Analyzing Factors Enabling Colorectal Cancer Screening Adherence in Korean Americans Using the Andersen’s Behavioral Model of Health Services Utilization

Abstract

Purpose/Objectives:

Korean Americans (KAs) report suboptimal colorectal cancer (CRC) screening adherence. This study investigated factors that enable KAs to adhere to CRC screening guidelines using Andersen’s Behavioral Model of Health Services Utilization.

Design:

Cross-sectional survey using self-report measures of CRC screening behaviors.

Sample and Methods:

Purposive sampling was used to recruit 433 KAs aged 50–75 from the Atlanta metropolitan area who completed questionnaires measuring predisposing (i.e., gender, age, marital status, educational attainment), enabling (income, health insurance, regular annual check-ups, doctor recommendation, English proficiency, CRC knowledge, self-efficacy for CRC screening, decisional balance in CRC screening), and need (family cancer history, self-reported health status) factors associated with CRC screening.

Findings:

A multiple logistic regression model including all 14 predictor variables revealed that several enabling factors (i.e., income, regular annual health check-ups, doctor’s recommendation, self-efficacy, and decisional balance) independently predicted increased CRC screening adherence in KAs. No predisposing or need factors independently predicted CRC screening.

Conclusions and Implications for Psychosocial Providers or Policy:

To increase CRC screening adherence among KAs, psychosocial interventions should target on improving their self-efficacy and decisional balance regarding CRC screening, while policy interventions should focus on promoting health providers’ CRC screening recommendations during routine health check-ups.

Introduction

Routine colorectal cancer (CRC) screening can lower the risks of CRC occurrence and deaths from the disease through early detection and the subsequent removal of precancerous polyps.¹ A recent national survey reported CRC screening rates have almost doubled from 34% in 2000 to 63% in 2015, and incidence and mortality of CRC have gradually decreased among the general U.S. population over that same period.² While CRC screening guidelines vary by health organizations and evolve to reflect new scientific evidence, the commonly recommended guidelines include an annual stool-based blood test, sigmoidoscopy every five years, or colonoscopy every ten years for men and women aged 50 to 75 at average risk.³

Despite screening advantages and guidelines and overall decrease in CRC incidence and mortality for the general U.S. population, CRC incidence and mortality for Korean Americans (KAs) have increased and persisted, respectively, over the past two decades. CRC for KAs is the second most commonly diagnosed cancer and the leading cause of death.⁴ According to one systematic review,⁵ only 25% – 50% of KAs had ever had a stool-based blood test, sigmoidoscopy, or colonoscopy, and about 10% – 40% underwent sigmoidoscopy or colonoscopy within the last five or ten years, which falls far below the national target (70.5%) set by Healthy People 2020. Furthermore, CRC screening rates of KAs are lower compared to other ethnics groups, including 65% among non-Hispanic whites, 62% among Blacks, and 50% among Hispanics.⁶ According to U.S. Census data,⁷ over 1.7 million KAs live in the U.S., becoming one of the fastest growing minority groups, and about 71% were born in Korea, while about 25% arrived in 2000 or later. However, KAs have reported disparities in cancer screening, addressing their unmet needs for healthcare services.^8,9

Non-compliance with CRC screening guidelines poses heightened risks for late stage diagnosis of CRC and low five-year survival rates.¹⁰ Therefore, it is critical to examine factors that influence adherence to CRC screening recommendations among KAs. Moreover, there is a need for a theory-guided investigation of the factors that explain and enhance KAs’ compliance with the CRC screening guidelines. Thus, the present study employed the Andersen’s Behavioral Model of Health Services Utilization (hereafter, Andersen’s Model) as a theoretical framework to examine the factors contributing to CRC screening adherence among KAs.

Theoretical framework

The Andersen’s Model was conceptualized to explain how people access and use health care services. As shown in Figure 1, this model posits that people utilize health care under the condition that their predisposing characteristics, enabling resources, and need factors all function fully together.¹¹

Figure 1.

