大腸直腸癌 以病患為基礎的品質指標遵從度與長期存活關係的探討

Relationships between Case-based Quality Measures and Long-term Survival for Colorectal Cancer Patients

張原肇(Yun-Jau Chang)

指導教授 ： 鍾國彪

共同指導教授 ： 賴美淑(Mei-Shu Lai)

國立臺灣大學/公共衛生學院/健康政策與管理研究所/博士(2011年)

https://doi.org/10.6342/NTU.2011.01209

摘要

背景: 在台灣，大腸直腸癌近年來已高居癌症死亡率第三位，2009年一整年即有4531人死於大腸直腸癌，是一重要公共衛生議題。而癌症照護品質也近十年來受到重視，但醫療從業人員對品質測量是否能改善病患結果，常持懷疑態度。台灣大腸直腸癌核心測量指標已經完成，核心測量指標是否能正確預測病患預後仍有待驗證。
目的: 探討大腸直腸癌的長期存活因素，和探討個別品質測量指標或綜合分數(composite process score)與病患5年存活率的相關程度，並研究影響淋巴腺(其中個別指標)的影響因素。
資料和方法: 我們收集2003年到2004年台灣癌症診療資料庫(TCDB)內被診斷為大腸直腸癌新發個案且接受大腸切除術的病人(Cohort 1 和子群 Cohort 2、Cohort 3與Cohort 4) 和一家地區醫院(Cohort 5，非TCDB組)的大腸直腸癌的手術病人，最重要的觀察點是的5年存活率(包括全體、無疾病和疾病別存活率)；次要觀察點是取出且檢查的淋巴結的數目。首先在準備階段，需要選定的是一組合適的品質指標(我們以已發表的十七個核心測量指標為藍圖)。而最終選擇的十一個指標是由二回合十三人次的專家會議決定，本文所稱的指標即指此。包含這些被選出指標(十一個)的專家問卷調查表，被郵寄到事先邀請且首肯的全國各個醫療專家。收集回信後，便可計算指標權重(使用於本研究的分析層級程序法，及模糊分析層級程序法)。本研究所討論的五個主要議題包括:1.探討醫院因素是否會影響大腸直腸癌病人手術後的5年存活率;2.探討個別測量指標的遵從率與病患個人5年存活率之間的關聯性;3.探討綜合品質分數 (運用六個方法：「全有或無法」、「70% 標準法」、「平均加權法」、「分析層級程序法」、「模糊分析層級程序法」和「主成分分析法」)和病患個人5年存活率之間的關聯性;4.不同門檻的淋巴結檢查數目對病患個人5年存活率之間的影響;5.使用多階層模型，探討醫院因素是否在期別低估上有相關性。
結果: Cohort 1包含在29醫院的 4973位接受大腸直腸癌手術切除的癌症病患，平均年齡65.0歲，Cohort 5由單一醫院的122接受大腸直腸癌手術切除的癌症病人組成，平均年齡68.5歲。在TCDB的病人，不管任何癌症期別，對病患個人5年癌症別存活的預後因素是年齡、性別、共病症、腫瘤腫瘤分化等級。在期別0&I病人，只有年齡、腫瘤腫瘤分化等級、醫|評鑑等級、年度大腸直腸癌切除手術量與病患個人5年癌症別存活有相關。第二期病人，年齡共病症、手術邊緣無殘存腫瘤和適當的淋巴結清除是與病患個人5年癌症別存活有關。在第三期病人， 年齡、性別、手術邊緣無殘存腫瘤和腫瘤分化等級是病患個人5年癌症別存活的預後因素。在Cohort 5病人不論期別，其中腫瘤分化等級、手術邊緣無殘存腫瘤、術前 CEA 血液濃度、住院時間長短與病患個人5年癌症別存活有相互關聯。在個別指標與結果相關性方面，有二個指標(P1, T4)不管在任何組，都與5年存活結果無相關。其它六個指標則與5年存活結果有相關，其中 P2, T3 和 T6 最明顯。綜合品質分數(六種不同的加權方法)也與病患個人5年存活互相有關係。其中「平均加權法」、「分析層級程序法」、「模糊分析層級程序法」和「主成分分析法」可以呈現不錯的相關性，綜合分數越高，病患5年存活率也越好。然而「全有或無法」卻沒有顯著相關性。在淋巴結數目與長期存活相關性方面，我們發現在控制年齡、性別、共病症、癌症病理分期後，我們發現不管門檻設在多少(5-25)，淋巴結數目都與長期存活相關，但門檻設為18時，統計模型配適度最好。而期別低估的多階層模型顯示醫院因素(除了年手術量)不會影響淋巴結的數目。
結論: 對於癌症照護的品質測量不僅僅只是能夠表現照護品質而已，並且可與病患個人5年存活率之間有關聯性，不管討論的目標是個別品質指標或是綜合品質分數。另外，淋巴結檢查數目的模型證明淋巴結檢查數目18對長期存活結果有最佳模型配適度。將來的研究應包括其他癌症照護的品質指標與病患個人存活間的相關性，和醫師、醫院因素與品質測量指標遵從性的探討。

