Colorectal carcinoma screening among Hispanics and non-Hispanic whites in a rural setting
Colorectal carcinoma ranks as the second most common cancer and the second leading cause of cancer death in the United States. Hispanics are less likely than their non-Hispanic white counterparts to have ever received a fecal occult blood test (FOBT) or sigmoidoscopy/colonoscopy. Little is known about the barriers to screening in the Hispanic population.
The authors used baseline data from a community randomized trial of cancer prevention to compare screening prevalence and the associations between reported barriers and screening participation between Hispanics (n = 137) and non-Hispanic whites (n = 491) age ≥ 50 years.
Hispanics were less likely than non-Hispanic whites to have ever received an FOBT (P = 0.003) or sigmoidoscopy/colonoscopy (P = 0.001). No significant difference across ethnic groups was observed in the prevalence of recent screening using FOBT (29.8% for Hispanics vs. 34.5% for non-Hispanic whites; P = 0.41), but recent use of sigmoidoscopy/colonoscopy was lower for Hispanics (24.1% for Hispanics vs. 33.7% for non-Hispanic whites; P 0.06). Lacking health care coverage or having few years of education were directly associated with failure to ever receive an FOBT or sigmoidoscopy/colonoscopy.
Interventions to improve adherence to colorectal carcinoma screening recommendations among Hispanics should target initial screening examinations, particularly among those lacking health care coverage or having low levels of education. Cancer 2005. © 2005 American Cancer Society.
Colorectal carcinoma ranks as the second most common cancer and the second leading cause of cancer death in the United States.1–3 Although Hispanics in the United States experience a lower incidence of colorectal carcinoma than their non-Hispanic white counterparts (43.8 per 100,000 for Hispanic vs. 64.1 for non-Hispanic white males; and 29.4 for Hispanic vs. 47.2 for non-Hispanic females), incidence increases with time in the United States. The greatest increase in colorectal carcinoma incidence among Hispanics occurs between the first and second generation.4 The United States has and will continue to experience dramatic increases in Hispanic ethnicity,5 particularly related to immigration from Latin America. Therefore, developing effective cancer prevention interventions in this group will become an increasing priority.
Hispanics are more likely to have larger colorectal tumors or more advanced-stage disease at the time of diagnosis, compared with non-Hispanic whites.6, 7 For these and other reasons, Hispanics experience higher mortality rates from colorectal carcinoma than non-Hispanic whites6 (Unpublished data). A key factor believed to explain the later stage at diagnosis is the underutilization of screening services for colorectal carcinoma.8
Routine screening reduces the incidence of advanced-stage colorectal carcinoma. Although there is some variation in recommendations, all organizations concerned with colorectal carcinoma emphasize the importance of screening beginning at age 50. The U.S. Preventive Services Task Force, for example, recommends annual fecal occult blood testing (FOBT) and/or sigmoidoscopy at unspecified intervals9; the American Cancer Society10 recommends an annual FOBT, sigmoidoscopy every 5 years, and colonoscopy every 10 years; and the American College of Gastroenterology recommends colonoscopy every 10 years or flexible sigmoidoscopy every 5 years plus an annual FOBT.11
Screening for colorectal carcinoma is performed with increasing frequency among the general population.12 However, data from several studies suggest that screening among Hispanics lags behind that of non-Hispanic whites.12 Data from the 2002 Behavioral Risk Factor Surveillance System of the Centers for Disease Control show that 34.5% of age-eligible Hispanics had ever received an FOBT compared with 46.7% of non-Hispanic whites.12 A similar pattern is observed when comparing proportions of lifetime participation in colonoscopy or sigmoidoscopy screening among Hispanics and non-Hispanic whites (38.9% vs. 49.9%, respectively).12
Although much is known about Hispanic barriers associated with screening participation for other types of cancer, barriers related to use of screening services for colorectal carcinoma are poorly understood. A limited number of previous investigations have shown that across racial and ethnic groups, those with less education or income are less likely to be screened.13 Focus group data from a sample of first-degree relatives of Hispanic and non-Hispanic white patients with colorectal carcinoma identified several barriers to screening, such as fear of finding cancer, pain associated with sigmoidoscopy, feeling violated, not knowing where to obtain screening, and difficulties in obtaining an appointment. The prevalence of sociodemographic characteristics and screening barriers are reported and the associations of sociodemographic characteristics and screening barriers to screening participation are described for Hispanics and non-Hispanic whites. Few previous studies on this topic have been conducted among Hispanics, and the current study, to our knowledge, is the first investigation among Hispanics in a rural setting to assess the influence of barriers on colorectal carcinoma screening participation. This is especially important as residents of rural areas might have more limited access to health care or experience a unique set of structural barriers to obtain needed services.