A Theoretical Framework of Andersen’s Behavioral Model of Health Services Utilization

This model has been widely used to investigate cancer screening use among KAs. For example, Lee and colleagues analyzed the merged data from the 2001, 2003, and 2005 California Health Interview Surveys and found KAs are less likely to utilize CRC screening compared to some other Asian American subgroups, Pacific Islanders, and non-Latino whites, after adjusting for predisposing, enabling, and need factors.¹² With guidance by the Andersen’s Model, Lee and colleagues also assessed factors related to breast cancer screening use among KAs in Midwest.¹³ This cross-sectional survey study found that age (predisposing factor), educational attainment, and knowledge of screening methods (enabling factor) were associated with breast cancer screening uptake among KAs. Furthermore, in a recent cross-sectional study using the Andersen’s Model, An and colleagues investigated breast cancer screening among KAs in the Atlanta metropolitan area and found that predisposing factors (i.e., age and marital status), enabling factors (i.e., income and annual checkup), and need factors (i.e., family cancer history) were associated with KAs’ breast cancer screening.¹⁴ These findings provided useful information in understanding how these three factors contribute to cancer screening utilization among KAs. However, further studies are needed to better understand the associations of CRC screening use with each of the factors of the Andersen’s Model among KAs, and how psychological and cultural variables (e.g., self-efficacy and attitudes and beliefs about CRC screening) are associated with CRC screening use among KAs.

Furthermore, literature on CRC screening across populations also sheds light on how variables in the predisposing, enabling, and need factors were connected to CRC screening.

Predisposing factors describe an individual’s demographic characteristics and social structure which are relevant to health care use.¹⁵ Prior research demonstrated that increased age was related to greater use of cancer screening,¹⁶ while gender has mixed results.^17,18 Also, married individuals were more likely than the unmarried to use health care,¹⁹ and those with higher educational attainment tended to utilize cancer screening more compared to those with lower educational attainment.²⁰ In this study, predisposing factors included gender, age, marital status, and educational attainment.

Enabling factors are defined as factors that allow people to lower financial and structural barriers to health care, thereby increasing the odds of accessing health care.²¹ Existing literature shows that increased income is associated with greater use of cancer screening.²² Moreover, having adequate coverage of health insurance positively affects access to cancer screening services,²³ and having regular medical check-ups or doctor’s recommendation for cancer screening is also linked to CRC screening.²⁴ Research has also indicated the key role of cultural and psychological factors in enabling CRC screening access and use. For example, language barriers limit individuals’ ability to navigate and access health care system.²⁵ Individuals with limited English proficiency are less likely to adhere to health care, such as getting prescription, medical treatment, and cancer screening, compared to those with adequate English proficiency.²⁶ Studies on cancer screening utilization among minority groups revealed that knowledge and self-efficacy regarding cancer screening are positively connected to the utilization of cancer screening.^27,28 Individuals’ attitudes and beliefs pertaining to cancer screening are also significantly associated with getting screened.²⁹ If individuals have negative attitudes or beliefs about cancer screening, these attitudes or beliefs hinder them from using the screening service.³⁰ In this study, the enabling factors included income, health insurance, regular annual health check-ups, doctor’s recommendation, English proficiency, CRC screening knowledge, self-efficacy for CRC screening, and attitudes and beliefs (i.e., decisional balance: pros vs. cons) about CRC screening.

Need factors refer to individuals’ perceived needs for health care and self-perceived health status.²⁵ Individuals who have had a family member diagnosed with CRC are more likely to get screened than those who have never had.³¹ Furthermore, individuals who are concerned about their own health tend to seek cancer screening.²⁷ The perceived needs for health care play a role in mobilizing individuals toward seeking CRC screening.³² The need factors under consideration in this study were family cancer history and self-reported health status.

This study addressed the following research questions:

Research Question 1. What are the rates of adherence to the CRC screening guidelines among KAs?

Research Question 2. What are the factors that predict adherence to the CRC screening guidelines for KAs?

Methods

Study design and data collection

A cross-sectional study was conducted in the Atlanta metropolitan area between May 2015 and February 2016. Self-identified KA aged 50 to 75 years and resident in the state of Georgia (GA) was eligible for inclusion in this study. Georgia is the third fastest growing KA population from 2000 to 2010, with an increase of 86.3%.³³ Furthermore, it is the 8^th state with the largest Korean population in the United States,³⁴ and the Atlanta metropolitan area ranks 6^th among all metropolitan areas in numbers of Korean population with a dramatic growth rate.⁷ The participants reporting cancer history were excluded from data analyses because their data might add significant bias to the results, and these participants had their own unique schedule of cancer screening as follow-up. Moreover, based on the screening guidelines from the American Cancer Society and the U.S. Preventive Services Task Force, participants who were below 50 or older than 75 were excluded from the study.

