JMIR Publications

ABSTRACT

Background:

Medical knowledge graph provides explainable decision support, helping clinicians with prompt diagnosis and treatment suggestions. However, in real-world clinical practice, patients visit different hospitals seeking various medical services resulting in fragmented patient data across hospitals. With data security issues, data fragmentation limits the application of knowledge graphs because single-hospital data cannot provide complete evidence for generating precise decision support and comprehensive explanation. It is important to study new methods for knowledge graph systems to implement in multicenter information-sensitive medical environment, utilizing fragmented patient records for decision support while maintaining data privacy and security.

Objective:

The goal of this study is to propose an electronic health record (EHR)-oriented knowledge graph system for collaborative reasoning on multicenter fragmented patient medical data, while preserving data privacy.

Methods:

The system was deployed in each hospital and used a unified semantic structure and OMOP vocabulary to standardize the local EHR dataset. The system transforms local EHR data into semantic formats and performs semantic reasoning to generate intermediate reasoning findings. The generated intermediate findings used hypernym concepts to isolate original medical data. The intermediate findings and hash-encrypted patient identities are synchronized through a blockchain network. The multicenter intermediate findings are collaborated for final reasoning and clinical decision support without gathering original EHR data.

Results:

The system was evaluated with an application study of utilizing multicenter fragmented EHR data to warn non-nephrology clinicians of unconsidered chronic kidney disease (CKD) patients. The study covered 1185 patients in non-nephrology departments from 3 hospitals. The patients visited at least 2 of the hospitals. Among them, 124 patients were found meeting CKD diagnosis criteria by collaborative reasoning of multicenter EHR data, while single-hospital data of these patients could not support the identification of CKD. The assessment by clinicians indicated that 85.71% of the patients were CKD positive.

Conclusions:

The proposed system was able to effectively utilize multicenter fragmented EHR data for clinical application. The application study showed the clinical benefits of the system with prompt and comprehensive decision support.