MATERIALS AND METHODS
Study Design and Sample
Data are from a community randomized trial, designed to assess the effectiveness of a comprehensive cancer prevention intervention in 20 communities in the Lower Yakima Valley, Washington State.14 Briefly, the Hispanic population is concentrated in Yakima County, where it constitutes 24% of the county population and nearly one-half of the state's total Hispanic population. The lower part of the county, as well as part of its neighboring county (Benton), is called the Lower Yakima Valley. An estimated 65% of Hispanics in the area are employed in the agriculture industry as field workers, applicators of farm chemicals, and other farm-related tasks.15
Adult residents from all 20 communities in the Yakima Valley were eligible to participate in a baseline assessment, and the following procedures were used to recruit respondents. For each community, we used census blocks to define geographic boundaries for sample selection. Address lists, purported to be household addresses, were purchased from bulk mailing companies and overlaid on the census block maps to identify households from which a random sample could be drawn. Eligibility criteria required that the respondent be ≥ 18 years, have lived in the household for at least the past week, and be willing and able to respond to questions. A sample of 160 households was drawn from each community, with the expectation of having approximately 100 usable households per community after businesses, empty dwellings, vacant lots, and other ineligible addresses were excluded. Within each randomly selected household, one adult was interviewed. If ≥ 2 eligible adults lived in the household, the first adult to have a birthday after December 31st was selected for the interview. From a sample of 2862 addresses, 2345 households were known to be eligible to participate in the study. The remaining addresses were vacant buildings (n = 190), organizations (e.g., schools, churches, fire stations) (n = 162), nonexistent dwellings (n = 109), and businesses (n = 56). A total of 1795 individuals completed the interviews. There was no answer after ≥ 5 visits in 141 households, and no adult home after ≥ 5 visits in 45 households, for a conservative response rate of 76.5%. The response rate of the known eligible households (n = 2159) was 83.1%.
All interviews were conducted by 22 locally hired and trained bilingual, biliterate project staff between October 1, 1998 and January 31, 1999. Procedures for the current study were approved by the institutional review board at the Fred Hutchinson Cancer Research Center. All respondents gave verbal consent to participate and were given a small incentive (i.e., a fanny pack) upon completion of the interview.
The survey instrument was a 100-item questionnaire, which included 7 sections: acculturation, health care access, smoking behavior, eating patterns, pesticide exposure, demographics, and beliefs about cancer and use of cancer screening tests.14 However, because the current report focuses on colorectal carcinoma screening compliance, we limit our description of items to those measures. Interviewers asked all male and female residents ≥ 50 years about 2 types of colorectal carcinoma screening. Respondents who reported having had an FOBT within the past 2 years were classified as being “in compliance.” Similarly, if residents ≥ 50 years reported having received a sigmoidoscopy within the previous 5 years, they were compliant, but they were not compliant if the sigmoidoscopy was received > 5 years, or if they had never received a sigmoidoscopy. In piloting our instrument, we found that our respondents could not differentiate a sigmoidoscopy from a colonoscopy even when we used a visual of the procedure and described the differences. For that reason, we report sigmoidoscopy/colonoscopy together.