Purposive sampling was performed to recruit study participants from the KA community. Recruitment list of KA community organizations, including senior centers, churches/temples, and associations was developed. Each organization was contacted by the research team via phone and email. Organizations who agreed to participate selected a preferred time and place to determine when the self-report survey questionnaires would be administrated by the research team. During each session the nature and purpose of the study were explained, and those who were interested in the study completed the informed consent process prior to participating in the self-administered surveys. On average, the survey took approximately 45 minutes. The study was approved by the local university institutional review board.

A total of 433 eligible KAs completed the survey (measures described below). All measures in English were translated into Korean using back-translation to assure comparability and equivalence in the meaning of measures. Two bilingual faculty at an English literacy department and a bilingual community health professional reviewed the translation. A pilot test of the Korean questionnaire was implemented with six Korean American men and women in order to finalize by integrating their feedbacks. While participants were allowed to choose either an English or Korean questionnaire, all participants self-administered the Korean questionnaire.

Measures

The outcome variable of CRC screening was assigned as a dummy variable with a value of 1 if the participant had up-to-date CRC screening or a value of 0 if otherwise. CRC screening was considered as up-to-date if the individual reported having: annual stool-based blood testing (e.g., Fecal Occult Blood Test [FOBT]), sigmoidoscopy in the past five years, or colonoscopy in the past ten years. The description of the screening tests was provided to participants before they were asked whether they had completed the test.

The predisposing factors in the study include dichotomized sociodemographics of gender, (female vs. male), age (50–64 vs. 65–75), marital status (never married/other vs. married/partnered), and educational status (below Bachelor’s degree vs. Bachelor’s degree or above).

The enabling variables consisted of categorical and continuous variables. The categorical enabling variables include income ( < $20,000, $20,000 – $39,999, $40,000 – $59,999, $60,000 – $79,999, $80,000 – $99,999, ≥ $100,000), health insurance (‘yes’ or ‘no’), regular annual health check-ups (‘yes’ or ‘no’), doctor recommendation for CRC screening (‘yes’ or ‘no’), and English proficiency (‘very bad,’ ‘bad,’ ‘moderate,’ ‘good,’ ‘very good’); the continuous enabling variables include CRC screening knowledge, self-efficacy, and decisional balance.

CRC screening knowledge

CRC screening knowledge was measured by the composite mean score of both the seven items of CRC Knowledge Assessment Survey developed by Sanchez et al.³⁵ and the eleven items of CRC Familiarity Scale developed by Han et al.³⁶ The CRC Knowledge Assessment Survey measures general CRC-related knowledge using four items about CRC screening knowledge and three items about CRC risk factor knowledge with values of 1 for ‘true’ or 0 for ‘false.’ Example items include: “Both men and women at age of 50 should begin colorectal cancer screening” and “Colorectal cancer has been linked to the heavy use of alcohol.” The CRC Familiarity Scale assesses participants’ familiarity with the terms related to CRC screening (i.e., adenomas, benign, malignant, and sigmoid); they are terms often used by health care providers over the course of CRC screening and essential for patients to understand for completing the screening procedure. The five-point Likert scale ranges from 0 for ‘do not know at all’ to 4 for ‘know very well.’ The CRC Familiarity Scale items were recoded to 0 for ‘do not know at all’ or 1 for ‘know a little,’ ‘know somewhat,’ ‘know well,’ and ‘know very well.’ In addition to the CRC Knowledge Assessment Survey, as Han and colleagues suggested, the CRC Familiarity Scale was employed in this study to assess more comprehensive CRC knowledge in the CRC screening-specific context at health care settings.³⁶ The total CRC knowledge composite score was calculated as the mean value of both scales, with the higher score indicating better CRC knowledge. The Cronbach’s alpha for this CRC screening knowledge scale was 0.8650.

