# Cost Effectiveness of Colorectal Cancer Screening Interventions with Their Effects on Health Disparity Being Considered

## Article information

Cancer Res Treat. 2016;48(3):1010-1019
Publication date (electronic) : 2015 December 28
doi : https://doi.org/10.4143/crt.2015.279
1Health Insurance Policy Research Institute, Korea National Health Insurance Service, Wonju, Korea
2Department of Preventive Medicine and Institute of Health Services Research, Yonsei University, Seoul, Korea
Correspondence: Eun-Cheol Park, MD, PhD  Department of Preventive Medicine and Institute of Health Services Research, Yonsei University, 50-1 Yonsei-ro, Seodaemun-gu, Seoul 03722, Korea  Tel: 82-2-2228-1862 Fax: 82-2-392-8133 E-mail: ecpark@yuhs.ac
Received 2015 July 26; Accepted 2015 December 16.

## Abstract

### Purpose

The purpose of this study was to evaluate the cost effectiveness of colorectal cancer screening interventions with their effects on health disparity being considered.

### Materials and Methods

Markov cohort simulation was conducted with the cycle/duration of 1/40 year(s). Data came from the results of randomized trials and others. Participants were hypothetical cohorts aged 50 years as of year 2013 in 16 Korean provinces. The interventions until the age of 80 were annual organized fecal occult blood test (FOBT) (standard screening), annual FOBT with basic reminders for provinces with higher mortalities than the national average (targeted reminder) and annual FOBT with basic/enhanced reminders for all provinces (universal reminder 1 and 2). The comparison was non-screening, the outcome was quality-adjusted life years, and only medical costs for screening and treatment were considered from a societal perspective. The Atkinson incremental cost effectiveness ratio (Atkinson ICER), the incremental cost effectiveness ratio adjusted by the Atkinson Inequality Index, was used to evaluate the cost effectiveness of the four interventions with their impacts on regional health disparity being considered.

### Results

Health disparity was smallest (or greatest) in non-screening (or the standard screening). The targeted reminder had smaller health disparity, and smaller Atkinson ICER with respect to standard screening, than did the universal reminder 1 and 2.

### Conclusion

The targeted reminder might be more cost effective than the universal reminders with their effects on health disparity being considered. This study helps to develop promotional effort for colorectal cancer screening with both the greatest cost effectiveness and the smallest health disparity

## Introduction

Colorectal cancer is the fourth most common cause of cancer death in the world (0.69 million, 8.5%) [1]. Early detection and treatment are essential for reducing high mortality from colorectal cancer, given that adenomas and early colorectal cancers are usually small and asymptomatic [2]. However, participation in colorectal cancer screening shows a wide variation across different socioeconomic or regional conditions, e.g., from 53% in Oklahoma to 72% in Delaware in the United States during 2006-2010 [2], from 33% in the most deprived area to 67% in the least deprived area in England for the year 2008 (organized screening) [3], and from 22% in Gyeongnam (a rural province) to 28% in Daejeon (a metropolitan province) in South Korea (Korea hereafter) for the year 2012 (organized screening) [4]. This disparity in colorectal cancer screening might exacerbate disparity in health status, making more contribution to those with higher participation in this intervention.

However, the current cost effectiveness analysis of colorectal cancer screening focuses on improving population health and ignores health disparity generated by this intervention. But recent studies show that the purposes of maximizing population health and minimizing health disparity often contradict with each other: the promotion of a “targeted” reminder designed to increase participation in organized colorectal cancer screening among deprived populations with ethnic diversity minimizes health disparity, whereas the promotion of a “universal” reminder designed to increase participation in organized colorectal cancer screening among the entire population maximizes population health in the United Kingdom [5]. For this reason, several researchers started to develop distributional cost effectiveness analysis (DCEA) of healthcare interventions, a combination of cost effectiveness analysis and health disparity examination.

This study might be the first DCEA of cancer screening interventions in East Asia. Socioeconomic disparity in organized colorectal cancer screening has slightly decreased in Korea since 2005, e.g., 10.3% vs. 16.9% for medicaid vs. insured in 2006, then 22.2% vs. 26.1% in 2012 [4]. However, regional disparity in this intervention still persists in the nation, i.e., from 22% in Gyeongnam (a rural province) to 28% in Daejeon (a metropolitan province) for 2012 [4]. Indeed, Korea has much lower national participation in organized colorectal cancer screening than does the United Kingdom, i.e., 26% in 2012 vs. 52% in 2008 [3,4]. In this context, this study evaluated the cost effectiveness of organized colorectal cancer screening interventions (i.e., screening with various types of reminders), with their impacts on regional health disparity being considered.