Barriers to Cancer Screening
We used the categories defined by Aday et al.16, 17 to construct 10 items with reasons for not obtaining cancer screening tests, such as, “I don't get cancer screening tests because they cost too much.” Response to each item was scored on a 4-point Likert scale (1 = strongly agree to 4 = strongly disagree). However, because cell sizes were very small for strongly agree and strongly disagree, we collapsed strongly agree with agree and strongly disagree with disagree for analysis. These decisions did not change the outcome of the analysis or our interpretation of the results. We divided these screening attitudes/barriers into two subscales, which we called structural barriers and personal barriers. The scale and the subscales had high internal consistency (Cronbach α = 0.90).
Level of acculturation was assessed using a four-item scale developed by Coronado et al.18 and validated using the current study sample. The items are language spoken, language used for thought, birthplace, and ethnic self-identification.
We obtained data on age, education, race/ethnicity, and access to health insurance. For these analyses, we included 137 Hispanic and 491 non-Hispanic participants who were age eligible to answer the questions about colon carcinoma screening. We used descriptive statistics to characterize the study population and beliefs and attitudes about cancer and cancer screening. We calculated the percentage of subjects who had ever received, and who had recently received FOBT or sigmoidoscopy/colonoscopy among the total sample, Hispanics, and non-Hispanic whites. Chi-square tests were used to assess the differences in prevalence between Hispanics and non-Hispanic whites. P < 0.05 is statistically significant (based on a 2-sided test). We present the proportion of participants who reported barriers to screening. To determine associations of sociodemographic characteristics and perceived barriers to cancer screening tests with lifetime colorectal carcinoma screening participation, we calculated odds ratios (OR) using logistic regression. We present raw results, as well as results adjusted for factors known to influence cancer screening behavior such as age, gender, income, access to health insurance, smoking, and residential community. All analyses were conducted with SAS software (version 6.12, Cary, NC).
Of the 1795 individuals who were interviewed, 628 were ≥ 50 years. Table 1 gives the demographic and lifestyle characteristics of this subset. The mean age of study participants was 66 years and 59% were female. Approximately 22% of respondents to this part of the study were Hispanic. Among Hispanics, 41.6% were highly acculturated. Approximately one-fourth of all study respondents had received only an eighth-grade education or less. However, among Hispanics, more than three-fourths had received an eighth-grade education or less. Many participants were living at poverty level; 30% of all participants ≥ 50 years had a household income < $15,000 per year. However, 66% of Hispanics had a household income < $15,000 per year. Approximately 60% were married or living as married. Among respondents, 94% of non-Hispanics had health insurance compared with 73% for Hispanics. Medical insurance includes Medicaid and the Washington State Basic Health Plan, an insurance plan designed specifically for low-income families.
Table 1. Demographic Characteristics of the Study Population
|Age (mean ± SD)||66.1 (11.3)||66.4 (11.3)||64.8 (11.0)|
| 50–59||219||168 (34.2)||51 (37.2)|
| 60–69||166||123 (25.1)||43 (31.4)|
| ≥70||243||200 (40.7)||43 (31.4)|
|Gender|| || || |
| Male||258||206 (42.0)||52 (38.0)|
| Female||370||285 (58.0)||85 (62.0)|
|Acculturation|| || || |
| High|| || ||57 (41.6)|
| Low|| ||N/A||80 (58.4)|
|Educationa|| || || |
| <8th grade||167||62 (12.7)||105 (76.6)|
| High school diploma||280||257 (52.6)||23 (16.8)|
| Some college||103||95 (19.4)||8 (5.8)|
| College degree or higher||76||75 (15.3)||1 (<1.0)|
|Income (dollars/year)1|| || || |
| <15,000||189||122 (33.7)||67 (65.7)|
| 15,000–25,000||102||79 (21.8)||23 (22.5)|
| 26,000–50,000||109||97 (26.8)||12 (11.8)|
| >50,000||64||64 (17.7)||0|
|Marital status|| || || |
| Married||374||293 (60.3)||81 (59.6)|
|Smoking status|| || || |
| Current smoker||90||78 (15.9)||12 (8.9)|
|Have health insurance||560||460 (93.9)||100 (73.0)|
As shown in Table 2, 41% of Hispanics had ever had an FOBT compared with 56% for non-Hispanic whites (P = 0.003) and the difference remained significant after adjustment for age group (P = 0.006). Hispanics were less likely than non-Hispanic whites to have ever received a sigmoidoscopy/colonoscopy (age-adjusted 26.4% vs. 43.8%; P value < 0.001). Recent use of FOBT also varied by ethnicity, with a slightly lower proportion of Hispanics than non-Hispanic whites having received an FOBT in the past 2 years, although the differences were nonsignificant. Nearly one-fourth of Hispanics and one-third of non-Hispanic whites had received a sigmoidoscopy/colonoscopy in the past 5 years. The difference was statistically significant (P < 0.05) and attenuated only slightly after adjustment for age (P = 0.06).
