Rates and predictors of colorectal cancer screening by race among motivated men participating in a prostate cancer risk assessment program

Authors

  • Michael J. Hall MD, MPH,

    1. Cancer Prevention and Control Program, Fox Chase Cancer Center, Philadelphia, Pennsylvania
    2. Department of Clinical Genetics, Fox Chase Cancer Center, Philadelphia, Pennsylvania
    Search for more papers by this author
  • Karen Ruth MS,

    1. Cancer Prevention and Control Program, Fox Chase Cancer Center, Philadelphia, Pennsylvania
    Search for more papers by this author
  • Veda N. Giri MD

    Corresponding author
    1. Cancer Prevention and Control Program, Fox Chase Cancer Center, Philadelphia, Pennsylvania
    2. Department of Clinical Genetics, Fox Chase Cancer Center, Philadelphia, Pennsylvania
    • Department of Clinical Genetics, Fox Chase Cancer Center, 333 Cottman Avenue, Philadelphia, PA 19111

    Search for more papers by this author
    • Fax: (215) 728-4061


Abstract

BACKGROUND:

Screening by fecal occult blood test and lower endoscopy has lowered colorectal cancer (CRC) mortality, but compliance gaps persist. Of concern are possible disparities in uptake of CRC screening between white and African American men. The goal of this study was to assess for disparities in uptake of CRC screening among men participating in a high-risk prostate cancer clinic. If present, such disparities could support hypotheses for further research examining racial differences in awareness and patient preferences in undergoing CRC screening.

METHODS:

Baseline data on a racially diverse cohort of men aged 50 to 69 years at increased risk of prostate cancer collected via the Prostate Cancer Risk Assessment Program at Fox Chase Cancer Center were analyzed. Predictors of uptake of CRC screening were assessed using multivariate logistic regression.

RESULTS:

Compared with whites, African American men had statistically significantly lower uptake of fecal occult blood testing (African American 49.0% vs white 60.7%, P = .035), lower endoscopy (African American 44.1% vs white 58.5%, P = .011), and any CRC screening (African American 66.2% vs white 76.3%, P = .053). Predictors of uptake of lower endoscopy among African American men included older age (odds ratio [OR], 3.61; 95% confidence interval [CI], 1.87-6.97), family history of CRC (OR, 3.47; 95% CI, 1.30-9.25), and insurance status (OR, 1.90; 95% CI, 1.04-3.46).

CONCLUSIONS:

Despite awareness of cancer risk and motivation to seek prostate cancer screening through a specialized prostate cancer risk assessment program, evidence supporting compliance gaps with CRC screening among men was found. Tailored messages to younger African American men with and without a family history of CRC are needed. Cancer 2011;. © 2011 American Cancer Society.

In 2010, more than 145,000 new cases of colorectal cancer (CRC) were diagnosed in Americans, with more than half of these in men. African American men have been shown to have higher incidence of CRC and poorer survival than Caucasian counterparts, highlighting the need to improve uptake of CRC screening across the diverse US population.1, 2

The proportion of Americans reporting ever having been screened for CRC with fecal occult blood test (FOBT) and lower endoscopy has increased dramatically in recent years,3, 4 but compliance gaps associated with disparities in CRC screening uptake persist. Studies of predictors of CRC screening compliance have identified a variety of demographic, family history, and sociodemographic barriers to screening compliance among the general public and in men in particular.3-11

Compliance with CRC screening relative to the pursuit of other preventive health care has also been examined in various contexts. Studies of the National Health Interview Survey and Behavioral Risk Factor Surveillance System (BRFSS)3-5, 10, 12-14 have found that individuals who have a primary care provider or seek preventive care are more complaint with CRC screening.3, 4 Among men, prostate-specific antigen (PSA) screening for prostate cancer has been shown to be the strongest predictor of CRC screening.10, 14, 15

