J Am Board Fam Med: first published as 10.3122/jabfm.2008.05.070266 on 4 September 2008. Downloaded from http://www.jabfm.org/ on 6 May 2026 by guest. Protected by copyright.
Factors Associated with Racial/Ethnic Differences in
Colorectal Cancer Screening
Navkiran K. Shokar, MD, MPH, Carol A. Carlson, BA, and Susan C. Weller, PhD
Introduction: Racial/ethnic differences in colorectal cancer (CRC) screening rates are thought to ac-
count, in part, for the racial/ethnic differences in CRC disease burden. The purpose of this study was to
examine which factors mediate racial/ethnic differences in CRC screening.
Methods: Five hundred sixty participants attending a primary care clinic, aged 50 to 80 years, and of
African-American, Hispanic, or non-Hispanic white race/ethnicity were interviewed. The goal was to as-
sess the contribution of sociodemographic characteristics, knowledge, beliefs about CRC, and the health
care experience with their primary care doctor to racial/ethnic differences in CRC screening. The out-
come variable was self-reported screening. All analyses were weighted; bivariate testing and multivari-
ate logistic regression was conducted.
Results: The response rate was 55.7%, with no sociodemographic differences noted between respon-
dents and nonrespondents. Respondents were African-American (n ⴝ 194), Hispanic (n ⴝ 162), and
non-Hispanic white (n ⴝ 204); 64.5% were aged 50 to 64 years; 63.1% were women; 96.9% were in-
sured; and over half reported a total annual income of less than $25,000. Overall 62.5% were current
with CRC screening: 67.5% of non-Hispanic whites, 54.3% of African-Americans, and 48.6% of Hispanics
(P < .001). A doctor’s recommendation (odds ratio, 3.86); awareness of screening (odds ratio, 3.32);
older age (odds ratio, 2.88); greater education (odds ratio, 2.02); and perceived susceptibility (odds
ratio, 1.74) contributed to racial/ethnic differences in CRC screening.
Conclusions: Interventions to address CRC screening disparities among racial/ethnic groups should
focus on the health care setting and patient education about CRC screening; differences in attitudes and
beliefs seem to be less important. (J Am Board Fam Med 2008;21:414 – 426.)
Colorectal cancer (CRC) is the second leading However, despite the evidence and recommen-
cause of cancer deaths in the United States.1 The dations, screening rates remain low9 –23 and are
American Cancer Society anticipates 148,810 consistently lower in minority groups.9,11,12,20 –24
new cases and 49,960 deaths from this disease in To understand low screening rates, investigators
2008.2 Although CRC affects all segments of the have examined the association between CRC
population, minority groups experience a greater screening and a multitude of factors including de-
burden of disease, with the highest incidence and mographics, health care access, knowledge, atti-
mortality occurring among African-Americans.1 tudes and beliefs, medical history, and other pre-
CRC screening facilitates early detection and ventive behaviors.11–13,15,18 –20,22–38 Studies and
treatment and has resulted in a reduction in CRC interventions focusing on disparities in CRC
incidence and mortality; it is therefore widely screening have tended to focus on a single racial/
recommended for people aged 50 and older.3– 8
ethnic group39 – 43 and have shown mixed results.
There remains a need for studies that examine a
wide array of factors simultaneously in diverse mul-
This article was externally peer reviewed.
Submitted 29 November 2007; revised 14 March 2008; tiethnic populations and that compare associations
accepted 1 April 2008. across racial/ethnic groups to determine which un-
From the Department of Family Medicine (NKS, CAC,
SCW) and the Department of Preventive Medicine and derlying factors explain the differences in CRC
Community Health (SCW), University of Texas Medical screening across groups.
Branch, Galveston.
Funding: John Sealy Memorial Endowment Fund for Bio- The information gained will guide interven-
medical Research NCI K07 CA107052-01 tions designed to reduce racial/ethnic disparities
Conflict of interest: none declared.