CRC screening self-efficacy

CRC screening self-efficacy was measured by the mean score of 12 items of perceived self-efficacy for CRC screening adopted from Luszczynska and Schwarzer.³⁷ This included three subscales for a stool-based blood test, sigmoidoscopy, and colonoscopy with four items each. A seven-point Likert scale ranging from 1 for ‘strongly disagree’ to 7 for ‘strongly agree’ was used to measure abilities of the participants to overcome potential obstacles anticipated in the screening test. A sample item asks “I am able to perform a stool-based blood test (or sigmoidoscopy or colonoscopy) regularly even if I will have to make a detailed plan describing how to remember about the test.” The Cronbach’s alpha for the 12 items was 0.8737.

CRC screening decisional balance

A total of 54-item decisional balance scale for a stool-based blood test, sigmoidoscopy, and colonoscopy, developed by Costanza et al.,³⁸ was adopted to measure attitudes and beliefs (pros vs. cons) about CRC screening. An example item asks “Having regular stool-based blood test (or sigmoidoscopy or colonoscopy) gives me peace of mind about cancer,” and each item has a binary value of 0 for ‘no’ or 1 for ‘yes.’ The Cronbach’s alpha for the 54 items of CRC decisional balance was 0.9150.

Particularly, the scores of the three enabling variables (i.e., knowledge, self-efficacy, and decisional balance) were log transformed to observe the effect of the log transformed variable’s percentage change.

Lastly, the need variables in this study included family cancer history and self-reported health status. The family cancer history was measured with the question of “Has any of your family (i.e., parents, grandparents, siblings, or close relatives) ever had cancer of any kind?” and options of ‘yes’ and ‘no.’ The health status was assessed with the question of “How do you rate your current health?” and options of ‘very bad,’ ‘bad,’ ‘moderate,’ ‘good,’ ‘very good.’

Statistical analysis

Descriptive statistics summarized the sample’s key characteristics, and bivariate analyses were conducted to examine the associations of CRC screening adherence with each predisposing, enabling, and need factor. Two-sided Pearson Chi-squared test for categorical variables and two-sided t-test for continuous variables were used to evaluate the associations. A multiple logistic regression model with heterogeneity robust standard errors was applied to investigate the factors that were significantly associated with the dichotomous outcome of CRC screening adherence. A 5% significance level was used as a criterion for all statistical tests in the study. Data were analyzed with Stata/SE 14.2.

Results

Sociodemographic characteristics

Table 1 shows sociodemographic characteristics of the participants. Approximately 62% of the participants were female, and about 75% were aged 50 to 64. The majority (84%) of the participants were married or partnered, and 59% had bachelor’s degree. More than a third (34.8%) earned annual household income of $60,000 or above. Approximately 43% of the participants reported having regular annual health check-ups, 74% had health insurance, and 49% received doctor’s recommendation to be screened. Half of the participants reported having a family cancer history. More than a third (33.6%) reported their health status as ‘very good’ or ‘good,’ while 58% indicated they had moderate health status. Only 8% reported having ‘bad’ or ‘very bad’ health status. There were 12% of the participants who reported their English proficiency level as ‘good’ or ‘very good,’ while 40% had moderate English proficiency and less than a half (47.8%) reported having ‘bad’ or ‘very bad’ English proficiency.

Table 1.

Sociodemographic characteristics and their differences by CRC screening adherence (N = 433)