Among other factors, research programs directed to the study of prostate cancer risk and public health campaigns have raised awareness of prostate cancer and prostate cancer screening among the general public. It is currently unknown whether participation in a multidisciplinary prostate cancer screening program would be associated with compliance with CRC screening, and whether particular patient characteristics might be associated with compliance. Recent data suggest high rates of CRC screening noncompliance among participants in a free PSA screening program. The current study builds on previous research to examine self-reported CRC screening compliance in a racially diverse cohort of men at increased risk for prostate cancer enrolled in a tertiary care multidisciplinary prostate cancer screening program, the Fox Chase Cancer Center Prostate Cancer Risk Assessment Program. We hypothesize that baseline differences by race in uptake of CRC screening in men will be present in our cohort despite self-selection as a group of individuals interested in cancer prevention/screening and with demonstrated access to preventive health care. Our goal was to examine rates of compliance with guidelines for CRC screening in this group as a whole, and if racial differences in screening compliance were found, to examine predictors of disparities in screening that may be uniquely found in this population.

MATERIALS AND METHODS

Prostate Cancer Risk Assessment Program Procedures and Eligibility

The design of the Prostate Cancer Risk Assessment Program has been described previously.16, 17 Briefly, the Prostate Cancer Risk Assessment Program was established in 1996 with the objectives of providing prostate cancer education, risk assessment, and early detection for men at high risk and to research genetic and epidemiologic factors contributing to the increased susceptibility to prostate cancer. Eligibility criteria include men aged 35 to 69 years with any of the following: 1) 1 first-degree relative with prostate cancer, 2) 2 second-degree relatives with prostate cancer on the same side of the family, 3) any African American man regardless of a family history of prostate cancer, or 4) men with known mutations in BRCA1 or BRCA2. Recruitment of Prostate Cancer Risk Assessment Program participants has been reported in detail16 and includes physician referrals, radio advertisements in the greater Philadelphia area, referrals by family/friends, and community events/health fairs. In addition to Fox Chase Cancer Center, participants are currently also recruited at 2 community Fox Chase Cancer Center Partner hospitals. Insurance information was not routinely collected on patients before November 2003, as all services were provided to participants free of charge (insurance-blind period). Beginning in November 2003 (insurance-mandatory period), health insurance became a requirement to receive prostate cancer screening and follow-up services through the Prostate Cancer Risk Assessment Program. Approximately 16% (n = 27 of 164 patients accrued after November 2003) were uninsured but were able to participate in the Prostate Cancer Risk Assessment Program because of the availability of limited external funding to cover the costs of prostate cancer screening during the insurance-mandatory period.

Prostate Cancer Risk Assessment Program Measures and Data Management

The variables assessed in this analysis included race, age at entry into the Prostate Cancer Risk Assessment Program, marital status, education, income, body mass index (BMI), history of colon polyps, family history of a first-degree relative with CRC, insurance status (insurance-blind period vs insurance-mandatory period), and history of CRC screening with FOBT or sigmoidoscopy/colonoscopy. Race and age are self-reported and obtained via phone at the time of enrollment in the Prostate Cancer Risk Assessment Program. Since November 2003, health insurance coverage has also been verified at intake. Marital status, education, income, history of colon polyps, family history of cancer, and colon cancer screening history are obtained by self-report using a Health History Questionnaire completed by participants and acquired at the initial Prostate Cancer Risk Assessment Program screening visit. BMI is calculated using the participants' height and weight, which are measured at their initial visit and at each subsequent follow-up visit. All data are entered via Oracle Forms 6i (client-server) or 10g (Web-enabled) applications used by the Population Studies Facility and Prostate Cancer Risk Assessment Program staff at Fox Chase Cancer Center. The data are maintained in an Oracle 10g database.

Informed consent is obtained from all Prostate Cancer Risk Assessment Program participants, and the Prostate Cancer Risk Assessment Program study has been approved by the institutional review board at Fox Chase Cancer Center.