Corresponding author: Navkiran K Shokar, MD, MPH, in CRC screening. The purpose of the present
Department of Family Medicine, University of Texas Med- study, therefore, was to (1) simultaneously exam-
ical Branch, 301 University Boulevard, Galveston, TX
77555-1123 (E-mail: ). ine the association between a wide range of vari-
414 JABFM September–October 2008 Vol. 21 No. 5 http://www.jabfm.org
, J Am Board Fam Med: first published as 10.3122/jabfm.2008.05.070266 on 4 September 2008. Downloaded from http://www.jabfm.org/ on 6 May 2026 by guest. Protected by copyright.
ables and CRC screening in a multiethnic popu- previous studies or were found to be important in
lation of African-American, Hispanic, and non- our previous qualitative work in the same popula-
Hispanic white participants attending a primary tion44 and if they were relevant for the practice
care clinic, and (2) to determine the contribution setting. Items were developed to be culturally and
of these factors to racial/ethnic differences in linguistically appropriate, and a Spanish language
CRC screening. version of the items was developed using standard
methods.45 The final instrument consisted of items
Methods organized into the following categories: (1) socio-
Patients and Setting demographic characteristics, (2) knowledge and be-
Patients were recruited from a University-based liefs about CRC and CRC screening, (3) medical
family medicine clinic in Southeast Texas during a history, and (4) health care experience variables.
16-month period in 2004 and 2005. Although sur- The outcome variable was self-reported history of
veys conducted over large scale regions offer infor- CRC screening.
mation about a larger population, a primary care
setting was chosen as the site for this study because Measures
multiple topics could be handled in-depth and most The outcome variable, self-reported CRC
CRC screening delivery is initiated in outpatient screening, was assessed with validated items
primary care settings. adapted from Vernon et al.46 A detailed descrip-
The clinic serves a diverse mix of racial/ethnic tion of each test preceded each question; descrip-
groups from both urban and semirural areas; tions were pilot tested for comprehension and
there is an excess of 40, 000 visits per year. The feedback was incorporated into the final version.
racial/ethnic distribution of patients over the age Current screening was determined by whether
of 50 years is 66% non-Hispanic white, 24% the patient reported that they had undergone any
African-American, and 10% Hispanic. More than of the recommended tests, for any reason accord-
90% are insured. To be eligible, patients had to ing to guidelines current at the time, as follows:
be 50 to 80 years of age and of non-Hispanic annual home fecal occult blood testing or flexible
white, African-American, or Hispanic race/eth- sigmoidoscopy every 5 years; or annual fecal oc-
nicity. Individuals with a history of CRC or high cult blood testing plus flexible sigmoidoscopy
risk of CRC (familial adenomatous polyposis syn- every 5 years; or double contrast barium enema
drome, hereditary nonpolyposis CRC, or ulcer- every 5 years; or colonoscopy every 10 years.6,7 The
ative colitis) were excluded. sociodemographic items were adapted from a na-
A stratified sampling scheme, balanced by tional survey47 and elicited information about the
race/ethnicity; age (⬍65, ⱖ65); and sex was in- patient’s age, educational level, sex, income, and in-
stituted to increase the statistical power for com- surance type. Race and ethnicity was self-reported
parisons across racial/ethnic groups and the older and elicited with a 2-part question consistent with
age group. Interviewers recruited patients for Federal criteria.48
each stratum until the target number was A 21-item knowledge test was developed cov-
reached. Interviewers were bilingual (Spanish ering CRC prevalence (1 item), symptoms (5
and English). Interviewers approached patients items), risk factors (13 items), screening test
attending the clinic for any reason and invited availability (1 item), and treatment (1 item).
them to participate in the study. The study was These items were developed de novo based on
approved by the University of Texas Medical the findings from prior studies and qualitative
Branch Institutional Review Board and informed interviews in this population.44,49,50 An addi-
written consent was obtained from each patient. tional question assessed awareness of screening.
