脑卒中康复评定知识图谱与评定推理的构建研究

Design of Rehabilitation Knowledge Graph for Post-stroke Assessment and Assessment Reasoning

目的:利用结构化的康复医学知识,构建一套脑卒中康复的知识图谱,进而通过算法给出康复评定与治疗的建议。方法:提出了一种针对脑卒中康复评定的知识图谱构建方法,其特点是全面纳入了《脑卒中康复评定指南》1.0版中涉及的康复评定量表,并给出了如何与临床指南、WHO标准、医学文献等数据源进行语义关联。基于该方法构建了卒中康复评定知识图谱(Rehabilitation Knowledge Graph,RKG)。结果:RKG现已具备146个卒中评定量表,每个量表具备42个属性,并与ICF和ICD-11相关的术语条目建立了近3000条语义连接。测试表明,RKG图谱可以实现对卒中康复评定知识的多线索查询,如某卒中患者心理评定总时间超出承受范围。结论:RKG具备了整合结构化脑卒中康复评定知识的能力,对康复临床决策支持的智能化以及康复医学技术推广至基层具有重要意义。

Objective:This study aims to leverage structured rehabilitation medical knowledge,and build a knowledge graph for stroke rehabilitation,which provides algorithm-based recommendations for rehabilitation assessment and treatment.Methods:This paper proposes a methodology for constructing a knowledge graph for stroke rehabilitation assessment,which is characterized by inclusion of the rehabilitation assessment scale involved in the 1.0 edition of the Stroke Rehabilitation Assessment Guidelines and clarifications such as WHO standards Data Sources.Based on this methodology,we constructed a stroke rehabilitation knowledge graph(RKG).Results:RKG includes 146 publicly searchable stroke assessment scales,with 42 attributes for each clinical scale,and we established nearly 3,000 ICF and ICD-11 related items.RKG showed its capability of multi-cue query on stroke rehabilitation assessment.It may generate clinical insights,e.g.the total time of psychological assessment of stroke patients exceeds the tolerated maximum.Conclusion:RKG is capable of synthesizing structrured knowledge on stroke rehabilitation assessment,which is significant for intelligent clinical support and the facilitatation of new technology for community-level rehabilitation.