Prostate Cancer Risk Assessment Program Participants

As of April 1, 2010, 812 participants were enrolled in the Prostate Cancer Risk Assessment Program, and 339 men were included in this analysis. Exclusions were as follows: men outside of the recommended age range for CRC screening (younger than 50 years) (n = 310), men of other self-reported race (n = 8), no health history questionnaire completed (n = 33), men with incomplete data on marital, education, or income status (n = 27), and men with a reported history of colitis, as these men were likely undergoing colon surveillance for colitis (n = 8).

Statistical Analyses

We used Fisher exact test to compare differences by race in categorical baseline variables, including demographic factors, CRC risk factors, and screening uptake. We used multivariable logistic regression to identify predictors of reported CRC screening including age, family history of CRC, marital status, and either income or education, which were highly correlated. Separate models were run for each outcome (FOBT screening, colonoscopy screening, and any CRC screening) within each racial group (African American and white), and only significant variables were kept in the final model, using P < .05 as level of statistical significance and P < .10 as borderline significant. To account for the transition from the insurance-blind to the insurance-mandatory time period, a binary variable was created indicating the time periods before and after the health insurance requirement was instituted for Prostate Cancer Risk Assessment Program participation (November 2003). To explore the possible impact of the American Gastroenterological Association recommendation for expanded screening of African American men aged >45 years in 2006, additional data on African American men aged >45 years enrolled in the Prostate Cancer Risk Assessment Program after 2006 are also examined. Analyses for this paper were generated using SAS/STAT software, version 9.1 of the SAS System for Windows (SAS Institute, Cary, NC).

RESULTS

Demographic Factors

Demographic characteristics of participants included in the current analysis are reported in Table 1. African American men represented 60.2% (204 of 339) and white men 39.8% (135 of 339) of the study cohort. CRC screening and demographic data were stratified by race, and baseline differences are reported and compared by Fisher exact test. White participants were more likely to be married (P < .0001), more likely to have an advanced or college degree (P < .0001), and reported a higher annual income (P < .0001).

Table 1. Demographic Characteristics of 339 Prostate Cancer Risk Assessment Program Participants Age 50+ Years by Race
Baseline CharacteristicsAfrican American, n=204White, n=135Comparison, Fisher Exact Test, P
No.%No.%
  • Abbreviations: BMI, body mass index; CRC, colorectal cancer; FDR, first-degree relative; FOBT, fecal occult blood test; GED, general equivalency diploma; HS, high school; SDR, second-degree relative.

  • a

    Before November 15, 2003: insurance-blind period.

  • b

    After November 15, 2003: insurance-mandatory period.

Age, y    .1015
 50 to 5914269.68260.7 
 60 to 696230.45339.3 
Marital status    <.0001
 Married11556.411081.5 
 Divorced/widowed/separated6330.92014.8 
 Never married/single2612.853.7 
Education    <.0001
 ≤HS/GED5326.02921.5 
 Some college7938.72518.5 
 Tech/bachelor4220.65037.0 
 Graduate3014.73123.0 
Income    <.0001
 <$30K6431.41712.6 
 $30-50K6933.83022.2 
 $50-75K4120.13928.9 
 >$75K2311.33525.9 
 Refused/do not know73.41410.4 
BMI, kg/m2    .0003 (missing=7)
 Underweight, <18.510.500 
 Normal, 18.5 to <252512.64029.9 
 Overweight, 25 to <3010050.56347.0 
 Obese ≥307236.43123.1 
History of polyps    .1677 (missing=31)
 Yes2715.23123.9 
 No14279.89371.5 
 Do not know95.164.6 
Family history of CRC (any FDR or SDR)    .0214
 Yes2411.82921.5 
 No18088.210678.5 
Insurance status     
 No insurance requirementa10551.57051.91.000
 Insurance requiredb9948.56548.1 
FOBT    .0354
 Yes10049.08260.7 
 No, do not know, missing10451.05339.3 
Colonoscopy/sigmoidoscopy    .0108
 Yes9044.17958.5 
 No, do not know, missing11455.95641.5 
Any CRC screening    .0525
 Yes13566.210376.3 
 No6933.83223.7 

CRC Risk Factors

Differences in the distribution of body mass indices by race were seen (Table 1), with obesity more common among African American than white participants (36.4% vs 23.1%, P < .0001). White participants were more likely to report a family history of CRC (21.5% vs 11.8%, P = .021).