關鍵字

品質測量 ； 大腸直腸癌 ； 綜合過程分數 ； 分析層級程序法 ； 模糊分析層級程序法 ； 主成分分析法 ； 5年存活率 ； 期別低估 ； 多階層迴歸

並列摘要

Background: Performance measurement for medical care has gone through several decades. Only recently, the emphasis extended to cancer care quality. Thogh colorectal cancer ranks third as leading cuase of cancer deaths, relationships between performance measures and patient’s outcomes had never been investigated. In addition, the minimum 12 lymph node is recommended as benchmark of cancer care for colorectal cancer to prevent understaging recent but concern about its adequacy follows.
Objectives: The main five topics discussed in this study encompass: 1. prognostic factors of 5-year survival; 2. association between adherence to individual quality measures and 5-year survival; 3. association between composite process scores (by six different aggregating algorithms) and 5-year survival; 4. impact of different threshold of number of lymph node on 5-year survival; 5. hospital factors influence on understaging.
Materials and Methods: We identified 2003-2004 TCDB patients who were diagnosed with colorectal cancer undergoing bowel resection (Cohort 1 and subgroup Cohort 2, Cohort 3 as well as Cohort 4) and a regional hospital (Cohort 5, non-TCDB group) Primary end point was 5-year survival rate (including overall, disease-free and disease specific survival). The set of quality measures was initially finalized by two rounds of expert meetings. Expert questionnaires including these eleven indicators were mailed to specialists nationwide to get indicator weights (i.e., weight of AHP and Fuzzy AHP in this study), which would be used in the subsequent computation.
Results: Cohort 1 included 4973 colorectal cancer patients undergoing colectomy treated at 29 hospitals with mean age 65.0 years old, and Cohort 5 included 122 colorectal cancer patients with mean age 68.5 years old. For TCDB patients regardless of stages, prognostic factors were age, sex, comorbidity, tumor grade, negative surgical margin, adequate lymphadectomy and accreditation status are prognostic for 5-year disease-specific survival. For stage 0& I of TCDB patients, age, grade, accreditation status and annual surgical volume of colorectal surgery correlated 5-year disease- specific survival. For stage II patients, age, comorbidity, negative surgical margin and adequate lymphadectomy are prognostic for 5-year disease-specific survival. For stage III, age, gender, negative, grade and surgical margin are prognostic for 5-year disease-specific survival. For non-TCDB patients, comorbidity, tumor locations, stages, negative surgical margin and pre-OP CEA, albumin, hemoglobin level as well as hospital stays correlated with 5-year disease-specific survival. When it comes to quality measures, two indicators (P1, T4) didn’t correlate 5-year survival in any Cohort. The others were correlated with patients’ 5-year survival, among which P2, T3 and T6 always work well. Composite process score also correlate 5-year survival by various aggregating algorithms, among which Equal Weighting, AHP, Fuzzy AHP and PCA works well. But the significant association with 5-year survival didn’t happen to rectal cancer patient by any method. Number of lymph node study showed cut-off of 18 provided the most model fitting in association with 5-year survival. Multilevel model study of understaging revealed hospital factors (accreditation status and ownership type) didn’t influence lymph node examined.
Conclusions: Quality measures for cancer care cannot only indicate cancer care quality but also correlate with patients” 5-year survival either by individual or composite quality score. The number of lymph node examined was also associated with patients’ long-term outcome. Further researches include validation of quality measures on other major cancer care and physician as well as hospital factors on adherence to quality measures.

並列關鍵字

quality measures ； colorectal cancer ； composite process scores ； analytic hierarchy process ； fuzzy analytic hierarchy process ； principal component analysis ； 5-year survival rate ； understaging ； multilevel regression

參考文獻

Andre, T., Boni C., Mounedji-Boudiaf L., Navarro M., Tabernero J., Hickish T., et al. (2004). Oxaliplatin, fluorouracil, and leucovorin as adjuvant treatment for colon cancer. New England Journal of Medicine, 350(23), 2343-2351.

Andre, T., Boni C., Navarro M., Tabernero J., Hickish T., Topham C., et al. (2009). Improved overall survival with oxaliplatin, fluorouracil, and leucovorin as adjuvant treatment in stage II or III colon cancer in the MOSAIC trial. Journal of the National Cancer Institute Monographs, 27(19), 3109-3116.

Auerbach, A. D., Hilton J. F., Maselli J., Pekow P. S., Rothberg M. B.Lindenauer P. K. (2009). Shop for quality or volume? Volume, quality, and outcomes of coronary artery bypass surgery. Annals of Internal Medicine, 150(10), 696-704.

Badri, M. A. (1999). Combining the analytic hierarchy process and goal programming for global facility location-allocation problem. International Journal of Production Economics, 62(3), 237-248.

Baxter, N. N., Virnig D. J., Rothenberger D. A., Morris A. M., Jessurun J.Virnig B. A. (2005). Lymph node evaluation in colorectal cancer patients: A population-based study. Journal of the National Cancer Institute, 97(3), 219-225.