The need and key points for patient matching in clinical studies using the National Database of Health Insurance Claims and Specific Health Checkups of Japan (NDB)

DOI Web Site
  • Kubo Shinichiro
    Department of Public Heath, Health Management and Policy, Nara Medical University
  • Noda Tatsuya
    Department of Public Heath, Health Management and Policy, Nara Medical University
  • Myojin Tomoya
    Department of Public Heath, Health Management and Policy, Nara Medical University
  • Higashino Tsuneyuki
    ICT Innovation Division, Mitsubishi Research Institute, Inc.
  • Matsui Hiroki
    Department of Clinical Epidemiology and Health Economics, School of Public Health, The University of Tokyo
  • Kato Genta
    Solutions Center for Health Insurance Claims, Kyoto University Hospital
  • Imamura Tomoaki
    Department of Public Heath, Health Management and Policy, Nara Medical University

Bibliographic Information

Other Title
  • レセプト情報・特定健診等情報データベース（ NDB ）の臨床研究における名寄せの必要性と留意点
  • レセプト ジョウホウ ・ トクテイケンシン トウ ジョウホウ データベース(NDB)ノ リンショウ ケンキュウ ニ オケル ナヨセ ノ ヒツヨウセイ ト リュウイテン

Search this article

Abstract

<p>[Background]</p><p>The National Database of Health Insurance Claims and Specific Health Checkups of Japan (NDB) contains monthly claims for reimbursement of medical expenses and is managed by two types of personal identification (IDs) of a patient. Errors occur in linking the information of two IDs to a patient (patient matching). In this study, we summarize important points to identify.</p><p>[Methods]</p><p>The 2013 NDB data were used to identify errors that interfere with patient matching and to reveal the number and frequency of such errors.</p><p>[Results]</p><p>Both ID information include life event-related changes, such as an insurance provider’s ID number, employer, employee (patient) career changes, retirement, and a patient’s name change due to marriage. In addition, different patients are often identified as the same patient, which is called as Type I errors, such as persons who have same surnames, given names and birthdays, and twins who are dependents with the same surname. Furthermore, variations in writing used at different healthcare providers and even typos when computerizing potentially identify one patient as different individuals (Type II errors). Our 1-year follow-up study revealed that changes in the IDs occur in approximately 11% of patients; hence, these patients are usually excluded from cohort studies. Type I and II errors also occurred simultaneously in approximately 0.8% of patients, suggesting that an estimated 1% of patients may be difficult to follow-up.</p><p>[Conclusion]</p><p>Problems with patient matching include life event-related changes in patient IDs used in the NDB and insufficiency of key variables in patient matching. A cohort of patients with different names, regardless of anonymity, may be used as training data to investigate the accuracy of matching. Further study is needed to improve the patient matching system used in the NDB.</p>

Journal

Related Projects

See more

Keywords

Details 詳細情報について

Report a problem

Back to top