Trends in breast cancer screening rates among Korean women: results from the Korean National Cancer Screening Survey, 2005-2020

OBJECTIVES Since 2002, the Korean government has provided breast cancer screening as part of the National Cancer Screening Program. This study reported trends in the screening rate among Korean women from 2005 to 2020, including organized and opportunistic screening for breast cancer. METHODS Data from the Korean National Cancer Screening Survey, an annual cross-sectional nationwide survey, were collected using a structured questionnaire between 2005 and 2020. The study population included 23,702 women aged 40-74 years with no history of cancer. We estimated the screening rate based on the current recommendation of biennial mammographic screening for breast cancer. In addition, a joinpoint trend analysis was performed for breast cancer screening rates among various subgroups. RESULTS In 2020, the breast cancer screening rate was 63.5%, reflecting an annual increase of 7.72% (95% confidence interval 5.53 to 9.95) between 2005 and 2012, followed by non-significant trends thereafter. In particular, a significant decrease in the breast cancer screening rate was observed in the subgroups aged 50-59 years old, with 12-15 years of education, and living in rural areas. CONCLUSIONS Although there has been substantial improvement in breast cancer screening rates in Korean women, the trend has flattened in recent years. Therefore, continual efforts are required to identify subgroups with unmet needs and solve barriers to the uptake of breast cancer screening.


INTRODUCTION
Trends in breast cancer screening rates among Korean women: results from the Korean National Cancer Screening Survey, 2005-2020 breast cancer mortality [3]. In Korea, both organized and opportunistic breast cancer screening programs have been developed. Since 2002, the National Cancer Screening Program (NCSP) in Korea has provided free or low-cost screening for breast cancer [4]. Individuals enrolled in the Medical Aid program and National Health Insurance (NHI) beneficiaries can receive services free of charge if their premium scale is less than 50%. In contrast, NHI beneficiaries in the top 50% of the premium scale pay 10% of the screening cost. Conversely, individuals who use opportunistic cancer screening programs pay all procedural expenses [2]. Although the Korean government provides organized cancer screenings, many Koreans still undergo opportunistic screenings despite the out-of-pocket costs [5].
Under the National Cancer Screening Guideline for breast cancer screening revised in 2015, the NCSP provides mammograms every 2 years for women aged ≥ 40 years [6,7]. Conversely, opportunistic breast cancer screening and the frequency and screening method used depend on individuals' choices and their doctors [4]. In this study, based on the national recommendations for breast cancer screening, we report the overall breast cancer screening rate from both organized and opportunistic screening in Korea. We also analyzed the changing trends in screening rates by various demographic and socioeconomic factors that might be associated with breast cancer screening.

Data source and study population
Since 2004, the National Cancer Center has conducted the Korean National Cancer Screening Survey (KNCSS), an annual national survey on screening behaviors and related factors for stomach, colorectal, liver, breast, and cervical cancers. The eligible population for the KNCSS comprised cancer-free men aged 40-74 years and cancer-free women aged 30-74 years, while cancer-free women aged 20-29 years have also been surveyed since 2015. The subjects were sampled based on the Resident Registration Population data using a stratified, multistage, random sampling procedure according to the residential area, gender, and age, to ensure that the survey participants were a representative population sample. The Korea National Statistical Office publishes the Resident Registration Population data annually after data are gathered from residents of the registration population on December 31 each year. Details of the sampling method have been described fully elsewhere [8]. We used the KNCSS data collected through face-to-face interviews, except for 2004, when data were collected via computer-assisted telephone interviews. Specialized research institutions recruited participants through in-person contact. At least 3 contact attempts were made for each household, and 1 person was selected from each household. All participants were fully informed and consented to participate in the survey for public purposes. According to the protocol of the NCSP, women aged > 40 years are eligible to undergo breast cancer screening. Therefore, this study included cancer-free women aged 40-74 who had participated in the KNC-SS between 2005 and 2020. A total of 23,702 women were included in the final analysis (Supplementary Material 1).

