臺灣博碩士論文加值系統

中文摘要
本研究目的主要是探討DRG127『心臟衰竭與休克』之12701及12702個案人口學、臨床特性之差異性及分析Tw-DRGs第三版支付方式中之三個方案對醫療機構之財務衝擊，並利用CART決策樹演算法依收益差發掘歸納規則，瞭解收益差之預測變數以及各權屬層級醫院之個案特性。所使用方法研究個案為2004年全民健康保險研究資料庫住院申報檔中DRG127個案(排除尚未出院之切帳申報案件)，DRG12701個案19,253筆；DRG12702個案5,431筆，共24,684筆，與醫事機構基本資料檔串檔，取得醫院層級別，權屬別變數。統計方法有描述性統計(Descriptive statistics)、chi-square test檢定、t-test檢定、對數線性模式(log-linear analysis)、二因子變異數分析(two-way ANOVA) 、二因子重複量變異數分析(two-way repeated ANOVA)等，且利用資料探勘CART決策樹分析影響醫院財務衝擊之因素。結果顯示DRG12701與DRG12702之平均診斷編碼個數、平均處置編碼個數及平均住院日均有顯著差異(P<0.0001)，呼吸治療或洗腎之處置有無與DRG12701或DRG12702有關聯(P<0.0001)，有呼吸治療或洗腎個案相對於無呼吸治療或洗腎個案有4.18倍勝算是歸戶於DRG12701，並且應用 CART決策樹分析在控制性別、年齡、診斷數、處置數等變數下，住院日與有無使用呼吸治療或洗腎會影響收益差變異的重要變數，亦發現住院日愈長及有使用呼吸治療或洗腎的個案對財務衝擊愈大。DRG12701與DRG12702有60%以上個案在區域層級以上醫院。DRG12701平均每住院人次收益差正值之醫院以方案一為適當方案。就DRG12702而言，醫學中心以方案二為適當方案，其他權屬別層級醫院則以方案一為較適當。Tw-DRGs第三版對醫學中心、區域醫院及地區教學醫院有負面財務衝擊尤以醫學中心影響最大。公立醫學中心在平均每住院人次負收益差值最大的規則集所佔案件數比，無論是DRG12701或DRG12702或方案都比其他醫學中心、各權屬區域及地區教學醫院高。雖不同層級醫院Tw-DRGs第三版有基本診療加成但仍彌補不了醫院的虧損。本研究建議在面對即將導入Tw-DRGs支付制度下，醫療院所臨床路徑之規劃、醫師主動參與、以及健保局針對疾病嚴重度分析需有完善準備，方能達成此支付制度實施之目標，亦即在於控制國內醫療費用成長之理想。
關鍵詞 : Tw-DRGs、心臟衰竭、CART決策樹 。

Abstract
DRG 12701 heart failure and shock with CC was the leading diagnosis-related group discharge diagnosis of MDC 5 in 2004 Taiwan NHI research data. The objective of this study sought to define the demographic and clinical characteristic-related differences between DRG 12701 and 12702 (heart failure and shock without cc). It also explored the hospital financial risk associated with these patients under the Tw-DRGs 3rd version. CART decision tree was conducted to build up the model to investigate the important variables that can explain the variation of difference in revenue between Tw-DRGs payment system and fee-for-services. Characteristics of hospitalized patient were further identified. Materials and methods consisted of data on DRG 127 inpatient discharge acquired from 2004 NHI research data warehouse, undischarged claim data were excluded. This study reviewed the data with regards to demographic and hospital characteristics, length of stay (LOS), charges, the number of