台灣非小細胞肺癌介入標靶藥物治療使用的醫療利用及療效分析

Target Therapy, Medical Resource Utilization and Cost Effectiveness Analysis of Non-Small Cell Lung Cancer in Taiwan

王麗桂(Li-Kuei Wang)

指導教授 ： 張肇松

高雄醫學大學/健康科學院/醫務管理學研究所碩士在職專班/碩士(2010年)

摘要

自2004年健保實施標靶藥物Gefitinib(Iressa)、Erlotinib(Tarceva)治療非小細胞肺癌
納入健保給付後，截至目前為止，國內關於標靶藥物癌症治療所使用之全民健保給付
之研究，仍然相當缺乏。目前國內相關研究多使用健保資料庫之費用科目及單一機構
醫院進行分析，然而，如此分析無法有效呈現標靶藥物治療成本之特性，有必要進行
更深入之分析。

本研究採回溯性之縱貫設計，資料源自2004至2006年台灣癌症中心資料庫、2004到
2009年健保資料庫，及2004到2009年承保死亡檔。Kaplan-Meier繪製存活曲線，Cox涉險
模型進行風險校正，評估病患介入標靶藥物治療的存活時間與醫療費用，並針對效果部
分進行直接醫療成本效果分析。

本研究的研究樣本共17,451位非小細胞肺癌病患。在性別方面，男性11,064(63%）、
女性6,387位（37%），年齡層以40-69歲8,598位(49%)最多，發現時已經是晚期第三期B3,713位 (21%)，第四期9,648位(55%)，達7成以上為癌症末期患者。男性仍以癌症晚期占49%最多，女性也有29%之多。介入標靶治療模式為2,996位，其平均住院天數為27天，門急住診總費用為台幣1,583,551元，總藥費1,022,884元，含標靶藥費420,307元，傳統治療門急住診總費用為572,911元，總藥費202,038元，而無積極治療分別為215,680元、39,912元；整體存活時間，介入標靶平均存活天數為825天，傳統治療464天及無積極治療220天，標靶治療1年存活率為92%，傳統治療46%及無積極治療18%。

線性複迴歸分析顯示在門住診醫療總費用，影響因子為治療模式、年齡、期別及共存疾病指標(CCI)，經校正後，整體解釋力為49.4%。截至2009年12月31日止，Cox等比例危害函數分析，顯示共存疾病指標(CCI)每增加1分，其死亡風險增加11%，女性相對於男性的死亡風險只有0.68倍，傳統治療死亡風險相對於介入標靶治療高出2.23倍，而無積極治療其死亡風險增加5.22倍；在癌症期別狀態之危害函數當中，顯示非小細胞肺癌晚期的患者其死亡風險是早期患者的7.5倍，年齡越大死亡風險越高，標靶治療與傳統療的成本效果差異比為每日2,799元。

本研究顯示影響非小細胞肺癌患者的醫療成本因子為期別、年齡、治療模式與共存疾病指標(CCI)；在成本效果研究顯示，需要較多的醫療成本，才能延長1年的存活時間，成本效果差異亦符合臨床實證醫學上，在不同治療模式間的治療效果差別，足見以醫療經濟觀點的成本效果估算，可有效佐證臨床醫療及醫療經濟兩者間的關連性。

關鍵字

非小細胞肺癌 ； 標靶藥物 ； 存活分析 ； 台灣癌症中心資料庫

並列摘要

The National Health Insurance (NHI) has included drugs used for target therapy to
treat Non-Small Cell Lung Cancer (NSCLC) such as Gefitinib (Iressa) and Erlotinib (Tarceva) in the reimbursement list since 2004. However, research evaluating what should be included in the reimbursement is still relatively poor. The current studies to this aspect primarily use the itemized charges and individual hospitals in the NHI database for the analysis, yet they frequently fail to reflect the average cost characteristics of target therapy for cancer treatment, which therefore arouse the need for further investigation.

The present study adopts the cross-sectional design of retrospective studies, which encompass The data is from Taiwan Cancer Database(TCDB) from 2004 to 2006, National Health Insurance Research Database(NHIRD) from 2004 to 2009, and Taiwan’s Death Registries from 2004 to 2009. Kaplan-Meier analysis was used to construct survival curves with Cox proportional hazard model to evaluate target survival day with the spent cost, and measure the direct effectiveness of the invested medical cost.

The samples included in the current research are 17,451 NSCLC patients, of which 11,064 (63%) are male and 6,387(37%) are female. The most common ages range from 40 to 69, which take 8,598 (49%) out of the total. 3,713(21%) are diagnosed stage III, with 9,648 (55%) are stage IV. Overall, more than 70% of the selected samples are at final stages, of which the male (49%) are apparently more than the female (29%). There are 2,996 patients participating in target therapy, with total hospitalization of 27 days, total clinical visit and emergency cost of $NT 1,583,551, and total drug expense of $NT 1,443,192, including those for target therapy of $NT 420,307. When the total survival days are compared, target survival days are 825 days, whereas those with conventional therapies are 464 days and those with non-aggressive therapies are 220 days. Prognosis with target therapy is 92%, compared to that of conventional therapies 46% and non-aggressive therapies 18%, respectively.

The multiple linear progression analysis indicates that, with the impact factors of treating mode, age, stage, and Charlson Comorbidity Index (CCI), the gross explanation power of total clinical visit and emergency cost reaches 49.4% after calibrated. From the functional analysis of Cox proportional hazard model that includes data collected until December 31st, 2009, it shows that each increment of CCI results in 11% increase of death risk. The death risk of the female is 0.68 of that of the male. In terms of treating modes, the death risk of conventional therapies is 223% and that of non-aggressive ones is 522% higher than that of target therapy, respectively. In the hazard functions of stage, the death risk of patients at late stages is 7.5 folds of that at early stages, which is also proportional to the ages. Finally, the difference of cost effectiveness between target therapy and conventional therapies is $NT 2,799 each day.

In conclusion, the present study shows that the major impact factors of the cost of medical care for NSCLC patients are stage, age, treating mode, and CCI. The analysis of cost effectiveness indicates that each survival year to be extended requires significantly more cost of medical care. The difference of this matter also matches the outcome of different treating modes that has been observed in clinical medicine. It then postulates that the relationship between clinical medicine and medical economics can be effectively validated when medical cost effectiveness is estimated under the viewpoint of medical economics.

並列關鍵字

Non-Small Cell Lung Cancer (NSCLC) ； Target Therapy ； Survival

參考文獻

1. Parkin DM, Bray F, Ferlay J, Pisani P. Global cancer statistics, 2002. CA Cancer J Clin 2005;55:74-108.

2.行政院衛生署：衛生統計系列(一)死因統計。http://www.doh.gov.tw/statistic/data/衛生統計叢書2/95/上冊/index.htm。引用2007/12/08。

3. WHO. Statistical Information System (WHOSIS).Available at: http://www.who.int/whosis/en/index.html. Accessed March 18, 2008.

4.賴美淑、邱秀渝、唐秀治、鍾國彪、謝碧晴：癌症中心品質提升評估整合型

研究計畫期末報告。行政院衛生署國民健康局委託研究計畫，計畫編號