DATA QUALITY EVALUATION OF 5 MUNICIPAL AND PROVINCIAL CLAIMS DATABASES OF FUJIAN PROVINCE IN CHINA
Author(s)
Chen C1, Xie X2, Li L2, Ye X2, Ding R3, Huang S3, Zhang Y4
1Fujian Medical University, Fuzhou, China, 2Pfizer Investment Co., Ltd., Beijing, China, 3Fujian Social Health Insurance Association, Fuzhou, China, 4Fujian Social Health Insurance Management Center, Fuzhou, China
OBJECTIVES Claims databases provide one important source of data for Real-World Evidence generations. Therefore, data quality is an important factor affecting the generation of the data value. In this study, we assessed the data quality of the regional claims from Fujian in China. METHODS We retrieved data of 395,950 patients with an Arteriosclerotic Cardiovascular Disease diagnosis, from the health insurance claims database of Fujian province (including 5 municipal and provincial claims databases) from 2005-2018. The database contains the following data elements: patient information, date, place and type of service, provider specialty, diagnosis information (international classification of diseases, ICD), prescription, and all charges and reimbursements associated with each claim. We assessed the data quality in three dimensions: Completeness, Concise Representation, and Consistent Representation. RESULTS For Completeness, 20% of the ICD code in the database is missing. In addition, 6% and 32% of claims are lacking admission dates and discharge dates, respectively. Moreover, the prescription information: supply days, frequency, and drug doses are frequently missing from the database. Furthermore, the medical record homepages are not available for some of the patients. For Consistent Representation, the first issue is some of the pharmacy claims data cannot be linked to the hospital claims data, due to inconsistent identifier information. Secondly, some medication information lacks a unified format. For Concise Representation, there may have been mistakes made inputting the unit price, and amount of the drug. CONCLUSIONS Based on our assessment, some strategies should be implemented to improve the data quality. Firstly, pharmacy claims data should be linked with hospital claims data. Secondly, standardized rules should be developed to ensure the physicians input information accurately during the doctor/patient visit. Improving data quality of the insurance claims database would be of great significance in generating accurate, valuable evidence, and for reimbursement management, as well.
Conference/Value in Health Info
2019-11, ISPOR Europe 2019, Copenhagen, Denmark
Code
PCV147
Topic
Health Policy & Regulatory, Organizational Practices, Real World Data & Information Systems
Topic Subcategory
Best Research Practices, Data Protection, Integrity, & Quality Assurance, Health & Insurance Records Systems, Insurance Systems & National Health Care
Disease
Cardiovascular Disorders, Multiple Diseases