diagnoses and procedures, and ICD codes of mechanical ventilation and hemodialysis. The results were analyzed by the chi-square test, the t-test, log-linear analysis, two-way ANOVA, two-way repeated ANOVA that were performed with SPSS 12.0 version. CART decision tree were performed with Clementine 7.2 version. It was found out that the mean number of diagnoses, procedures and LOS of DRG 12701 was statistically different from DRG 12702 (p< 0.0001). The more patients with the procedure “mechanical ventilation and hemodialysis” were assigned into DRG12701 (odds ratio, 4.18, p<0.0001). After controlling for the other covariates, CART decision tree showed that the determinants of difference in revenue were LOS and the procedures “mechanical ventilation and hemodialysis.” More than 68 per cent of DRG 127 patients were cared at regional hospitals and medical centers. In terms of negotiation protocol of DRG 12701 for hospitals, protocol 1 was appropriate among three protocols if the mean of difference in revenue per admission of hospital was positive, otherwise protocol 2 was appropriate for hospitals. However, the appropriate protocol was determined by hospital accreditation as to DRG 12702, protocol 2 was appropriate for tertiary hospitals; protocol 1 was appropriate for district, district teaching, and regional hospitals. Although Tw-DRGs 3rd version had evolved to explain the quality of hospitals, this result demonstrated that tertiary, regional, and district teaching hospitals for treating DRG 127 patients had difficulty in obtaining appropriate reimbursement. We conclude therefore that while changing to the Tw-DRGs system, the lack of taking into account of the severity of heart failure was associated with negative financial impact of hospitals. Of all the different hospital levels, public medical centers have the highest negative difference in revenue. Measures of severity of illness need to be reviewed and incorporated for Tw-DRGs refinement to ensure equitable funding to tertiary, regional, and district teaching hospitals.
Key words: Tw-DRGs, heart failure, CART decision tree