Interviews were conducted in a private room We utilized the Health Belief model to examine
around the time of the doctor visit and lasted attitudes and beliefs about CRC screening be-
approximately 45 minutes. cause it provides a useful organizing framework
for explaining screening behavior.15,18,27,29 –35,50 –54
Survey Development Scales were developed specifically for this multi-
Variables were chosen for inclusion in the study if ethnic population and were tested to assess the
they had been correlated with CRC screening in constructs of perceived susceptibility (4 items),
doi: 10.3122/jabfm.2008.05.070266 Racial/Ethnic Differences in Colorectal Cancer Screening 415
Factors Associated with Racial/Ethnic Differences in
Colorectal Cancer Screening
Navkiran K. Shokar, MD, MPH, Carol A. Carlson, BA, and Susan C. Weller, PhD
Introduction: Racial/ethnic differences in colorectal cancer (CRC) screening rates are thought to ac-
count, in part, for the racial/ethnic differences in CRC disease burden. The purpose of this study was to
examine which factors mediate racial/ethnic differences in CRC screening.
Methods: Five hundred sixty participants attending a primary care clinic, aged 50 to 80 years, and of
African-American, Hispanic, or non-Hispanic white race/ethnicity were interviewed. The goal was to as-
sess the contribution of sociodemographic characteristics, knowledge, beliefs about CRC, and the health
care experience with their primary care doctor to racial/ethnic differences in CRC screening. The out-
come variable was self-reported screening. All analyses were weighted; bivariate testing and multivari-
ate logistic regression was conducted.
Results: The response rate was 55.7%, with no sociodemographic differences noted between respon-
dents and nonrespondents. Respondents were African-American (n ⴝ 194), Hispanic (n ⴝ 162), and
non-Hispanic white (n ⴝ 204); 64.5% were aged 50 to 64 years; 63.1% were women; 96.9% were in-
sured; and over half reported a total annual income of less than $25,000. Overall 62.5% were current
with CRC screening: 67.5% of non-Hispanic whites, 54.3% of African-Americans, and 48.6% of Hispanics
(P < .001). A doctor’s recommendation (odds ratio, 3.86); awareness of screening (odds ratio, 3.32);
older age (odds ratio, 2.88); greater education (odds ratio, 2.02); and perceived susceptibility (odds
ratio, 1.74) contributed to racial/ethnic differences in CRC screening.
Conclusions: Interventions to address CRC screening disparities among racial/ethnic groups should
focus on the health care setting and patient education about CRC screening; differences in attitudes and
beliefs seem to be less important. (J Am Board Fam Med 2008;21:414 – 426.)
Colorectal cancer (CRC) is the second leading However, despite the evidence and recommen-
cause of cancer deaths in the United States.1 The dations, screening rates remain low9 –23 and are
American Cancer Society anticipates 148,810 consistently lower in minority groups.9,11,12,20 –24
new cases and 49,960 deaths from this disease in To understand low screening rates, investigators
2008.2 Although CRC affects all segments of the have examined the association between CRC
population, minority groups experience a greater screening and a multitude of factors including de-
burden of disease, with the highest incidence and mographics, health care access, knowledge, atti-
mortality occurring among African-Americans.1 tudes and beliefs, medical history, and other pre-
CRC screening facilitates early detection and ventive behaviors.11–13,15,18 –20,22–38 Studies and
treatment and has resulted in a reduction in CRC interventions focusing on disparities in CRC
incidence and mortality; it is therefore widely screening have tended to focus on a single racial/
recommended for people aged 50 and older.3– 8
ethnic group39 – 43 and have shown mixed results.
There remains a need for studies that examine a
wide array of factors simultaneously in diverse mul-
This article was externally peer reviewed.
Submitted 29 November 2007; revised 14 March 2008; tiethnic populations and that compare associations
accepted 1 April 2008. across racial/ethnic groups to determine which un-
From the Department of Family Medicine (NKS, CAC,
SCW) and the Department of Preventive Medicine and derlying factors explain the differences in CRC
Community Health (SCW), University of Texas Medical screening across groups.
Branch, Galveston.
Funding: John Sealy Memorial Endowment Fund for Bio- The information gained will guide interven-
medical Research NCI K07 CA107052-01 tions designed to reduce racial/ethnic disparities
Conflict of interest: none declared.