Variable definition
The KNCSS survey collected information on the history of screening for 5 types of cancer (stomach, liver, colorectal, breast, and cervical cancer) and socio-demographic characteristics using a structured questionnaire. The major questions about the participant's cancer screening experiences are provided in Supplementary Material 2. The questions were as follows: "Have you ever undergone (cancer type) screening?", "Which screening method was used?", "When did you last undergo (cancer type) screening with this method?", and "How did you pay for the screening?" Based on the screening payment method, the screened respondents were classified according to whether they underwent organized or opportunistic screening. If the payment method was any of the following, the respondent was considered to have undergone organized screening: "National Health Insurance/National Health Insurance Corporation (partially self-paid), " or "Public health center/government (total free of charge). " Respondents with other payment methods were considered to have undergone opportunistic screening. Socio-demographic factors such as age, monthly household income level, education level, and location of residence were considered. The age groups were 40-49 years, 50-59 years, 60-69 years, and 70-74 years. The monthly household income level was classified as low, middle, or high, according to the annual tertile. The educational level was divided into 3 groups according to the number of years of education received: ≤ 11 years (less than high school graduate), 12-15 years (high school graduate to less than 4-year college graduate), and ≥ 16 years (4-year college graduate or more). Finally, residential areas were classified as metropolitan, urban, or rural.

Statistical analysis
To analyze the breast screening rates according to the recommendations, we estimated the proportion of respondents who reported having had a mammographic screening within the past 2 years, according to the NCSP Breast Cancer Screening Protocol (Table 1). To yield unbiased estimates, we calculated the breast cancer screening rates using survey sampling weights. Changes in breast cancer screening rates were estimated using joinpoint regression [9]. Both the APC and average annual percent change (AAPC) were calculated with 95% confidence intervals (CIs). The results were summarized as APCs using a linear model of each screening rate's raw values. The AAPC is based on the weighted average of the APC, calculated across the joinpoints. Breast cancer screening trends were also calculated across the selected demographic and geographic subgroups. We adopted a maximum of 2 possible joinpoints for all analyses and selected the model that best explained the data trends [10]. Based on these results, we characterized breast cancer screening trends. The trend was described as stable if there was a change of 0.5% or less per year (-0.5% ≤ APC ≤ 0.5%) and if the APC was not statistically signifi-cant. If the APC changed by ≥ 0.5% per year (APC < -0.5% or APC > 0.5%) and the APC was not statistically significant, it was described as a non-significant change. A change with a statistically significant APC > 0 was described as increasing. A change to a statistically significant APC < 0 was described as decreasing [11]. Statistical analysis was performed using SAS version 9.4 (SAS Institute Inc., Cary, NC, USA) and Joinpoint version 4.8.0.1 (National Cancer Institute, Bethesda, MD, USA).

Ethics statement
This study was approved by the Institutional Review Board of National Cancer Center (IRB approval No. NCC2019-0233). All participants were fully informed and consented to participate in the survey for public purposes.

RESULTS
The distribution of the sociodemographic characteristics of the annual survey respondents is provided in Supplementary Material 3. The breast cancer screening rate initially increased from 38.4% in 2005 to 71.0% in 2012 (APC, 7.72%), followed by non-significant changes (Figure 1). In 2020, the breast cancer screening rate was 63.5% (Table 2), of which the proportions of the women who underwent organized and opportunistic screening were 58.3% and  . Contrastingly, the opportunistic screening rate showed a decreasing trend (APC, -6.0% from 2005 to 2020). In all subgroups, the organized screening rates increased significantly, except in the age group of 70-74 years and the rural population. Until 2012, the screening rate for breast cancer significantly increased in almost all subgroups, followed by a plateau. However, since 2012, the breast cancer screening rate has decreased in women aged 50-59 years, having 12-15 years of education, and living in rural areas. In women aged 50-59 years, breast cancer screening rates showed a statistically non-significant increase until 2012, and then significantly decreased (APC, -1.62%; 95% CI, -3.07 to -0.14) (Figure 2A). In the low-income group, breast cancer screening rates increased at an annual average of 11.34% between 2005 and 2010, but there was no significant change in the rates afterward (APC, -1.08%; 95% CI, -2.74 to 0.61) ( Figure 2B