目錄
第一章 緒論 1
第一節 研究背景與動機 1
第二節 研究目的 4
第三節 研究特色與貢獻 5
第四節 名詞解釋 6
一、診斷關聯群 6
二、心臟衰竭 6
三、DRG12701及DRG12702 6
四、收益差 6
五、決策樹 7
六、分類迴歸樹CART(Classification and Regression Tree) 7
第二章 文獻探討 9
第一節 各國DRGS診斷關聯群支付制度之實施 9
第二節 DRGS診斷關係群支付制度之優缺點 10
第三節 DRGS診斷關係群支付制度所造成之衝擊 11
第四節 醫療資源耗用之影響因素 14
一、性別 14
二、年齡 14
三、疾病特性及其型態 15
四、醫院特性 16
第五節 DRGS診斷關聯群支付制度對DRG127心臟衰竭與休克之影響 17
第三章 研究方法 20
第一節 研究架構 20
第二節 資料來源與前置處理 22
第三節 資料分析 28
一、統計分析 28
二、CART (分類迴歸樹) 建模 28
三、分析軟體工具 30
第四章 研究結果 31
一、DRG127研究案件病人特質、臨床與醫院特性 31
二、各權屬層級醫院DRG127方案一、方案二、方案三平均每住院人次收益差分析 34
三、DRG12701收益差CART決策樹分析 52
四、DRG12702收益差CART決策樹分析 66
第五章 結論與建議 80
第一節 研究結論 80
一、研究個案特性 80
二、各權屬層級醫院之財務衝擊 81
三、Tw-DRGs第三版支付制度之方案比較 82
第二節 研究限制 85
第三節 研究建議 85
一、臨床路徑導入 86
二、CART決策樹有助於醫院策略之制訂 86
三、Tw-DRGs應納入疾病嚴重度及醫院教學任務之影響因素 86
第四節 後續研究方向 87
一、擴大不同DRG之研究個案 87
二、增加塑模工具的使用 87
三、增加輸入變數 87
四、縱貫面分析 87
五、心臟衰竭疾病嚴重度分級 88


表目錄
表1-1 2004年MDC5循環系統疾病前十個DRGs排行 4
表3-1建模變數定義 29
表4-1 DRG127研究個案病人特質、臨床與醫院特性資料摘要 33
表4-2 各權屬層級醫院DRG12701方案一、方案二與方案三
平均每住院人次收益差值 36
表4-3 DRG12701各權屬層級醫院方案平均每住院人次收益差成對比較 37
表4-3 DRG12701各權屬層級醫院方案平均每住院人次收益差成對比較(續) 38
表4-3 DRG12701各權屬層級醫院方案平均每住院人次收益差成對比較(續 39
表4-3 DRG12701各權屬層級醫院方案平均每住院人次收益差成對比較(續) 40
表4-3 DRG12701各權屬層級醫院方案平均每住院人次收益差成對比較(續) 40
表4-4 DRG12701各方案各權屬層級平均每住院人次收益差事後檢定顯著性 42
表4-5 各權屬層級醫院DRG12702方案一、方案二與方案三
平均每住院人次收益差值 45
表4-6 DRG12702各權屬層級醫院方案平均每住院人次收益差成對比較 46
表4-6 DRG12702各權屬層級醫院方案平均每住院人次收益差成對比較(續) 47
表4-6 DRG12702各權屬層級醫院方案平均每住院人次收益差成對比較(續) 48
表4-6 DRG12702各權屬層級醫院方案平均每住院人次收益差成對比較(續) 48
表4-6 DRG12702各權屬層級醫院方案平均每住院人次收益差成對比較(續) 50
表4-7 DRG12702各方案各權屬層級平均每住院人次收益差事後檢定顯著性 51
表4-8 DRG12701方案一收益差CART樹狀結構末梢節點規則集摘要 55
表4-9各權屬層級醫院DRG12701方案一收益差決策樹末梢節點之分佈及平均每住院人次收益差two-way ANOVA檢定 56
表4-10 DRG12701方案二收益差CART樹狀結構末梢節點規則集摘要 59
表4-11各權屬層級醫院DRG12701方案二收益差決策樹末梢節點之分佈及平均每住院人次收益差two-way ANOVA檢定 60
表4-12 DRG12701方案三收益差CART樹狀結構末梢節點規則集摘要 64
表4-13各權屬層級醫院DRG12701方案三收益差決策樹末梢節點之分佈及平均每住院人次收益差two-way ANOVA檢定 65
表4-14 DRG12702方案一收益差CART樹狀結構末梢節點規則集摘要 69
表4-15各權屬層級醫院DRG12702方案一收益差決策樹末梢節點之分佈及平均每住院人次收益差two-way ANOVA檢定 70
表4-16 DRG12702方案二收益差CART樹狀結構末梢節點規則集摘要 73
表4-17各權屬層級醫院DRG12702方案二收益差決策樹末梢節點之分佈及平均每住院人次收益差two-way ANOVA檢定 74
表4-18 DRG12702方案三收益差CART樹狀結構末梢節點規則集摘要 78
