Zhang Yili, Callaghan-Koru Jennifer A, Koru Güneş
Innovation Center for Biomedical Informatics, Georgetown University, Washington, DC 20007, United States.
Department of Internal Medicine, University of Arkansas for Medical Sciences, Fayetteville, AR 72703, United States.
JAMIA Open. 2024 Aug 1;7(3):ooae058. doi: 10.1093/jamiaopen/ooae058. eCollection 2024 Oct.
Various data quality issues have prevented healthcare administration data from being fully utilized when dealing with problems ranging from COVID-19 contact tracing to controlling healthcare costs.
(i) Describe the currently adopted approaches and practices for understanding and improving the quality of healthcare administration data. (ii) Explore the challenges and opportunities to achieve continuous quality improvement for such data.
We used a qualitative approach to obtain rich contextual data through semi-structured interviews conducted at a state health agency regarding Medicaid claims and reimbursement data. We interviewed all data stewards knowledgeable about the data quality issues experienced at the agency. The qualitative data were analyzed using the Framework method.
Sixteen themes emerged from our analysis, collected under 4 categories: (i) Defect characteristics: Data defects showed variability, frequently remained obscure, and led to negative outcomes. Detecting and resolving them was often difficult, and the work required often exceeded the organizational boundaries. (ii) Current process and people issues: The agency adopted primarily ad-hoc, manual approaches to resolving data quality problems leading to work frustration. (iii) Challenges: Communication and lack of knowledge about legacy software systems and the data maintained in them constituted challenges, followed by different standards used by various organizations and vendors, and data verification difficulties. (iv) Opportunities: Training, tool support, and standardization of data definitions emerged as immediate opportunities to improve data quality.
Our results can be useful to similar agencies on their journey toward becoming learning health organizations leveraging data assets effectively and efficiently.
在处理从新冠病毒接触者追踪到控制医疗成本等问题时，各种数据质量问题阻碍了医疗管理数据的充分利用。
（i）描述当前为理解和提高医疗管理数据质量所采用的方法和实践。（ii）探索实现此类数据持续质量改进的挑战和机遇。
我们采用定性方法，通过在一家州卫生机构就医疗补助申请和报销数据进行的半结构化访谈来获取丰富的背景数据。我们采访了所有了解该机构所经历数据质量问题的数据管理员。使用框架法对定性数据进行分析。
我们的分析得出了16个主题，分为4类：（i）缺陷特征：数据缺陷表现出变异性，常常难以发现，且会导致负面结果。检测和解决这些问题通常很困难，所需工作常常超出组织边界。（ii）当前流程和人员问题：该机构主要采用临时的、手动的方法来解决数据质量问题，导致工作受挫。（iii）挑战：关于遗留软件系统及其所维护数据的沟通和知识匮乏构成挑战，其次是各组织和供应商使用的不同标准以及数据验证困难。（iv）机遇：培训、工具支持和数据定义标准化成为改善数据质量的直接机遇。
我们的结果对于类似机构在有效且高效利用数据资产迈向学习型健康组织的过程中可能会有所帮助。
