基于迭代校正的病例临床复杂度度量方法研究  

An Iteration-Based Episode Clinical Complexity Estimation Method

作　　者：陶丽[1] 王纯青 张自力[1] 陈波[2] TAO Li;WANG Chunqing;ZHANG Zili;CHEN Bo(School of Computer and Information Science,Southwest University,Chongqing 400715,China;The First Affiliated Hospital of Chongqing Medical University,Chongqing 400700,China)

机构地区：[1]西南大学计算机与信息科学学院,重庆400715 [2]重庆医科大学附属第一医院,重庆400700

出　　处：《西南大学学报（自然科学版）》2022年第5期185-191,共7页Journal of Southwest University(Natural Science Edition)

基　　金：国家自然科学基金面上项目(61976181);西南大学中央高校基金基本科研业务费项目(XDJK2019D018).

摘　　要：现有病例临床复杂度(episode clinical complexity,ECC)模型在评估临床复杂程度时主要关注诊断数目的不同,忽略了不同诊断类型造成的差异.本文提出了一种基于迭代校正思想,将诊断数目和诊断类型同时纳入病人费用和病情复杂水平评估过程的ECC改进模型(iteration-based ECC,IECC).基于重庆某大型三甲医院2015-2019年呼吸系统、循环系统的病案首页数据的实验结果表明,根据IECC方法估算得到的住院费用与真实费用的相关性有较大提升,说明本文所提方法能够更好地评估病人临床复杂水平,为成功实施以疾病诊断相关分组(diagnosis related groups,DRGs)为核心的付费制度提供有力支持.How to extract features from high-dimensional sparse data and obtain new features with rich information is an important problem that are faced by many application fields such as hospital managements and product recommendations. Episode clinical complexity score(ECCS, i.e., the extracted new feature) has higher correlations with patients’ actual costs. Episode Clinical Complexity(ECC) model is a classic method to quantify the complexity of treatment based on the assumption that a larger number of diagnoses need more resource for treatment, and thus should be classified into higher complexity levels. However, the ECC model assumes that the diagnosis type has little effect on the medical cost, which is not consistent with the real situation. This paper proposes an Iteration-Based ECC(IECC) model, which incorporates the number and the type of diagnoses into the assessment process of clinical complexity levels. We used the first page data in medical records of inpatients with respiratory or circulatory problems in a 3 A-Grade hospital in Chongqingduring the years of 2015 and 2019 to evaluate the effectiveness of the IECC model to estimate the clinical complexity with the ECC model. The experiment results showed that the estimated medical cost based on IECC model has higher associations with the real medical cost than that of calculated based ECC model. This implies that IECC model can better quantify the episode clinical complexity, and thus can provide a stronger support for the successful implementation of DRGs.

关 键 词：病案首页数据 高维稀疏数据 病例临床复杂度度量 迭代校正 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]