表4-19各權屬層級醫院DRG12702方案三收益差決策樹末梢節點之分佈及平均每住院人次收益差two-way ANOVA檢定 79

圖目錄
圖3-1 資料探勘過程 20
圖3-2研究架構 ---------------------------------------------------------------------------------------- 21
圖3-3研究樣本選取流程圖 27
圖4-1 DRG12701方案一收益差CART樹狀結構 54
圖4-2 DRG12701方案二收益差CART樹狀結構 58
圖4-3 DRG12701方案三收益差CART樹狀結構 63
圖4-4 DRG12702方案一收益差CART樹狀結構 68
圖4-5 DRG12702方案二收益差CART樹狀結構 72
圖4-6 DRG12702方案三收益差CART樹狀結構 77

參考文獻:
Ⅰ中文文獻
中央健康保險局。(西元2006年3月10日)。健保局財務狀況，中央健康保險局網站。見 http://www.nhi.gov.tw。
于美德（2002 ）。心導管檢查醫療資源耗用分析 。 碩士論文。台北：國立陽明大學醫務管理研究所 。
田月枝（2000）。市場競爭對醫療費用的影響—以全膝關節置換術為例。碩士論文。台北：國立陽明大學醫務管理研究所。
江東亮、文羽苹、謝瓊慧（2000）。醫師與病床供給對醫療費用總額之影響。行政院衛生署八十九年度委託研究計畫期末報告。
余秘a（2002）。住院醫療服務市場之供給誘發需求現象探討。碩士論文。台北：國立台灣大學衛生政策與管理研究所。
吳仁佑、葉玲玲、閻愷正、黃達夫（2002）。甲狀腺切除論病例計酬之甲狀腺癌與良性甲狀腺疾患醫療資源耗用分析。台灣衛誌，21(6)：387-393。
吳明彥（ 2005）。 德國健保總額與台灣醫院總額制度之省思 。
台北：95年醫院協會醫院總額專題研討會 。
吳旭志、賴淑貞譯（2001）。 資料探勘現論與實務，顧客關係管理的技巧與科學。
李芳年（2000）。急診病人的資源耗用---以"台北某區域教學醫院為例”。碩士論文。台北：國立陽明大學醫管所。
金家玉 （2001）。市場競爭對醫院產出表現的影響。碩士論文。台北：國立陽明大學公共衛生研究所。
張樂心、鄭守夏、楊銘欽、江東亮（2004）。醫院權屬別與醫療收費。台灣衛誌，23(2)：130-140。
莊念慈、黃國哲、釧阞Y、郭乃文、魏中仁（2004）。醫院因應總額支付制度之策略方案及其相關因素探討。台灣衛誌，23(2):150-158。
陳佳琪 （1998）。健保論病例計酬門診手術與住院手術醫療資源耗用分析--以疝氣修補術為例。碩士論文。台中:中國醫藥學院醫務管理研究所。
陳婉茗、吳肖琪 （2005）。模擬DRGs 實施對我國醫院的財務衝擊。台灣衛誌 ，24(4)：306-314。
曾憲雄、蔡秀滿、蘇東興、曾秋蓉（2005）。資料探勘。台北旗標出版股份有限公司。
黃勝雄（1999）。論病例計酬制實施前後對醫療資源利用之影響-以某區域醫院閉鎖性骨折開放性復位術為例。．碩士論文。台中:中國醫藥學院醫務管理研究所。
楊志良等(1994) 。全民健康保險有關醫療行為規制問題探討（POD）,台北：研考會。
楊志良（1995）。健康保險，台北巨流圖書公司。
劉夷生(1999)。全民健康保險實施論病例計酬對醫療服務之影響-以前列腺切除術為例。碩士論文。台中:中國醫藥學院醫務管理研究所。
蔡依珍、黃偉堯（2004）。醫院對診斷關聯群之因應行為--以美國為例。醫院期刊，37(3)：23-31。
簡德年（2001）。 智慧資本構面下企業危機診斷模式之建構-類神經網路，分類迴歸樹與鑑別分析方法之應用。碩士論文。台北：國立台北大學商業自動化與管理研究所。
Ⅱ英文文獻
Anselmino, M., Petrelli, A., Gnavi, R., Bobbio, M. & Trevi, G. (2005). Hospital admissions for congestive heart failure in Piedmont, Italy:1996-2001 trend. Italian Heart Journal Supplement, 6(1):42-52.
Arnold, J.M.O., Liu, P., Demers, C., Dorian, P., Gianetti, N., Haddad, H., Howlett, J., Ignazewski, A. & Jong, P. (2006). Canadian cardiovascular society consensus conference recommendations on heart failure 2006: diagnosis and management. Canadian Journal of Cardiology, 22(1):23-45.
