Browse > Article

Efficient DRG Fraud Candidate Detection Method Using Data Mining Techniques  

Lee, Jung-Kyu (Department of Health Policy and Management, Seoul National University College of Medicine)
Jo, Min-Woo (Department of Preventive Medicine, College of Medicine, University of Ulsan)
Park, Ki-Dong (Department of Health Policy and Management, Seoul National University College of Medicine)
Lee, Moo-Song (Department of Preventive Medicine, College of Medicine, University of Ulsan)
Lee, Sang-Il (Department of Preventive Medicine, College of Medicine, University of Ulsan)
Kim, Chang-Yup (Graduate School of Public Health, Seoul National University)
Kim, Yong-Ik (Department of Health Policy and Management, Seoul National University College of Medicine)
Hong, Du-Ho (Department of Health Policy and Management, Seoul National University College of Medicine)
Publication Information
Journal of Preventive Medicine and Public Health / v.36, no.2, 2003 , pp. 147-152 More about this Journal
Abstract
Objectives : To develop a Diagnosis-Related Group (DRG) fraud candidate detection method, using data mining techniques, and to examine the efficiency of the developed method. Methods ; The Study included 79,790 DRGs and their related claims of 8 disease groups (Lens procedures, with or without, vitrectomy, tonsillectomy and/or adenoidectomy only, appendectomy, Cesarean section, vaginal delivery, anal and/or perianal procedures, inguinal and/or femoral hernia procedures, uterine and/or adnexa procedures for nonmalignancy), which were examined manually during a 32 months period. To construct an optimal prediction model, 38 variables were applied, and the correction rate and lift value of 3 models (decision tree, logistic regression, neural network) compared. The analyses were peformed separately by disease group. Results : The correction rates of the developed method, using data mining techniques, were 15.4 to 81.9%, according to disease groups, with an overall correction rate of 60.7%. The lift values were 1.9 to 7.3 according to disease groups, with an overall lift value of 4.1. Conclusions : The above findings suggested that the applying of data mining techniques is necessary to improve the efficiency of DRG fraud candidate detection.
Keywords
Diagnosis-Related Groups; Fraud; Decision trees; Neural networks;
Citations & Related Records
연도 인용수 순위
  • Reference
1 Sokol L, Garcia B, Rodriguez J, West M, Johnson K. Using data mining to find fraud in HCFA health card claims. Top Health Inf Manage 2001; 22(1):I-13   PUBMED
2 Hand D, Mannila H, Smyth P. Principlesof data mining. London: The MIT Press; 200I. p. 327-365
3 김광용. 데이터마이닝 기법의 성과평가 및 새로운 위험분류측정에 관한 실증적 연구. 보험개발연구 2001; 12(2): 133-166
4 Korcok M. Medicare, Medicaid fraud a billion-dollar art form in the US. CMAJ 1997; 156(8): 1195-1997
5 박일용，안철경. 보험사기 성향 및 규모추정. 보험개발원 보험연구소; 1999
6 김헌수. 보상전문가의 지식을 이용한 보험사기의 조기경보 모형의 개발에 관한 연구. 한국리스크관리학회; 1998
7 서울대학교 의과대학 의료관리학교실 . DRG 지불제도 시범사업 평가 및 개선방안 연구; 2000, (205-300쪽)
8 Suhn MO, Chae YM, Lee HJ, Lee SH, Kang SH, Ho SH. An application of data mining approach to CQI using the dis-charge summary. J Korean soc med inform 2000; 6(4): 1-13 (Korean)
9 Ho SH, Chae YM, Choi MY, Song MR. Development of critical pathway for the cesarean section using data mining. J Korean soc med inform 2002; 8(2): 41-68 (Korean)
10 Carter GM, Newhouse JP, Relles DA. How much change in the case mix index in DRG creep? J Health Econ 1990.; 9(4): 411-428   DOI   ScienceOn
11 Santon TH. Fraud-and-Abuse Enforcement in Medicare: Finding Middle Ground. Health Affair 2001; 20(4): 28-42   DOI   ScienceOn
12 Kim ON, Kim YH, Kang SH, Kim SH. A study on effects of critical pathway prac-tices by using BSC and datamining me-thod. J Korean soc med inform 2002; 8(2): 51-68(Korean)
13 강현철，한상태，최종후 김은석，김미경. 데이터마이닝-방법론 및 활용. 자유아카데미; 2001, (84-86쪽)
14 Hsia DC, Ahern CA, Ritchie BP, Moscoe LM, Krushat WM. Medicare reimburse-ment accuracy under the prospective payment system, 1985 to 1988. JAMA 1992; 268(7): 896-899   DOI   ScienceOn
15 보건복지부. 건강보험요양급여행위 및 그상대가치점수(보건복지부 고시 제20이 -70호) ; 2001