Integrating clinicians, knowledge and data: expert-based cooperative analysis in healthcare decision support.
Integrating clinicians, knowledge and data: expert-based cooperative analysis in healthcare decision support.
复制标题
DOI：
10.1186/1478-4505-8-28
复制
发表时间：
2010-09-30
影响因子：
4
通讯作者：
Salvador-Carulla L
中科院分区：
文献类型：
--
作者：
Gibert K;García-Alonso C;Salvador-Carulla L
卫生系统中的决策支持是一项非常困难的任务，因为所涉及的过程和结构具有固有的复杂性。本文介绍了一种新的基于专家的合作分析（EbCA）混合方法，该方法将明确的先验专家知识纳入数据分析方法，并引出隐含或隐性专家知识（IK），以改善医疗保健系统的决策支持。EbCA已应用于两个不同的案例研究，显示了它的可用性和通用性：1)基于EbCA-数据包络分析（EbCA- dea）估计的技术效率的小心理健康领域基准测试；2)基于功能依赖的基于规则聚类（ClBR）的精神分裂症病例混合。在这两种情况下，对使用定性明确先验知识的经典程序进行了比较。贝叶斯预测效度测量用于与专家小组结果的比较。用类内相关系数（Intraclass Correlation Coefficient）和kappa检验两种情况的总体一致性。EbCA是一种由6个步骤组成的新方法：1)数据收集和数据准备；2)“先验知识”（PEK）的获取和“先验知识库”（PKB）的设计；3) pkb导向分析；4)支持解释工具来评估结果和检测不一致（这里可能会引出隐性知识- ik）；5)在PKB中加入诱导的IK，重复直到满意的解决方案；6)决策支持的后处理结果。EbCA有助于将PEK纳入两种不同的分析方法（DEA和聚类），分别用于评估小型精神卫生领域的技术效率和基于功能依赖的精神分裂症病例组合。经典方法获得的结果差异主要与使用EbCA可以引起的IK有关，并且对两种情况下的决策都有重大影响。本文介绍了EbCA，并展示了用PEK作为提取复杂健康领域相关知识的手段来完成经典数据分析的便利性。EbCA的主要优点之一是IK的迭代引出。显性和隐性或隐性专家知识对于指导对卫生系统研究中发现的非常复杂的决策问题进行科学分析至关重要。
Decision support in health systems is a highly difficult task, due to the inherent complexity of the process and structures involved. This paper introduces a new hybrid methodology Expert-based Cooperative Analysis (EbCA), which incorporates explicit prior expert knowledge in data analysis methods, and elicits implicit or tacit expert knowledge (IK) to improve decision support in healthcare systems. EbCA has been applied to two different case studies, showing its usability and versatility: 1) Bench-marking of small mental health areas based on technical efficiency estimated by EbCA-Data Envelopment Analysis (EbCA-DEA), and 2) Case-mix of schizophrenia based on functional dependency using Clustering Based on Rules (ClBR). In both cases comparisons towards classical procedures using qualitative explicit prior knowledge were made. Bayesian predictive validity measures were used for comparison with expert panels results. Overall agreement was tested by Intraclass Correlation Coefficient in case "1" and kappa in both cases. EbCA is a new methodology composed by 6 steps:. 1) Data collection and data preparation; 2) acquisition of "Prior Expert Knowledge" (PEK) and design of the "Prior Knowledge Base" (PKB); 3) PKB-guided analysis; 4) support-interpretation tools to evaluate results and detect inconsistencies (here Implicit Knowledg -IK- might be elicited); 5) incorporation of elicited IK in PKB and repeat till a satisfactory solution; 6) post-processing results for decision support. EbCA has been useful for incorporating PEK in two different analysis methods (DEA and Clustering), applied respectively to assess technical efficiency of small mental health areas and for case-mix of schizophrenia based on functional dependency. Differences in results obtained with classical approaches were mainly related to the IK which could be elicited by using EbCA and had major implications for the decision making in both cases. This paper presents EbCA and shows the convenience of completing classical data analysis with PEK as a mean to extract relevant knowledge in complex health domains. One of the major benefits of EbCA is iterative elicitation of IK.. Both explicit and tacit or implicit expert knowledge are critical to guide the scientific analysis of very complex decisional problems as those found in health system research.
登录
查看更多内容
DOI：
10.1682/jrrd.2004.02.0016
发表时间：
2004-11-01
影响因子：
--
作者：
Annicchiarico, R;Gibert, K;Caltagirone, C
通讯作者：
Caltagirone, C
影响因子：
6
作者：
Humphris, GM
通讯作者：
Humphris, GM
DOI：
10.1007/bf02196348
发表时间：
1994-09-01
期刊：
The journal of primary prevention
影响因子：
--
作者：
Duncan, D F
通讯作者：
Duncan, D F
影响因子：
4
作者：
Fretheim, Atle;Munabi-Babigumira, Susan;Lewin, Simon
通讯作者：
Lewin, Simon
影响因子：
3.6
作者：
Cinca, CS;Molinero, CM
通讯作者：
Molinero, CM