Brawled, E., Fuci, A.SO., Kasper, DO.I., Hauser, SO.I., Longo, DO.I. & Jameson, I.I. (2001). Disorders of the heart: heart failure. Harrison’s Principle of Internal Medicine (15th edition). New York: McGraw-Hill. Vol.1:1318
Brunnelli, C., Spallarossa, P., Pasdera, A., Bezantge, G.P., Zorzet, F. & Rossettin, P. (1998). Treatment aspects of Unstable Angina. Costs and payments for DRG. Cardiologia, 43(1):67-75.
Chazan, J.A., London, M.R. & Pono, L.(1992). The impact of diagnosis-related groups on the cost of hospitalization for end-stage renal disease patients at Rhode Island Hospital from 1987 to 1990. American Journal of Kidney Disease,19(6), 523-25.
Chesney, J.D. (1990). Utilization trends before and after PPS. Inquiry, 27(4):376-81
Feinglass, J. & Holloway J.J. (1991). The initial impact of the Medicare prospective payment system on U.S. health care: a review of the literature. Medical Care Research and Review, 48(1):91-115.
Forgione, D.A., Vermeer, T.E., Surysekar, K., Wrieden, J.A. & Plante, C.A. (2004). The Impact of DRG-Based Systems on Quality of Health Care in OECD Countries. Journal of Health Care Finance, 31(1): 41-54.
Furstenberg, T., Bunzemeier, H., Roeder, N. & Reinecke, H. (2005). Cardiovascular medicine in the German diagnosis-related group (G-DRG) system 2005. Zeitschrift für Kardiologie, 94(5):360-70.
Gilman, B.H. (2000). Hospital response to DRG refinements: the impact of multiple reimbursement incentives on in-patients length of stay. Health Economics, Vol.9:277-94.
Heerey, A., McGowan, B., Ryan, M. & Barry, M. (2002). Microcosting versus DRG in the provision of cost estimates for use in pharmacoeconomic evaluation. Excerpt Rev. Pharmacoeconomics and Outcomes Research., 2(1):29-30.
Ho, K.K., Anderson, K.M., Kannel, W.B., Grossman, W. & Levey, D. (1993). Survival after the onset of congestive heart failure in Framingham Heart Study subjects. Circulation, 88:107-15.
Hoffman, J.M., Doctor, J.N., Chan, L., Whyte, J., Jha, A. & Dikmen, S. (2003). Potential impact of the new Medicare prospective payment system on reimbursement for traumatic brain injury inpatient rehabilitation. Archives of Physical Medicine and Rehabilitation, 84(8):1165-72.
Hostetter, T.H. (2004). Chronic kidney disease predicts cardiovascular disease. New England Journal of Medicine, 351(13):1344-46.
Hughes, J.S., Lichtenstein, J. & Fetter, R.B. (1990). Procedure codes: potential modifiers of diagnosis-related groups. Health Care Finance Review., 12(1):39-46.
Harjai, K.J., Boulous, L.M., Smart, F.W., Turgut, Y., Krousel Wood, M.A., Stapleton, D.D., Mehra, M.R., Murgo, J.P. & Ventura, H.O. Clinical Outcomes Following Hospitalization for Heart Failure. The American Journal of Cardiology, Vol. 82:82-85.
Kugelmass, A.D., Cohen, D.J., Brown, P.P., Simon, A.W., Becker, E.R. & Culler, S.D. (2006). Hospital resources consumed in treating complications associated with percutaneous coronary interventions. American Journal of Cardiology, 97(3):322-27.
Louis, D.Z., Yuen, E.J., Braga, M., Cicchedtti, A., Rabinowitz, C., Laine, C. & Gonnella, J.S. (1999). Impact of a DRG-based hospital financing system on quality and outcomes of care in Italy. Journal of Health Services Research and Policy, 34(1 Pt 2):405-15.
Lynk, W.J. (2001). One DRG, One Price? The Effect of Patient Condition on Price Variation Within DRGs and Across Hospitals. International Journal of health Care Finance and Economics, 1(2)：111-37.