Table 2. Participation in FOBT and Sigmoidoscopy/Colonoscopy Screening, by Ethnicity
|FOBT screening|| || || || || || |
| Ever had||55.7||40.6||0.003||55.6||41.3||0.006|
| Had within past 2 yrs||34.5||29.8||0.32||34.1||30.2||0.41|
| Had > 2 yrs ago||20.9||9.9||0.005||20.7||9.8||0.006|
|Sigmoidoscopy/colonoscopy|| || || || || || |
| Ever had||44.4||26.9||<0.001||43.8||26.4||<0.001|
| Had within past 5 yrs||33.7||24.1||<0.05||32.8||23.8||0.06|
| Had > 2 yrs ago||10.0||3.0||0.013||9.9||3.0||0.015|
When we evaluated the association between FOBT or sigmoidoscopy/colonoscopy screening and age, education, and availability of health insurance, the models were consistently and positively associated with screening compliance (Table 3). Hispanics were less likely to have received FOBT when the raw OR was evaluated, and less likely to have received sigmoidoscopy/colonoscopy when considering all variables. There was a strong, statistically significant inverse association between being a current smoker and having ever received a sigmoidoscopy/colonoscopy, although the association was no longer significant once other factors were included in the model.
Table 3. Associations of Demographic Characteristics with Compliance with FOBT and Sigmoidoscopy/Colonoscopya
|Age (yrs)|| || || || |
| 60–69||1.52 (0.93–2.47)||1.44 (0.86–2.41)||2.48 (1.51–4.05)||2.65 (1.58–4.45)|
| >70||1.77 (1.14–2.76)||1.63 (1.01–2.64)||2.58 (1.64–4.07)||2.72 (1.65–4.47)|
|P value for trend||0.02||0.06||0.05||<0.001|
|Gender|| || || || |
| Male||1.24 (0.85–1.79)||1.33 (0.91–1.95)||0.82 (0.57–1.19)||0.84 (0.58–1.24)|
|Race/ethnicity|| || || || |
| Non-Hispanic white||1.0||1.0||1.0||1.0|
| Hispanic||0.44 (0.27–0.74)||0.63 (0.33–1.24)||0.63 (0.39–1.01)||0.52 (0.28–0.98)|
|Education|| || || || |
| <8th grade||1.0||1.0||1.0||1.0|
| High school graduate||1.28 (0.81–2.03)||1.36 (0.77–2.41)||1.19 (0.74–1.90)||1.08 (0.61–1.91)|
| Some college||1.05 (0.58–1.90)||1.07 (0.54–2.13)||1.83 (1.04–3.21)||1.79 (0.93–3.46)|
| College degree||1.21 (0.64–2.30)||1.24 (0.58–2.64)||2.50 (1.37–4.55)||2.59 (1.25–5.34)|
|P value for trend||>0.05||>0.05||0.01||0.06|
|Current smoker|| || || || |
| Yes||0.65 (0.36–1.15)||0.76 (0.41–1.40)||0.43 (0.23–0.80)||0.58 (0.30–1.13)|
|Health insurance|| || || || |
| Yes||2.97 (1.32–6.66)||2.93 (1.19–7.19)||3.96 (1.68–9.35)||2.78 (1.06–7.27)|
Beliefs about cancer and cancer screening tests varied by ethnicity (Table 4). In general, non-Hispanic whites experienced fewer structural barriers that prevented them from obtaining cancer screening tests, compared with Hispanics. Approximately one-fourth of Hispanics reported that they could not take time off work to obtain cancer screening tests, compared with 14% of non-Hispanic whites. Similarly, 25.8% of Hispanics reported that lack of transportation to the clinic or physician prevented compliance with cancer screening tests, but only 10% of non-Hispanic whites reported transportation problems. Approximately 16% of Hispanics reported that clinic staff who do not speak Spanish is a barrier to obtaining a cancer screening test. There were few differences in personal barriers to obtaining cancer screening tests across ethnic groups. A slightly greater proportion of Hispanics than non-Hispanic whites reported that the high cost of screening tests was a barrier. Only a few non-Hispanic whites reported that embarrassment would prevent them from obtaining cancer screening tests, but 10.6% of Hispanics agreed that embarrassment would prevent them from obtaining the needed screening, although these differences are unlikely to be statistically significant.