Variable	na (%) or (Mean, SD)	CRC screening adherenceb
		n (%) or Mean (SD)	p-valuec
Predisposing Factor
Gender
Female	267 (61.7%)	117 (43.8%)	0.3123
Male	166 (38.3%)	81 (48.8%)
Age (Years) (58.8, 7.22)
50 − 64	325 (75.1%)	132 (40.6%)	0.0002
65 − 75	108 (24.9%)	66 (61.1%)
Marital Status
Never married or other	69 (16.0%)	29 (42.0%)	0.4910
Married or partnered	361 (84.0%)	168 (46.5%)
Education
< Bachelor’s degree	175 (41.4%)	70 (40.0%)	0.0510
≥ Bachelor’s degree	248 (58.6%)	123 (49.6%)
Enabling Factor
Income
< $20,000	63 (15.9%)	38 (60.3%)	0.0254
$20,000 – $39,999	99 (24.9%)	38 (38.4%)
$40,000 – $59,999	97 (24.4%)	36 (37.1%)
$60,000 – $79,999	70 (17.6%)	37 (52.9%)
$80,000 – $99,999	40 (10.1%)	20 (50.0%)
≥ $100,000	28 ( 7.1%)	15 (53.6%)
Health Insurance
No	109 (25.6%)	27 (24.8%)	<0.0001
Yes	317 (74.4%)	169 (53.3%)
Regular Annual Health Check-Ups
No	244 (57.0%)	69 (28.3%)	<0.0001
Yes	184 (43.0%)	128 (69.6%)
Doctor Recommendation
No	205 (51.4%)	40 (19.5%)	<0.0001
Yes	194 (48.6%)	137 (70.6%)
English Proficiency
Very Bad / Bad	207 (47.8%)	82 (39.6%)	0.0089
Moderate	175 (40.4%)	84 (48.0%)
Very Good / Good	51 (11.8%)	32 (62.8%)
Psychosocial Factor
CRC Screening Knowledge	433 (0.52, 0.26)	0.58 (0.26)	<0.0001
CRC Screening Self-Efficacy	420 (0.49, 0.18)	0.56 (0.17)	<0.0001
CRC Screening Decisional Balance	422 (0.70, 0.19)	0.76 (0.17)	<0.0001
Need Factor
Family Cancer History
No	216 (50.0%)	90 (41.7%)	0.0822
Yes	216 (50.0%)	108 (50.0%)
Self-Report Health Status
Very Bad / Bad	36 ( 8.4%)	21 (58.3%)	0.2888
Moderate	248 (57.9%)	114 (46.0%)
Very Good / Good	144 (33.6%)	63 (43.8%)

^aThe total sample size of each variable may not be the same as the total sample size of the study due to missing values.

^bCRC screening adherence based on the CRC screening guidelines (any of annual stool-based blood testing, sigmoidoscopy every five years, or colonoscopy every ten years).

^cp-values for Pearson Chi-squared test for categorical variables and t-test with unequal variances for continuous variables.

Table 1 also shows that participants’ adherence to the CRC screening guidelines differed depending on the predisposing factor of age (p = 0.0002) and significant enabling factors of household income levels (p = 0.0254), health insurance (p < .001), regular annual health check-ups (p < .001), doctor’s recommendation (p < .001), and English proficiency levels (p < .01). Participants’ CRC screening adherence also had significant differences in scores of knowledge (p < .001), self-efficacy (p < .001), and decisional balance (p < .001). However, none of the need factors were significantly associated with CRC screening adherence.

CRC screening adherence

As shown in Table 2, less than half (45.7%, n = 198) of the participants reported adhering to the screening guidelines. Among the 198 participants who were up-to-date with their screening, most participants had been screening using colonoscopy (43%), 18% had been screened using sigmoidoscopy, and 8% had been screening using a recommended stool-based blood test.

Table 2.

CRC screening adherence (N = 433)

CRC screeninga	n (%)
Meets any of the CRC screening test guidelines	198 (45.73%)
- FOBT (in last year)	36 (8.31%)
- Sigmoidoscopy (in last 5 years)	80 (18.48%)
- Colonoscopy (in last 10 years)	188 (43.42%)

^aCategories are not mutually exclusive.

Factors predicting CRC screening adherence

Table 3 shows the results from the multiple logistic regression analyses conducted with predisposing, enabling, and need factors for CRC screening adherence. The model fits the data well (${\chi }_{\left(8\right)}^2=8.95$, p = .3464). None of the predisposing factors were significantly and independently associated with CRC screening adherence. However, for enabling factors, having annual household income of $80,000 to $99,999 had significant association with having less CRC screening adherence (OR = 0.29, p < .05, 95% CI [0.09, 0.99]). Having regular annual health check-ups (OR = 4.78, p < .001, 95% CI [2.40, 9.54]), having doctor’s recommendation (OR = 8.26, p < .001, 95% CI [4.41, 15.45]), greater self-efficacy score (OR = 2.82, p < .05, 95% CI [1.11, 7.20]), and greater decisional balance score (OR = 4.61, p < .01, 95% CI [1.49, 14.30]), respectively, were also significantly associated with more adherence to CRC screening, while controlling for other variables. Particularly, the odds ratio was 2.82 for 1% change in the self-efficacy score, and 4.61 for 1% change in the decisional balance score. None of the need factors were significantly related to CRC screening adherence.