## Materials and Methods

### 1. Participant, intervention, comparison, and outcome

Markov cohort simulation was conducted for hypothetical cohorts aged 50 years as of the year 2013 in each of 16 Korean provinces. The interventions until the age of 80 were (1) annual fecal occult blood test (FOBT) (standard screening), (2) annual FOBT with basic reminder letters for eight provinces with higher mortalities from either all causes or colorectal cancer than the national average (i.e., Busan [metropolis], Gangwon, Chungbuk, Chungnam, Jeonbuk, Jeonnam, Gyeongbuk, and Gyeongnam [rural areas]) [6] (targeted reminder), (3) annual FOBT with basic reminder letters for all provinces (universal reminder 1), and (4) annual FOBT with enhanced reminder letters (i.e., personal reminder letters with tailored information packages) for all provinces (universal reminder 2). The comparison was non-screening. The outcome measure was quality-adjusted life years (QALYs). The Markov states were (1) healthy, (2) polyps not detected by screening (P1), (3) polyps detected by screening (P2), (4) symptom-free early colorectal cancer not detected by screening (ECC1), (5) symptom-free early colorectal cancer detected by screening (ECC2), (6) symptomatic early colorectal cancer (ECC3), (7) symptom-free advanced colorectal cancer not detected by screening (ACC1), (8) symptom-free advanced colorectal cancer detected by screening (ACC2), (9) symptomatic advanced colorectal cancer (ACC3), (10) death from colorectal cancer, and (11) death from other causes (the true positive goes through colorectal cancer treatment without further colorectal cancer screening). The length of the cycle was 1 year and the length of the duration was 40 years between the ages of 50 and 90. Only medical costs for screening and treatment were considered from a societal perspective. The “Atkinson incremental cost effectiveness ratio (Atkinson ICER),” the incremental cost effectiveness ratio adjusted by the Atkinson Inequality Index (to be elaborated below) was used to evaluate the cost effectiveness of the four interventions with their impacts on health disparity being considered.

### 2. Model structure

Fig. 1 shows a simplified version of the Markov model for cohort simulation. In each cycle, the healthy can stay healthy, develops polys (P1/2) or dies from causes other than colorectal cancer. Without treatment (polypectomy), one with polyps (1) stays as he or she is, (2) develops early colorectal cancer with the annual polyp-ECC transition rate of 0.005 [2,7,8], i.e., P1 to ECC1, P2 to ECC2, or (3) dies from causes other than colorectal cancer. Without treatment, one with early colorectal cancer (1) stays as he or she is, (2) develops advanced colorectal cancer with the mean ECC dwelling time of 2 years [2,7,8], i.e., ECC1 to ACC1, ECC2 to ACC2, ECC3 to ACC3, or (3) dies from causes other than colorectal cancer. Also, without treatment, the mean sojourn time from symptom-free to symptomatic colorectal cancer (i.e., from ECC1/2 to ECC3, from ACC1/2 to ACC3) is 5 years [9]. A new incidence of polyps is detected by a screening, while a new incidence of colorectal cancer is detected by either a screening or a symptom. The incidence rate of polyps or colorectal cancer is the same between screened and unscreened groups. Suspicious lesions detected by colonoscopy are biopsied and those detected by FOBT receive colonoscopy (and biopsy if applicable). One who receives polypectomy becomes healthy or dies from causes other than colorectal cancer. Patients who receive treatment either stay as cancer patients or die, and those who survived for 5 years after treatment have the same mortality rates with healthy people in the same age group. Table 1 summarizes the parameters, values and references for the model described above.

Simplified version of Markov cohort simulation model. P, polyp; ECC, early colorectal cancer; ACC, advanced colorectal cancer; CC, colorectal cancer.

Parameters, values, and references

### 3. Utility, screening, natural history, and survival after treatment and cost

Utilities for the 11 Markov states came from a systematic review of colorectal cancer utilities [10] and a cost effectiveness analysis of colorectal cancer screening with blood-based biomarkers [11]. Three percent were introduced as annual discount rates for utility. Values on screening-related variables and the natural history of colorectal cancer were obtained from Statistics Korea [6], previous studies on cancer registry data (for Korea [4,12] and for Germany [9]), existing literature on the cost effectiveness of colorectal cancer screening [2,7,8,13], and Bayesian calibration for the natural history of colorectal cancer [14]. Data sources on survival after treatment were randomized trials through 13 and 30 years of follow-up for the cost effectiveness of colorectal cancer screening in the United States [15,16]. The cost of a basic/enhanced reminder letter, FOBT, colonoscopy and biopsy per participant in Korea for 2013 were derived from randomized trials for the cost effectiveness of reminder letters for colorectal cancer screening in the United States and UK for 2005 and 2009, respectively [17,18], the Korea Ministry of Health and Welfare notifications and the Korea Health Insurance Review and Assessment Service guidelines on health insurance medical cost including drug components and materials for medical treatment [19].