McCarthy, E.J. (1999). Provider-owned health plans: El Dorado or Armageddon? Healthcare Financial Management, 53(11):35-38.
Montagne, O., Chaix, C., Harf, A., Castaigne, A. & Durand-Zaleski, I. (2000). Costs of Acute myocardial infarction in a tertiary center and nationwide in France. Pharmacoeconomics,17(6):603-09.
Munoz, E., Chalfin, D., Birnbaun, E., Mulloy, K., Cohen, J. & Wise, L. (1989). Hospital costs, resource characteristics, and the dynamics of death for hospitalized patients in Cardiology Diagnosis-Related Groups. Heart Lung, 18(2):164-71.
Omachonu, V.K., Suthummanon, S., Akein, M. & Asfour, S. (2004). Predicting length of stay for Medicare patients at a teaching hospital. Health Services Management Research, 17:1-12.
Pol, L.G. & Thomas, R.K. (2001). The demography of health and healthcare, 2nd edition, Plenum Press, New York.
Reinecke, H., Bunzdemeier, H., Furstenberg, T., Rothenburg, M., Bocker, D., Scheld, H.H., Breithardt, G. & Roeder, N.(2003). Evaluating the first German diagnosis-related groups (G-DRG) in cardiological patients: problems in the correct medical and economic grouping. Zeitschrift Fur Kardiologie, 92(7):581-94.
Reis, S.E., Holubkov, R.,Edmundowicz, D., McNamara, D.M, Zell, K.A., Detre, K.M. & Feldman, A.M. (1977). Treatment of patients admitted to the hospital with congestive heart failure: specialty-related disparities in practice patterns and outcomes. Journal of the American College of Cardiology, 30 (3):733-38.
Rich, M.W. & Freedland, K.K. (1988). Effect of DRG on three-month readmission rate of geriatric patients with congestive heart failure. American Journal of Public Health, 78(6):680-82.
Rosenthal, G.E. & Landefeld, C.S. (1993). Do older Medicare patients cost hospital more? Evidence from an academic medical center. Archives of Internal Medicine, 153(1):89-96.
Rosko, M.D. & Carpenter, C.E. (1994). The impact of intra-DRG severity of illness on hospital profitability: implications for payment reform. Journal of Health Politics, Policy and Law, 19(4),729-51.
Sumner, A.T. & Moreland, C.C. (1995). The potential impact of diagnosis elated group medical management on hospital utilization and profitability, Health Care Management Review, 20(2):92-100.
Taheri, P.A., Butz, D.A., Dechert, R. & Greenfield, L.J. (2001). How DRGs hurt academic health systems. Journal of American College of Surgery, 193(1):1-8.
Vergnenegre, A., Drust-Cabana, M. & Rymer, R.(1999). Interventional radiology medical applications of information systems: What consequences for the radiology department? Journal of Radiology, 80(6):565-68.
Wickramasinghe, N., Gupta, J.N.D. & Sharma, S.K. (2005). Creating knowledge-based Healthcare Organizations. Business Process Management Journal.
Yang, M.H., Lee, Y.Y., Tsai, L.M., Fan, C.P. & Lin, C.Y. (2006). Possible impact of the third Taiwan DRGs on cardiac arrhythmic patients in tertiary hospital. Health for all-challenges of hospitals in 21st century, International hospital federation Asian Pacific regional conference, abstract book, p.18.
Yasunaga, H., Ide, H., Imamura T. & Ohe, K. (2005). Impact of the Japanese Diagnosis Procedure Combination-based Payment System on Cardiovascular Medicine-related Costs. International Heart Journal, 46(5):855-66.
Yoong, S.L. (2001). Casemix reimbursement: A Singapore Children’s hospital perspective. Annals Academy of Medicine Singapore, 30(4 supplement):36-40.
Yoshikawa, A., Bhattacharya, J. & Vogt, W.B. (1996). Health Economics of Japan. Tokyo: University of Tokyo Press