Table 3.

Logistic regression analysis^a: CRC screening adherence^b

	OR	95% CI
Predisposing Factor
Male (Ref: Female)	0.87	(0.45 − 1.69)
Age (Ref: 50−64 years)	1.04	(0.99 − 1.10)
Marital Status (Married or partnered) (Ref: Never married or other)	1.28	(0.54 − 3.02)
Education (Bachelor’s degree) (Ref: < Bachelor’s degree)	1.05	(0.53 − 2.07)
Enabling Factor
Income (Ref: < $20,000)
$20,000 – $39,999	0.45	(0.16 − 1.29)
$40,000 – $59,999	0.35	(0.12 − 1.05)
$60,000 – $79,999	0.81	(0.25 − 2.61)
$80,000 – $99,999	0.29*	(0.09 − 0.99)
≥ $100,000	0.35	(0.09 − 1.04)
Health Insurance	1.34	(0.65 − 2.76)
Regular Annual Health Check-Ups	4.78***	(2.40 − 9.54)
Doctor Recommendation	8.26***	(4.41 − 15.45)
English Level (Ref: Very Bad / Bad)
Moderate	0.87	(0.41 − 1.85)
Very Good / Good	0.82	(0.29 − 2.31)
CRC Knowledge Scorec	1.73	(0.83 − 3.61)
CRC Self-efficacy Scorec	2.82*	(1.11 − 7.20)
CRC Decisional Balance Scorec	4.61**	(1.49 − 14.30)
Need Factor
Family Cancer History	1.27	(0.69 − 2.35)
Self-reported Health Status (Ref: Very Bad / Bad)
Moderate	0.71	(0.23 − 2.15)
Very Good / Good	0.49	(0.15 − 1.58)
Number of Observations	343
Hosmer-Lemeshow goodness-of-fit test	8.95
Wald χ2 test	117.23***
Pseudo R2	0.4114

^aHeterogeneity robust standard errors are used.

^bCRC screening adherence based on the CRC screening guidelines (any of annual stool-based blood testing, sigmoidoscopy every five years, or colonoscopy every ten years).

^cNatural log of each score

^*p < 0.05

^**p < 0.01

^***p < 0.001

Discussion

This study found a low rate (45.7%) of CRC screening adherence among KAs residing in the Atlanta metropolitan area. The adherence rate in this study is similar to that of prior research on KAs, yet still lower than that of aggregated Asian Americans (47% – 58%), non-Hispanic white Americans (66%), and the general population of U.S. (59%).^5,6,22 This suboptimal CRC screening outcome can be understood by comparing the characteristics of KA participants who adhere to the screening guidelines with those who do not adhere.

This study found several enabling factors predicting participants’ adherence to the screening guidelines. In the analytic model of this study, doctor’s recommendation for CRC screening was the strongest predictor for CRC screening adherence. Research shows that health providers’ screening recommendation positively influences cancer screening use among Asian Americans, specifically Filipinos and Koreans in the U.S.³⁹ To enable Korean Americans to maintain routine CRC screening, however, it is necessary for the doctor’s recommendation to occur repetitively because CRC screening is not an one-time but rather a recurrent health-related behavior based on the established timeline.⁴⁰ Therefore, the finding also highlights the significant role of regular annual health check-ups in facilitating CRC screening adherence. Annual health check-ups can provide opportunities for patients to discuss screening guidelines. Health providers can leverage the visits of patients to clinics for annual health check-ups by providing flyers or brochures with a summary of CRC-related information (risk and preventive factors of CRC) and guidelines of CRC screening or by showing a short video with information on CRC screening test options, costs, and preparation and screening procedure types (e.g., biopsy, polyp removal) for patients are at a waiting room. Because the costs of CRC screening tests vary based on type and plan of health insurance and types of procedure, it is also critical for health providers to recommend patients to check with their health insurers.⁴¹

