參加率及大腸鏡轉介率影響糞便潛血免疫化學法篩檢之成本效益分析

Abstract

背景：根據台灣大腸直腸癌篩檢例行性報表指出，台灣歷年糞便潛血免疫化學法陽性個案轉介率大約65-75%不等，顯示每年約有30-40%的人糞便潛血免疫化學法是陽性且未進行大腸鏡確診，疾病篩檢是一整流程的健康介入工作，非單一單位決定後續所產生的結果。因此層層過程中每一個環節皆會影響到國家整體的篩檢成效，甚至是成本的支出會隨之變動。本研究將針對國家大腸直腸癌篩檢執行過程中最直接因素：篩檢涵蓋率及大腸鏡檢轉介率兩個因素，因此兩項皆是大規模大腸直腸癌篩檢過程中扮演相當重要角色。

材料和方法：本研究以我國2004-2009年資料為基礎所產生之參數，以模擬決策樹之方式估計篩檢涵蓋率及陽性個案大腸直腸癌轉介率之成本效益分析，並同時以敏感度分析進行兩個參數變動時對於該大腸直腸癌篩檢成本效益之影響。以台灣2004-2009年50-69歲為應參加篩檢固定世代，共計5417699人應該接受篩檢。根據大腸直腸癌症篩檢登記系統，2004-2014年共計有3074538人至少完成一次糞便潛血免疫化學法篩檢，篩檢涵蓋率(coverage rate)為56.7%。我們以Markov cycle tree應用於成本效益分析(CEA)。

統計分析：本研究以我國2004年50-69歲大腸直腸癌個案(包含篩檢及非篩檢大腸直腸癌個案)，且追蹤至2014年12月31日，以獲得我國大腸直腸癌10年存活率，並使用TreeAge軟件構建了Markov cycle tree進行大腸直腸癌自然病史及篩檢介入過程，將大腸癌早期和晚期的存活率納入篩檢程序結果的模擬中。

研究結果：可以發現雖然兩種情況之下（參與率和轉介率）都是約42%的人真正得到篩檢實質上的好處，亦即在增量效益(X軸)的散佈兩者相當，但成本卻是參與率的影響較大，節省的幅度較大。此狀況在參與率60%、轉介率80%與參與率80%、轉介率60%，兩種情況下之間的比較又再明顯一些。

結論：依本研究結果得知，參與率比陽性轉介率更具成本效益，建議衛生主管機關及醫療服務者：(1)在健保資源有限的情況下，擴大民眾參與，早期發現早期治療，以達到節制成本的目的。(2)建立參與率KPI，關鍵績效指標（Key Performance Indicators，KPI）目標策略，以提升病患照護品質。

BACKGROUND：According to the Taiwan Colorectal Cancer Screening Report, the referral rate of positive fecal immunochemical testing (FIT) in Taiwan is between 65%-75%, indicating that there are 30%-40% of individuals who did not submit to a colonoscopy confirmation examination following a positive FIT result. Disease screening is part of a systematic health intervention procedure and not an outcome that was determined by an isolated department. Therefore, each segment within the layered health mechanism would impact the screening outcome of the entire program, and even cause cost and expenditure to fluctuate. This study focused on the direct factors in the population-based colorectal cancer screening implementation processes, namely the coverage rate and the colonoscopy confirmation referral rate. Therefore, the two factors are significant in terms of a large-scale colorectal cancer screening program.

MATERIALS AND METHODS：This research used the parameters generated from Taiwanese colorectal cancer screening database between 2004 and 2009 as a basis and adopted a simulated decision tree approach to conduct a costbenefit analysis and estimate the coverage rate and the referral rate of positive cases of colorectal cancer. Also, this research used sensitivity analysis to study the effects on the cost and benefit of colorectal cancer screening with two parameters. This study identified 5,417,699 individuals who ought to have received screening among aged 50-69 population in 2004-2009, which are eligible for screening invitation. According to the colorectal cancer screening registry system, a total of 3,074,538 individuals completed at least one fecal immunochemical testing (FIT) in 2004-2014, with a coverage rate of 56.7%. The Markov cycle tree was applied for cost-effectiveness analysis (CEA).

STATISTICAL ANALYSIS：According to the scenario of colorectal cancer screening in Taiwan, the population aged 50-69 in 2004 following the screening processes and simulate their conditions up to the end of 2014 using those parameters based on the results of Taiwanese screening program. The transition probabilities of seven-state colorectal cancer natural history model were estimated based on screening outcome 2004-2009 and followed up to 2014, which were employed to construct the Markov cycle tree using TreeAge software . The Survival rate of colorectal cancer by early and late stages were incorporated into simulation of the outcome of screening program

RESULTS：The results show that, under both circumstances (attendance rate and referral rate), 42% of cases benefited from the screening procedure. In other words, in terms of the distribution of incremental benefit (X-axis), both circumstances are similar, but in terms of cost, there is a more significant impact on screening coverage rate. This outcome is even more apparent when the screening attendance rate is at 60% and the referral rate at 80%, or when the screening attendance rate is at 80% and the referral rate at 60%.

Conclusions： According to the results of this study, attendance rate was more cost-effective than referral rate. It is suggested that health administration and medical service providers perform the followings: (1) In case of limited health insurance resource, it is imperative to encourage public participation in early detection and early treatment to achieve the purpose of cost control; (2) A KPI strategy to achieve certain participation rate should be established to improve patient care quality.