A Study on the Improvement for big data utilization- Focused on health insurance claims data

Abstract

The rapid development of medical technology and information and communication technology (ICT) in Korea has led to the accumulation of vast amounts of information related to healthcare. The National Health Insurance Service (NHIS) and Health Insurance Review & Assessment Service (HIRA) in Korea collect and store health insurance claims data. Despite the excitement and recent interest in healthcare big data, few empirical studies have been conducted to evaluate the potential value of health insurance claims data. The following three methods were used to suggest strategies for optimal utilization of Korean health insurance claims data. First, Systematic Review was conducted of published studies related to Korean Health Insurance Claims Data. The PubMed and Cochrane database searches from 2007 to 2017. A total of 478 studies were included in the study after applying duplication and elimination criteria to the initial 3,951 search results. Second, comparative analysis was conducted to draw implications for using Korean Health Insurance Claims Big Data across countries (US, UK, Australia and Taiwan). Cross-country comparisons were performed based on horizontal as well as vertical comparison perspectives. Data analysis consisted of the constant comparison method. Third, a Delphi survey was conducted to 42 healthcare professionals working in National Health Insurance Service (NHIS), Health Insurance Review & Assessment Service (HIRA), and relevant agencies. The questionnaire content was intended to identify the obstacles to and policy priorities for the safe use of Health Insurance claims data. This study questionnaire was approved by the IRB Institutional Review Board at Yonsei University (IRB: 2-1040939-AB-N-01-2014-228). The results of the three methods of this study are as follows. First, 478 studies were selected as a result of systematic review. There were 55 studies (11.5%) between 2007 and 2012, and a total of 423 (88.5%) were found over the past five years (2013–2017). The HIRA database was used a little more often than NHIS database (HIRA: 51.9%, NHIS: 47.5%). The most frequent research type was health service utilization (41.4%), and 29 (6.9%) out of 478 cases were connected with external data. These data include the information from the cause of death data (12, 41.4 %), clinical data (9, 31.0%), cancer data (7, 24.1%), cost data (6, 20.7%), Surveillance data (2, 6.9%), other data (3, 10.3%). Second, this study shows the implications for policies in Korea through comparison of the big data utilization in the major countries. The experience of developed countries suggests important issues to be reflected in the formulation of strategies for national utilization of healthcare data; there is a national strategy and health and data governance was being built, it focuses on utilization of public interest objectives such as improvement of public health and medical quality, there is a balance between strengthening and balancing privacy and data security. Third, 13 policies that indicate four obstacles were included through the Delphi survey. Participants responded by rating the four obstacles in this order: legal immaturity for data use, lack of consensus on providing information, technical constraints on information sharing, and lack of government support. Policy priorities include policy for the “patient’s consent to data use,” a policy for legal revision for Health Insurance Big Data utilization, an institutional improvement policy for Health Insurance Big Data utilization, an institutional consent policy for data provision, technical privacy policies such as anonymization for data sharing, and a national governance establishment policy for health insurance claims data utilization. Finally, three strategies have been proposed for each issue derived from the three methodologies. First, it is necessary to establish “National Big Data Governance” for the successful utilization of health related big data. Second, it is necessary to develop legal institutional guidelines in the framework of the separate big data law (differentiation of personal information consent, development of legal and institutional guidelines). The method of consent should be improved to resolve the dilemma whereby utilizing and protecting personal information. Third, it is a strategy to revitalize healthcare research for big data linkage (development of personal information protection technology for data linkage, utilization of user - centered health insurance claim data). Although Korea is aware of the global trends of big data, negative opinions are still common about the view that the use of personal information is inevitable to improve the quality of life through public well-being and public health promotion. Clear legislative and institutional grounds for the use of Health Insurance Big Data are needed and government support for the proposed policy recommendations should be established.