Corresponding author: Navkiran K Shokar, MD, MPH, in CRC screening. The purpose of the present
Department of Family Medicine, University of Texas Med- study, therefore, was to (1) simultaneously exam-
ical Branch, 301 University Boulevard, Galveston, TX
77555-1123 (E-mail: ). ine the association between a wide range of vari-
414 JABFM September–October 2008 Vol. 21 No. 5 http://www.jabfm.org
, J Am Board Fam Med: first published as 10.3122/jabfm.2008.05.070266 on 4 September 2008. Downloaded from http://www.jabfm.org/ on 6 May 2026 by guest. Protected by copyright.
ables and CRC screening in a multiethnic popu- previous studies or were found to be important in
lation of African-American, Hispanic, and non- our previous qualitative work in the same popula-
Hispanic white participants attending a primary tion44 and if they were relevant for the practice
care clinic, and (2) to determine the contribution setting. Items were developed to be culturally and
of these factors to racial/ethnic differences in linguistically appropriate, and a Spanish language
CRC screening. version of the items was developed using standard
methods.45 The final instrument consisted of items
Methods organized into the following categories: (1) socio-
Patients and Setting demographic characteristics, (2) knowledge and be-
Patients were recruited from a University-based liefs about CRC and CRC screening, (3) medical
family medicine clinic in Southeast Texas during a history, and (4) health care experience variables.
16-month period in 2004 and 2005. Although sur- The outcome variable was self-reported history of
veys conducted over large scale regions offer infor- CRC screening.
mation about a larger population, a primary care
setting was chosen as the site for this study because Measures
multiple topics could be handled in-depth and most The outcome variable, self-reported CRC
CRC screening delivery is initiated in outpatient screening, was assessed with validated items
primary care settings. adapted from Vernon et al.46 A detailed descrip-
The clinic serves a diverse mix of racial/ethnic tion of each test preceded each question; descrip-
groups from both urban and semirural areas; tions were pilot tested for comprehension and
there is an excess of 40, 000 visits per year. The feedback was incorporated into the final version.
racial/ethnic distribution of patients over the age Current screening was determined by whether
of 50 years is 66% non-Hispanic white, 24% the patient reported that they had undergone any
African-American, and 10% Hispanic. More than of the recommended tests, for any reason accord-
90% are insured. To be eligible, patients had to ing to guidelines current at the time, as follows:
be 50 to 80 years of age and of non-Hispanic annual home fecal occult blood testing or flexible
white, African-American, or Hispanic race/eth- sigmoidoscopy every 5 years; or annual fecal oc-
nicity. Individuals with a history of CRC or high cult blood testing plus flexible sigmoidoscopy
risk of CRC (familial adenomatous polyposis syn- every 5 years; or double contrast barium enema
drome, hereditary nonpolyposis CRC, or ulcer- every 5 years; or colonoscopy every 10 years.6,7 The
ative colitis) were excluded. sociodemographic items were adapted from a na-
A stratified sampling scheme, balanced by tional survey47 and elicited information about the
race/ethnicity; age (⬍65, ⱖ65); and sex was in- patient’s age, educational level, sex, income, and in-
stituted to increase the statistical power for com- surance type. Race and ethnicity was self-reported
parisons across racial/ethnic groups and the older and elicited with a 2-part question consistent with
age group. Interviewers recruited patients for Federal criteria.48
each stratum until the target number was A 21-item knowledge test was developed cov-
reached. Interviewers were bilingual (Spanish ering CRC prevalence (1 item), symptoms (5
and English). Interviewers approached patients items), risk factors (13 items), screening test
attending the clinic for any reason and invited availability (1 item), and treatment (1 item).
them to participate in the study. The study was These items were developed de novo based on
approved by the University of Texas Medical the findings from prior studies and qualitative
Branch Institutional Review Board and informed interviews in this population.44,49,50 An addi-
written consent was obtained from each patient. tional question assessed awareness of screening.
Interviews were conducted in a private room We utilized the Health Belief model to examine
around the time of the doctor visit and lasted attitudes and beliefs about CRC screening be-
approximately 45 minutes. cause it provides a useful organizing framework
for explaining screening behavior.15,18,27,29 –35,50 –54
Survey Development Scales were developed specifically for this multi-
Variables were chosen for inclusion in the study if ethnic population and were tested to assess the
they had been correlated with CRC screening in constructs of perceived susceptibility (4 items),
doi: 10.3122/jabfm.2008.05.070266 Racial/Ethnic Differences in Colorectal Cancer Screening 415
