrials.gov ↗