CRC Screening Uptake

Fewer African American participants reported ever having had CRC screening compared with Caucasian participants (P = .053). African American men reported lower rates of ever having had an FOBT (49.0% for African American men vs 60.7% for Caucasian men, P = .035) or having undergone a screening sigmoidoscopy or colonoscopy (44.1% for African American men vs 58.5% for Caucasian men, P = .011) (Table 1). CRC screening uptake with respect to insurance requirements in the Prostate Cancer Risk Assessment Program are shown in Table 2. Before any insurance requirement (insurance-blind period), African American men reported lower rates of colonoscopy/sigmoidoscopy relative to white participants (relative risk [RR], 0.67; 95% confidence interval [CI], 0.48-0.92). Even in the insurance-mandatory period, fewer African American participants reported any CRC screening compared with white participants (RR, 0.84; 95% CI, 0.71-0.98).

Table 2. Colorectal Cancer Screening Rates by Race Prior to November 15, 2003 (Insurance-Blind Period) and After November 15, 2003 (Insurance-Mandatory Period)
ProcedureAAWhiteRelative Risk
No.Screened% ScreenedNo.Screened% ScreenedAA to White95% CI
  • Abbreviations: AA, African American; CRC, colorectal cancer; FOBT, fecal occult blood test.

  • a

    Statistically significant differences by race.

FOBT        
 Insurance-blind period1055249.5704361.40.8060.617-1.053
 Insurance-mandatory period994848.5653960.00.8080.608-1.073
Colonoscopy/sigmoidoscopy        
 Insurance-blind period1053937.1703955.70.667a0.482-0.923a
 Insurance-mandatory period995151.5654061.50.8370.698-1.098
Any CRC screening        
 Insurance-blind period1056561.9704868.60.9030.726-1.123
 Insurance-mandatory period997070.7655584.60.836a0.709-0.984a

Logistic Regression Analysis of Predictors of CRC Screening

Multivariable logistic regression was used to examine predictors of uptake of CRC screening in the Prostate Cancer Risk Assessment Program population by race (Table 3). Self-reported educational attainment as an indicator of socioeconomic status was chosen for logistic modeling rather than self-reported income. Individuals are more likely to inaccurately report income relative to education, and education has previously been identified as a stronger predictor of CRC screening uptake.18 Marital status was insignificant in all models (across all outcomes) and was therefore not included in the final models.

Table 3. Predictors of CRC Screening Among 339 Prostate Cancer Risk Assessment Program Participants by Race
PredictorFOBTColonoscopy/SigmoidoscopyAny CRC Screening
AAWhiteAAWhiteAAWhite
OR95% CIOR95% CIOR95% CIOR95% CIOR95% CIOR95% CI
  • Abbreviations: AA, African American; CI, confidence interval; CRC, colorectal cancer; FH, family history; FOBT, fecal occult blood test; OR, odds ratio.

  • a

    Statistically significant differences by race.

  • b

    Reference is high school education/general equivalency diploma.

  • c

    Insurance-mandatory period (after November 2003) versus insurance-blind period (prior to November 2003).

Age, 60-69 years vs 50-59 years1.250.68-2.301.610.75-3.433.61a1.87-6.97a1.550.73-3.291.630.82-3.221.910.76-4.79
Educationb            
Some college0.840.41-1.721.170.39-3.551.120.52-2.420.530.17-1.611.180.57-2.470.920.27-3.16
Tech/college1.330.58-3.021.940.74-5.071.820.75-4.410.800.31-2.052.74a1.07-7.02a1.710.56-5.22
Graduate1.410.57-3.551.020.36-2.861.320.49-3.541.640.55-4.902.160.79-5.922.020.58-7.06
FH of CRC, yes vs no2.320.93-5.781.690.67-4.213.47a1.30-9.25a1.380.57-3.352.640.85-8.271.690.55-5.16
Insurance statusc0.950.54-1.670.850.41-1.751.90a1.04-3.46a1.340.65-2.751.540.84-2.832.60a1.08-6.24a

