区块链和人工智能驱动的脑血管病诊疗和质量管理决策范式研究

Cerebrovascular disease is the leading cause of death and disability in adults in China. Previous studies of our research group have proved that the linear paradigm management process based on clinical pathway could improve the prognosis of patients with cerebrovascular disease, but there were certain shortcomings, such as increased workload of doctors, non-embedding of external decision-making elements, data security, and poor traceability. The applicant presided over the National Key Research and Development Project "Cerebrovascular Disease Clinical Research Big Data and Biological Sample Database Platform Construction and Key Technology Research", established more than 1.5 million clinical research big data platform for cerebrovascular disease, and developed a cerebrovascular disease diagnosis and treatment and quality management decision-making system based on artificial intelligence and a medical quality evaluation system based on blockchain. In view of the shortcomings of the linear paradigm management process above, and relying on the existing resources and technologies of big data for cerebrovascular diseases, the current study aims to integrate blockchain technology with decision-making algorithms for diagnosis, treatment and quality management of cerebrovascular diseases based on artificial intelligence. The method of data-driven cerebrovascular disease diagnosis and treatment and quality management decision-making paradigm was established, and an empirical demonstration was carried out in 28 hospitals with interconnected cerebrovascular disease databases. Relying on the National Center for Healthcare Quality Management in Neurological Diseases of the National Health Commission, it has been gradually generalized and applied in 2,497 quality control hospitals across the country, providing a paradigm for the diagnosis, treatment, and quality management of big data-driven chronic diseases (e.g. cerebrovascular diseases) in the context of management decision-making.

脑血管病是我国成人死亡和致残的首位病因。课题组前期研究证明基于临床路径等的线性范式管理流程可改善脑血管病患者预后，但存在增加医师工作量、外部决策要素未嵌入、数据安全及可溯源性差等问题。申请人主持国家重点研发计划“脑血管病临床研究大数据与生物样本库平台构建和关键技术研究”，建立超过150万的脑血管病临床研究大数据平台，研发了基于人工智能的脑血管病诊疗和质量管理决策系统和基于区块链的医疗质量评价系统。针对上述线性范式管理流程的短板，依托已有脑血管病大数据资源和技术，本研究拟将区块链技术和基于人工智能的脑血管病诊疗和质量管理决策算法融合，建立数据驱动的脑血管病诊疗和质量管理决策范式的方法，并在前期脑血管病专病互联互通的28家医院开展实证示范。依托国家卫生健康委神经系统疾病医疗质量控制中心在全国2497家质控医院逐步推广应用，为管理决策情境下大数据驱动的脑血管病等慢病的诊疗和质量管理方法提供范式。

在健康医疗大数据的时代背景下，新型的脑血管病诊疗管理决策范式呈现出人工智能和区块链等大数据核心技术驱动的全景式特点。针对传统管理流程线性范式为主的脑血管病诊疗和质量管理决策存在的临床医师工作量重、依从性不佳、数据安全和可溯源性差、多维交互动态与全要素参与不足等短板，在项目前期建立的脑血管病临床大数据和医疗质量改进模式基础上，利用前期建立脑血管病大数据资源平台和基于人工智能的脑血管病诊疗和质量管理决策系统算法融合的重大成果，通过人工智能的技术增强引入新视角，推动临床特征、影像特征、指南推荐、医疗质量和患者结局等新型变量全景交互，建立人工智能大数据驱动的脑血管病诊疗和质量管理决策范式的方法，并在前期脑血管病专病数据库互联互通的医院开展实证示范。研究纳入40家医院进行推广应用，评价脑血管病临床诊疗辅助决策系统对急性缺血性卒中患者医疗质量改进的作用，目前已经入组了4407例患者。该体系通过人工智能技术将优质医疗资源下沉到基层，规范脑血管病急性期诊治及二级预防管理，推进脑血管病诊疗的标准化和同质化。