Table 4. Attitudes and Beliefs about Cancer Screening among Hispanic and Non-Hispanic Whitesa
|Structural barriers|| || || || || || || || |
| Unable to take time off work||17 (10.8)||4 (11.1)||37 (12.3)||22 (26.5)||17 (10.9)||5 (16.1)||38 (12.4)||22 (24.2)|
| No transportation to physician or clinic||17 (10.6)||12 (31.6)||35 (11.4)||20 (23.5)||17 (10.7)||6 (19.4)||37 (11.9)||27 (28.4)|
| Takes too long to wait at the clinic||24 (15.2)||11 (28.2)||45 (15.1)||21 (25.6)||22 (14.3)||6 (20.7)||49 (16.1)||27 (28.4)|
| Clinic staff do not speak Spanish||N/A||6 (15.8)||N/A||15 (17.4)||N/A||3 (10.0)||N/A||18 (18.6)|
| I do not have child care||15 (14.0)||1 (4.0)||25 (12.1)||8 (12.5)||15 (14.9)||1 (4.6)||25 (11.6)||8 (11.8)|
|Personal barriers|| || || || || || || || |
| Screening tests cost too much||62 (40.3)||16 (42.1)||123 (42.6)||35 (44.3)||51 (34.7)||12 (42.9)||133 (44.2)||41 (44.6)|
| I don't like being touched||25 (15.7)||8 (20.5)||66 (22.1)||14 (16.3)||27 (17.3)||3 (9.7)||64 (21.0)||20 (20.6)|
| I am afraid the physician will find cancer||34 (21.0)||5 (12.8)||72 (23.5)||19 (22.1)||34 (21.5)||2 (6.5)||73 (23.2)||22 (22.7)|
| I am afraid the physician will find other diseases||24 (15.3)||4 (10.3)||52 (17.0)||15 (17.0)||24 (15.4)||1 (3.2)||53 (17.1)||18 (18.2)|
| I am embarrassed about what my friends and family might think||6 (3.7)||3 (7.7)||9 (3.0)||11 (12.5)||5 (3.1)||0 (0.0)||11 (3.5)||14 (14.1)|
We evaluated associations of barriers to cancer screening with reported lifetime screening participation (data not shown). Hispanics who reported being unable to take time off work, not having child care, being afraid that the physician will find cancer, being afraid that the physician will find another disease, or being embarrassed about what friends and family might think were slightly less likely than those who did not, to obtain an FOBT, although none of the associations were significant. Modest and nonsignificant inverse associations were found between being unable to take time off work, lacking transportation, the cost of screening tests, not liking being touched, being afraid that the physician might find other diseases, and being embarrassed about what their friends and family might think and receipt of FOBT screening among non-Hispanic whites. For lifetime sigmoidoscopy/colonoscopy screening among Hispanics, modest and nonsignificant inverse associations were found for being unable to take time off work, lacking child care, being afraid that the physician will find cancer, and being afraid that the physician will find other disease. Hispanics who reported that they were embarrassed by what their friends and family might think were 20% less likely than those who did not report this barrier to obtain sigmoidoscopy/colonoscopy. However, after adjustment for the potentially confounding variables, associations were no longer statistically significant. Among non-Hispanic whites, no differences were found in the use of sigmoidoscopy/colonoscopy comparing those who did and did not report a given barrier.
