中国医疗设备2025，Vol.40Issue(11)：38-43,6.DOI:10.3969/j.issn.1674-1633.20250371
基于关联网络的死亡及非医嘱离院患者高频诊断规律分析
Analysis of High-Frequency Diagnostic Patterns for Deceased and Uninstructed Discharged Patients Based on Association Networks
摘要
Abstract
Objective To explore the analysis of the diagnostic patterns of deaths and patients discharged without medical advice,and to provide a methodological basis for reducing the rate of patients discharged without medical advice and ensuring the quality and safety of patients.Methods The first-page diagnostic information of patients who died and were discharged without medical advice in 2024 was selected.Factor analysis and cluster analysis were used to initially screen the diagnostic clusters,and the Apriori algorithm was utilized to construct the association network model.The model fitting effect was evaluated through KMO and Bratlett sphere tests,and the diagnostic association groups were visually presented with the aid of load diagrams,lineage diagrams and network diagrams.Results The variables were classified in combination with the magnitudes of factor loadings,and 28 diagnostic groups were obtained.Through high-frequency diagnostic cluster analysis,eight effective cluster combinations were formed for comparison.Conclusion The association network model can fit well the diagnostic association groups of deceased and uninstructed discharged patients.By frequently diagnosing the associated incidence rate,early warning for high-risk patients with multiple organ dysfunction within the hospital can be achieved,thereby enabling the prevention and control of nosocomial infection risks and optimizing the allocation of medical resources.关键词
关联网络/因子分析/聚类分析/Apriori算法/死亡及非医嘱离院/诊断规律Key words
association network/factor analysis/cluster analysis/Apriori algorithm/death and discharge without medical advice/diagnostic rules分类
医药卫生引用本文复制引用
袁筱祺,高玮,孔雯,董笑..基于关联网络的死亡及非医嘱离院患者高频诊断规律分析[J].中国医疗设备,2025,40(11):38-43,6.基金项目
国家自然科学基金青年项目(82203742) （82203742）
上海申康医院发展中心诊疗技术推广及优化管理项目(SHDC12025628). （SHDC12025628）
