【摘要】 目的 本研究基于人工智能技术优化中医医院一体化信息管理平台，构建中医“防治康”（预防、治疗、康复）一体化数智化管控体系，解决胸痹病（冠心病）“防治康”全流程中存在的数据孤岛、辨证标准化不足、随访依从性低三大核心问题。方法 依托中山市原创性“三环耦合”理论模型（预防—治疗—康复），融合人工智能技术设计四级技术架构：(1)数据层整合多源设备，建立含352项标准化数据元的专病库；(2)平台层部署Hadoop数据中心与联邦学习框架；(3)AI算法层构建LSTM-XGBoost辨证模型及基于Apriori算法的古籍方剂推荐引擎；(4)应用层开发多终端系统贯通全链条。采用前瞻性对照设计，纳入972例胸痹病患者，按就诊科室诊疗路径分配为干预组（486例，应用该数智化体系）与对照组（486例，采用传统诊疗模式），通过多重插补法处理随访数据（迭代50次）。结果 两组基线特征（年龄、性别、病程、证型分布、合并症）均衡可比（P均>0.05），组间具有良好可比性。干预组30天再住院率为7.2%，较对照组的11.93%降低39.7%(P<0.05)；干预组中医证候积分（TCM-SSS）改善度为47.6±11.9分，较对照组的39.6±6.8分提高20.2%(P<0.001)；干预组例均医保费用较对照组显著降低14.8%(P<0.05)；干预组病历关键字段完整性达99.6%，较对照组的82.7%显著提升20.4%(P<0.001)。结论 该数智化管控体系通过数据标准化治理、人工智能辅助决策及全流程闭环管理，较传统模式一定程度改善胸痹病“防治康”效果，为中医优势病种数智化管控提供可复制范式。更多还原

【Abstract】 Objective To address three core issues in the full process of prevention, treatment, and rehabilitation(PTR) of Chest Bi Syndrome(coronary heart disease) — namely data silos, insufficient standardization of syndrome differentiation, and low follow-up compliance — this study optimized the integrated information management platform of Traditional Chinese Medicine(TCM) hospitals based on artificial intelligence(AI) technology and constructed a digital-intelligent management and control system for integrated PTR of TCM. Methods Relying on the original "three-ring coupling" theoretical model(prevention-treatment-rehabilitation) in Zhongshan City and integrating AI technology, a four-level technical architecture was designed:(1)The data layer integrated multi-source devices to establish a disease-specific database containing 352 standardized data elements;(2)The platform layer deployed a Hadoop data center and a federated learning framework;(3)The AI algorithm layer constructed an LSTM-XGBoost syndrome differentiation model and an ancient prescription recommendation engine based on the Apriori algorithm;(4)The application layer developed a multi-terminal system to connect the entire workflow. A prospective controlled design was adopted, enrolling 972 patients with Chest Bi Syndrome. According to the diagnosis and treatment pathway of their visited departments, patients were assigned to the intervention group(n=486, applying the digital-intelligent system) and the control group(n=486, receiving traditional diagnosis and treatment mode). Follow-up data were processed using multiple imputation(50 iterations). Results The 30-day readmission rate was 7.2% in the intervention group, which was 39.7% lower than that in the control group(11.93%, P < 0.05). The improvement in TCM syndrome score(TCM-SSS) was 47.6±11.9 in the intervention group, representing a 20.2% increase compared with the control group(39.6±6.8, P < 0.001). The average medical insurance cost per case was significantly reduced by 14.8% in the intervention group(P < 0.05). The completeness rate of key medical record fields was 99.6% in the intervention group, significantly higher than that in the control group(82.7%, P < 0.001).Conclusion Through standardized data governance, AI-assisted decision-making, and full-process closed-loop management, the digital-intelligent management and control system moderately improves the prevention, treatment, and rehabilitation outcomes of Chest Bi Syndrome compared with the traditional model. It provides a replicable paradigm for digital-intelligent management of TCM-dominant diseases.更多还原