In the current study, we used a random sample from a rural population to assess differences in prevalence of colorectal screening participation of Hispanics ≥ 50 years and non-Hispanic whites in the same age group. The findings suggest that Hispanics are less likely than non-Hispanic whites to have ever received an FOBT or sigmoidoscopy/colonoscopy. Our data further demonstrate that those with lower levels of education or lacking health care coverage are less likely to ever have been screened. Because there is considerable scientific interest in the factors that motivate Hispanics to participate in screening, the data are potentially important pieces of information that will assist in the development of intervention programs to improve screening adherence.
The relative lower prevalence of lifetime use of FOBT and sigmoidoscopy/colonoscopy among Hispanics is consistent with findings from national studies. Data from the Centers for Disease Control's National Health Interview Survey (NHIS) and Behavioral Risk Factor Surveillance System demonstrate that Hispanics have significantly lower proportions of lifetime screening for colorectal carcinoma than non-Hispanic whites.8, 12 Particularly noteworthy is the finding that Hispanic ethnicity was a significant predictor of screening in age-adjusted rates. Further, socioeconomic factors were related to screening prevalence. This suggests that the ethnicity—screening associations may be driven by socioeconomic factors. However, the correlation between ethnicity and socioeconomic status is high, suggesting we may be measuring a similar dimension of Hispanic ethnicity.
Contrary to expectation, the overall prevalence of lifetime FOBT screening reported in the current study (40.6% for Hispanics vs. 55.7% for non-Hispanic whites) was higher than the 2002 national estimates (34.5% for Hispanics vs. 46.7% for non-Hispanic whites).12 Of interest is when comparing incidence of unstaged and later-staged colorectal carcinoma in rural and urban communities, Hawley et al.19 observed a higher incidence in rural areas, suggesting lower screening prevalence.
It is noteworthy that among those who received an initial screening examination, no significant differences were found across ethnic groups in recent receipt of FOBT, which replicates the results from Perez-Stable et al.20 in which FOBT screening proportions within the past 2 years were 32.3% for Hispanics and 34.0% for non-Hispanic whites (nonsignificant). This suggests that initial screening may be a motivator for compliance with repeat screening.
Participants who lacked health care coverage or had completed fewer years of education were less likely than others to ever have participated in screening. This corroborates the findings of Swan et al.,21 who performed an analysis using data from the NHIS. Those data show that those least likely to have received a recent FOBT were those with no usual source of care, no health insurance, or were recent immigrants. Another study that revealed a direct association with education and income and screening participation was that of Hoffman-Goetz et al.13
Our data suggested an association between embarrassment about what friends and family might think and failure to receive a sigmoidoscopy/colonoscopy among Hispanics, but not among non-Hispanic whites. This finding is consistent with that reported by Bastani et al.,22 who carried out focus groups among an ethnically diverse sample of first-degree relatives of patients with colorectal carcinoma. Data from their study demonstrate that among Hispanic men specifically, the shame of being seen as sick or weak was a major disincentive to being screened for colorectal carcinoma. Other studies also have identified embarrassment as an impediment to colorectal screening.23, 24
The data for these analyses were collected using random sampling. Therefore, the data are generalizable to other rural communities in the area. The limitations of the current study should be noted. The low number of age-eligible participants limited our ability to assess the influence of certain factors, such as acculturation.
Our study is one of a few to evaluate the relation between screening barriers and receipt of screening services for colorectal carcinoma. The factors underlying the relative lower prevalence of lifetime screening participation among Hispanics deserve further investigation. Given the large proportion of Hispanics who lack health care coverage, it is likely that public approaches to expanding sources of coverage and care will need to be considered to reduce the disparity. Interventions designed to improve screening adherence in this group should target the initial screening examination, particularly among those lacking health care coverage or having few years of education.