Among African American participants, older age (odds ratio [OR], 3.61; 95% CI, 1.87-6.97), family history of CRC (OR, 3.47; 95% CI, 1.30-9.25), and enrollment during the insurance-mandatory period (OR, 1.90; 95% CI, 1.04-3.46) were significantly associated with increased rates of colonoscopy/sigmoidoscopy. Among African American participants, college education (OR, 2.74; 95% CI, 1.07-7.02) was significantly associated with any CRC screening, whereas enrollment in the insurance-mandatory period was associated among white participants (OR, 2.60; 95% CI, 1.08-6.24). No predictors emerged for uptake of FOBT by race.

An additional exploratory analysis to accommodate for changes in CRC screening guidelines by the American Gastroenterology Society in March 200519 recommending lowering the age of initiation of routine CRC screening to age 45 years did not significantly impact the study results.

DISCUSSION

Population-based studies performed in the United States have demonstrated persistent disparities in compliance with recommended CRC screening, despite recent nationwide gains in reported uptake of FOBT and lower endoscopy. Factors that may explain these differences are many, and may include racial/ethnic differences in beliefs relative to CRC screening (eg, fatalism, preferences for cancer screening), access (eg, medical insurance coverage or access to specialty physicians), and others (eg, awareness of CRC screening). Multivariate models examining disparities in CRC screening within federally insured Americans (the Medicare and Veteran's Administration populations) have not supported the existence of CRC screening disparities after accounting for other demographic and socioeconomic factors.20-22

The Fox Chase Cancer Center Prostate Cancer Risk Assessment Program population includes adult men from the Philadelphia region who have chosen to enroll in a clinical research program that offers education, risk assessment, and yearly screening for prostate cancer. Our results suggesting lower uptake of CRC screening in African American participants compared with white participants, particularly among younger African American participants, are intriguing for several reasons. The same men who chose to participate in the Prostate Cancer Risk Assessment Program because of a concern about prostate cancer demonstrated marked baseline deficiencies in compliance with CRC screening. One possible explanation for the findings is that Prostate Cancer Risk Assessment Program participants, composed of self-selected African American and white men at increased risk for prostate cancer, consider prostate cancer to be a greater personal health threat than CRC, and thus have chosen to pursue prostate cancer screening first and foremost. Public health messages emphasizing the importance of family history to health risk and the increased risk of prostate cancer among African American men relative to others may to some degree contribute to preventive behaviors more focused on 1 high-threat entity (here, prostate cancer) versus more comprehensive preventive health behavior. That is, the Prostate Cancer Risk Assessment Program men, in the face of limited time and resources, may choose to focus preventive efforts on their greatest perceived health risk. Details related to the CRC screening modality, such as variability in perceived invasiveness or ease of obtaining/completing screening (eg, PSA blood test and digital rectal exam), could explain documented differences in compliance with lower endoscopy. It should be noted that the reported data also represent baseline compliance with CRC screening; it is possible that known racial/ethnic differences in awareness of CRC screening (ie, awareness of CRC screening has generally been found to be higher among whites than African Americans) contributed to baseline compliance differences. It is standard practice to educate Prostate Cancer Risk Assessment Program participants during their clinical visit of all age-appropriate cancer screening, including CRC. A future analysis will examine 1-year follow-up compliance with CRC screening in this population.

