보건의료 빅데이터 이용 활성화를 위한 오픈 소스 데이터 분석 프로그램의 개발 = Development of Open-Source Data Analysis Program for Healthcare Big Data

다국어 초록 (Multilingual Abstract)

Objectives: An era of open and transparent information in Korean healthcare area is now underway. Health Insurance Review and Assessment Service has disclosed the National Patient Sample claims data since 2009 and National Health Insurance Service announced to disclose 9 year period cohort national health insurance claims database to healthcare stakeholders. Since Korea uses the fee-for-service scheme as basic payment system for all of medical treatments excepting for 7 common diseases which are run by Diagnosis Related Group, it is easy to identify the medical treatment practice and resource utility information for individual medical procedures. The use of SAS software is the generally accepted data analysis tool as the average data size of national health insurance claims data easily exceeds over 30 Giga Bytes. However, the data analysis using SAS is labor-intensive and time-consuming works and has a low accessibility due to its costly license fees. As the need to analyze the Healthcare Big Data faster and appropriately rises, demand for development of new data analysis tool is also significantly increasing.
Methods: Open-source big data analysis program with the name of BigPy was developed using Py- thon which is a high-level object oriented programming language. BigPy’s design philosophy empha- sizes on code readability and reusability, and its syntax allows users to express concepts in fewer lines of code than would be possible in statistical software such as SAS or R. Results: Bigpy program is com- posed of a series of data analysis macro and functions. The functions in BigPy can easily read, trim, sort, and merge the healthcare big data with database format and convert large dataset to a Hierarchical Data Format Version 5 file. Conclusion: Healthcare stakeholders now have access to promising new value of knowledge that is called Big Data. The efforts on Big Data analysis can address problems related to variability in healthcare quality and consequently improve healthcare treatments. The development of open-source data analysis is a noteworthy and promising methodology to handle the healthcare big data in a rapid and a cost-effective way.

참고문헌 (Reference)

1 고민정, "근거중심 보건의료의 시행을 위한 빅데이터 활용" 대한의사협회 57 (57): 413-418, 2014

2 HIRA, "Workshop for HIRA Claims Data Analysis"

3 Groves P, "The ‘big data’ revolution in healthcare" McKinsey&Company 2013

4 "The HDF Group"

5 HIRA, "Structure and Application Method for National Patient Sample Data"

6 HIRA, "SAS Manual for HIRA Claims Data Analysis"

7 Das S, "Ricardo: integrating R and Hadoop" ACM 987-998, 2010

8 "Python program language"

9 "Pandas-Python Data Analysis Library"

10 Srinivasan U, "Leveraging big data analytics to reduce healthcare costs" 15 : 21-28, 2013

11 "Korean Governments Version 3.0 Plan in 2014"

12 Kim RY, "Introduction to Health Insurance Review & Assessment Service-National Patient Sample (HIRA-NPS)" 6 : 33-47, 2012

13 Baasal A, "Healthcare Data Analysis using Dynamic Slot Allocation in Hadoop" 3 : 15-18, 2014

14 White T, "Hadoop: the definitive guide" O’Reilly Media, Inc. 2009

15 Lee JY, "Establishment of Sample-Cohort DB using NHIS Database"

16 Song TM, "Efficient Management of Big Data on Health & Welfare" 193 : 68-76, 2012

17 Song TM, "Big-Data Trend in Korean Healthcare and Application Research" 192 : 56-73, 2013

18 Feldman B, "Big data in healthcare hype and hope. Dr. Bonnie 360: Business Development for Digital Health"

19 Raghupathi W, "Big data analytics in healthcare: promise and potential" 2 : 3-, 2014

20 Lee JH, "Big Data Application Trend in Healthcare Sector" 32 : 63-75, 2014

21 Taylor RC, "An overview of the Hadoop/MapReduce/HBase framework and its current applications in bioinformatics" 11 (11): S1-, 2010

22 Barreto M, "A Spark-based workflow for probabilistic record linkage of healthcare data"