The Development of Dataset Tables for NDB Analyses

DOI Web Site
  • Fukuda Haruhisa
    Kyushu University Graduate School of Medical Sciences
  • Sato Daisuke
    Center of Outcomes Research and Economics Evalution for Health, National Institute of Public Health
  • Shiroiwa Takeru
    Center of Outcomes Research and Economics Evalution for Health, National Institute of Public Health
  • Fukuda Takashi
    Center of Outcomes Research and Economics Evalution for Health, National Institute of Public Health

Bibliographic Information

Other Title
  • NDB 解析用データセットテーブルの開発
  • NDB カイセキヨウ データセットテーブル ノ カイハツ

Search this article

Abstract

<p>Objectives: Although the National Database of Health Insurance Claims and Specific Health Checkups of Japan (NDB) was approved for third-party use in 2011, it remains underused for research purposes. There is therefore a need to improve the NDB’s usability to facilitate its applications in academic research and support its contributions to evidence-based health policy. This study aimed to construct NDB analytical datasets with high storage efficiency and specific applications for research in clinical epidemiology and health economics.</p><p>Methods: The NDB in this study comprises a random sample of 25% of all patients with at least one claims record in medical claims data or diagnosis procedure combination (DPC) claims data between April 2009 and December 2016. This database stores claims data for all medical services provided to these patients throughout the study period. We investigated the variables needed in analytical dataset tables that would facilitate research in clinical epidemiology and health economics. As medical claims data do not include information on discharge dates, we also examined methods for supplementing this information using other claims data. We first calculated the possible discharge dates using admission dates and the number of actual treatment days or dates of recorded treatments; these estimated discharge dates were then compared with the actual discharge dates obtained from the DPC claims data.</p><p>Results: We developed the following 11 analytical dataset tables that were organically linked using patient identification codes available in the NDB: Patient information (KAN), claims information (REC), disease information (SYO), medical services information (SIN), pharmaceutical information (IYA), device information (TOK), drug dispensing information (TYO), drug dispensing incentives information (TKA), DPC diagnostic group classifications (BUD), medical institution information (IRK), and hospital admissions information (ADM). With the exception of the IRK table, all tables could be mutually compared using patient identification codes. In the analysis of supplementary discharge dates for the medical claims data, estimates using the final date of recorded treatments (available in the SI, IY, TO, and CD files) were able to accurately identify the correct discharge date in 99.83% of all hospitalized cases.</p><p>Conclusion: The use of the analytical dataset tables developed in this study may help to establish an environment that facilitates the rapid initiation of research in clinical epidemiology and health economics using the NDB.</p>

Journal

Related Projects

See more

Keywords

Details 詳細情報について

Report a problem

Back to top