Moreover, self-efficacy for CRC screening and decisional balance toward screening were found to be the enabling factors predicting CRC screening adherence in the participants. Consistent with prior research,⁴² self-efficacy was significantly associated with CRC screening adherence in KAs. Self-efficacy refers to individuals’ perceived beliefs regarding conducting desirable health-related tasks they seek so self-efficacy for cancer screening is influenced by their previous screening experiences.⁴³ Therefore, the finding suggests the need of innovative strategies to engage KAs in the use of an easy-to-perform CRC screening test, such as FOBT, and to have positive experience with the test to promote continuous use. The initial engagement with and satisfactory completion of the screening test would enhance their self-efficacy for CRC screening and eventually result in adhering to the screening guidelines. On another hand, research shows that individuals’ initial engagement with health behavior is closely linked to their attitudes and beliefs (i.e., decisional balance) about the behavior.^30,44 Individuals with negative attitudes and/or beliefs pertaining to CRC screening are less likely than those with positive ones to engage with screening and even continue screening according to its guidelines.⁴⁵ As a result, educational interventions for CRC screening adherence among KAs should focus on promoting positive attitudes and beliefs regarding the screening. Because sigmoidoscopy and colonoscopy pose a relatively high burden in bowl preparation and sedation and more complex guidelines than other screening tests (e.g., mammography), it is useful for educational materials to include visual information of the whole screening procedure along with relevant screening guidelines through brochures-, DVDs-, or Internet-based simulations; also, the education should focus on enhancing KAs’ perceived self-efficacy of screening.

Finally, this study found that participants with annual household income of $80,000 to $99,999 were less likely to adhere to the CRC screening guidelines than those with other annual household incomes. While this finding needs further analyses on the relationship with collapsed income levels and income sources, an explanation for this can be partly because this income group might be working couples or small business owners without an appropriate health insurance and/or with busy working schedules for a living.^46,47 Thus, it would be difficult for them to secure sufficient time for screening uptake, thereby delaying screening uptake. This suggests that interventions for CRC screening adherence for this group need to be adjusted based on time efficiency for screening not cost effectiveness. Health providers can recommend them for a time-efficient CRC screening test like FOBT.

Limitations and future research directions

Although this study obtained a good sample size that reflected characteristics of KAs in GA, the significant findings cannot be generalized to all KAs nor can they be used to assert causal relationships between the variables of interest in this study. Future studies based on probability sampling would offer a clearer understanding of the characteristics that predict CRC screening adherence among KAs. Furthermore, while the self-report survey is the most widely used method for measuring cancer screening adherence in terms of its time efficiency, cost effectiveness, and actionable information, the information related to screening history might be influenced by participants’ recall and social desirability bias. Reducing social desirability bias may enhance validity and precision of the self-report cancer screening adherence. It would also be helpful to delineate each screening test and its reference periods in detail on the questionnaire. Lastly, while this study employed the Andersen’s Behavioral Model as a conceptual framework, different theories, such as the Transtheoretical Model or the Health Belief Model, could also explain the associations of the diverse factors that influence CRC screening adherence.

Conclusions

Using the Andersen’s Behavioral Model, this study evaluated the predisposing, enabling, and need factors associated with CRC screening guidelines adherence among KAs. Specifically, this study focused on assessing the participants’ CRC screening adherence rate, differences in characteristics of the participants by variables, and predictors of the screening adherence with an emphasis on the factors that can be modified to facilitate screening adherence in KAs who do not adhere to the guidelines.

The study found that KAs had poor CRC screening adherence, which was characterized by the predisposing and enabling factors, including age, household income levels, health insurance, regular annual health check-ups, doctor’s recommendation, English proficiency levels, CRC screening knowledge, self-efficacy for CRC screening, and decisional balance regarding CRC screening. The study also found several enabling factors (i.e., household income, regular annual health check-ups, doctor’s recommendation, self-efficacy, and decisional balance pertaining to CRC screening) to predict adherence to the screening guidelines among KAs.

The study confirmed prior research and added to the body of relevant literature by highlighting the psychosocial and cultural factors (i.e., self-efficacy and decisional balance regarding CRC screening) that enable KAs to adhere to CRC screening guidelines. The findings also highlight the significant role of health providers in improving CRC screening adherence among KAs by enhancing patient-doctor communications for CRC screening uptake and adherence. The results also shed light on interventions and practices aimed at promoting CRC screening adherence. Future research should focus on improving the self-efficacy for CRC screening and attitudes and beliefs regarding screening through leveraging these enabling factors among medically underserved KAs and focus on increasing their CRC screening adherence and reducing the disparities in their CRC incidence and mortality.