DISCUSSION
It is necessary to screen repeatedly at an optimal interval to detect breast cancer in its early stages. Therefore, cancer screening should be performed according to the recommended protocol. According to the recommendations, this statistical study reports time trends in breast cancer screening, considering both organized and opportunistic screening over the past 16 years. Furthermore, we analyzed the breast cancer screening trend in socio-demographic subgroups to determine whether specific subgroups had different trends. Between 2005 and 2020, the national breast cancer screening rate has increased considerably. It has been consistently maintained above 60% since 2010, increasing to 63.5% in 2020.
In response to reduced requirements for the designation of screening facilities in the NCSP for breast cancer in 2008, the number of screening facilities increased [12]. This increase in screening facilities would have contributed to the marked increase in NCSP participation for breast cancer screening, which was approximately 62% in 2017 [13]. Although the screening protocols are diverse, in recent years, the rates of breast cancer screening in accordance with the recommendations have flattened in the United States and England [14,15]. Korea' s breast cancer screening rates were slightly lower than those of the United States (77.2% in 2000, 74.2% in 2015) and England (77.2% in 2011, 75.0% in 2019) [14,15], AAPC, average annual percent change; CI, confidence interval. 1 Respondents were restricted to women aged 40-74 years who had undergone mammography screening within a period of 2 years. 2 Low-income, middle-income, and high-income groups were classified according to each year's tertile of household income. 3 Some columns do not sum to 100% because of missing data. *p<0.05.  [16]. We selected the joinpoint regression model that best described each data set in the subgroup analysis according to socioeconomic characteristics. Each group showed different changes in the breast cancer screening rate. The recent significant declining trend in the estimated screening rate among women aged 50-59 years might have been due to a temporary drop in the screening rate due to coronavirus disease 2019 (COVID-19) pandemic in 2020 [17]. Further investigations are needed to minimize the impact of COVID-19 for a more accurate estimation of trends in the breast cancer screening rate.
Income and education levels are significantly associated with participation in breast cancer screening [18]. In our results, as in previous studies [8,19], the breast cancer screening rate was relatively high in high-income households, and the recent trend in the breast cancer screening rates showed non-significant changes or remained stable for all household income levels. The fact that the screening rate did not decrease in low-income households may be due to the positive effect of free national cancer screening. According to the KNCSS data, the rate of free-of-charge breast screening in accordance with the recommendations, also called "organized screening, " showed a significant increase between 2005 and 2020. In contrast, opportunistic screening showed a significant decrease (Supplementary Material 6). This difference between the organized and opportunistic screening rates may be due to the relatively reduced demand for opportunistic screening, especially in low-income households, as systematic screening is well established.
Our results show that breast cancer screening increased significantly in urban areas between 2005 and 2020, but declined significantly after 2011 in rural areas, while there was no significant change in metropolitan areas after a significant increase between 2005 and 2010. Regional differences in breast cancer screening rates may be based on access to screening institutions. In Korea, mammography equipment is concentrated in large cities [12,20]. Therefore, it is important to resolve this distributional inequity in mammography facilities. The gradient of breast cancer screening rates between regions results from both differences in socioeconomic status and accessibility, so neither factor is a simple proxy for the other [21]. Therefore, these results should be interpreted with caution. Nevertheless, it has been reported that there is a gap in the rate of breast cancer screening in countries with and without an organized national screening program like that of Korea [22,23].
The KNCSS has reported nationwide breast cancer screening rates for over 16 years. However, there were some limitations to our study. First, there may have been recall bias as survey data were used. However, despite this methodological limitation of the survey, the KNCSS is used as the source for calculating representative indicators related to cancer screening in Korea [24]. Second, there is a possibility of sampling error. We adopted stratified multistage, random sampling according to the residential area, gender, and age. In addition, this nationwide, annual, population-based survey has been conducted since 2004. Therefore, the sampling error would gradually be minimized as the sample size increased and the data sufficiently represented the population. Third, the breast cancer screening rates were presented considering both organized and opportunistic screening, but we did not analyze trends in these 2 screening rates separately according to socioeconomic status because we performed this analysis to determine how well the breast cancer screening recommendations, which recommends biennial mammography for women aged 40 and older, was implemented, regardless of the payment method.
This study provided more comprehensive results by including an individual's opportunistic screening experience besides their organized programmatic screening experiences. We report breast cancer screening rates between 2005 and 2020. Despite the continuous expansion of the breast cancer screening program, the mortality and medical cost burden of breast cancer is expected to increase in the future due to an increase in the absolute number of breast cancer patients (5,839 cases in 1999 vs. 23,723 cases in 2018) [1,25]. Additionally, decreasing trends in the breast cancer screening rate were observed for women aged 50-59 years, having 12-15 years of education, and living in rural areas. Further studies are needed to identify the groups more likely to be excluded from screening benefits and the causes of disparities in screening, including socio-demographic changes and gradients in breast cancer screening rate trends.