Only 42% of US population was up to date with CRC screening in 2003.3 Data from the BRFSS have demonstrated lower screening rates among African American respondents compared with Caucasian respondents at each of the 2002, 2004, and 2006 survey time points (ie, FOBT within 1 year and/or sigmoidoscopy or colonoscopy within 10 years).23 However, studies in Medicare beneficiaries20 and individuals receiving health care through the Veterans Affairs Administration system12, 21 have not found racial/ethnic screening disparities. Several studies have examined predictors of CRC screening in the US population eligible for screening.3-6 Older age, family history of CRC, and being married have been positively associated with screening compliance in several studies. Men specifically may be less aware of and less likely to undergo FOBT relative to endoscopic screening.4, 5, 7-9 Men with lower educational attainment, however, may be specifically prone to screening noncompliance.10 Health insurance also appears to play an important role in screening behaviors,3, 4, 10, 11 whereas the role of income per se is less clear.11

Several studies have examined the relationship between the uptake of 1 cancer screening (eg, PSA screening for prostate cancer or mammography for breast cancer screening) and its impact on CRC screening.14, 15, 18 Examining data collected via the BRFSS, Carlos et al14 found that undergoing PSA screening was a significant predictor of CRC screening uptake. A similar association was found by this group with mammography18 and in an analysis of updated data by Wong and Coups.15 Our data are most readily compared with those recently published by Red et al.24 In their multivariate analysis of 331 adult men (60% African American) undergoing free prostate cancer screening through a clinic at Georgetown Medical Center, these investigators did not find race to be predictive of CRC screening compliance, but, similar to the analyses of the BRFSS by Carlos et al, found prostate cancer screening, health insurance, and having a regular primary care provider to be predictive of both ever having had and adherence to colorectal cancer screening. Of note, although similar in racial composition to our cohort, the participants in the Georgetown study are different from those examined in our research, as our population enrolls in the Prostate Cancer Risk Assessment Program for education, clinical surveillance, and research participation relative to prostate cancer risk assessment. Thus, noncompliance with CRC screening would seem even more surprising among our participants than in the Georgetown study. Taken together, these data hint at the complexity of decision making to pursue preventive medical care such as cancer screening among the public, with decisions seemingly related to awareness of screening, known behavioral risks, perceived threat of cancer, ease of testing, and time commitments. To this, our data extend observations to a small yet racially diverse, motivated, high-risk screening population, further enriching hypotheses relative to uptake of available preventive care by adults faced with a multitude of age-related health risks and health maintenance recommendations.

There are some limitations to this analysis that are important to consider when drawing conclusions from these data. Compliance data for FOBT and colonoscopy are self-reported, and thus are subject to reporting error. In particular, it is conceivable that patients may experience difficulty recalling a history of an FOBT, particularly as they age. However, the younger age of African American participants in this analysis would tend to diminish the differences seen. Because health insurance was not required for participation in the Prostate Cancer Risk Assessment Program before November 2003, we examined the impact of health insurance on uptake of CRC screening before and after November 2003. Although the time period after institution of a mandatory insurance requirement for the Prostate Cancer Risk Assessment Program is associated with higher CRC screening in African American and white participants, there were also temporal trends toward increasing lower endoscopy uptake nationwide. However, African American men still have lower rates of CRC screening compared with white participants in the Prostate Cancer Risk Assessment Program. Therefore, these results deserve further study. The generalizability of these findings may be limited because of the highly specialized nature of the Prostate Cancer Risk Assessment Program cohort. Nonetheless, that men with actual and perceived high-risk of prostate cancer do not complete routine age-appropriate CRC screening with significant variability by race remains a compelling and concerning observation.

Future studies are needed to better understand how individuals process health risk-related information, how they make decisions about their own health care, and how better to deliver recommendations for cancer screening in African American men. In particular, our results point to the need to tailor messages to younger, yet screening age-appropriate African American men to motivate CRC screening uptake.

FUNDING SOURCES

Supported by American Cancer Society grant MRSG-07-232-01-CPHPS to M.J.H. and Department of Defense grant W81XWH-09-1-0302 to V.N.G.; M.J.H., K.R., and V.N.G. receive additional research support through the Fox Chase Cancer Center Core Grant (National Cancer Institute P30 CA006927).

CONFLICT OF INTEREST DISCLOSURES

The authors made no disclosures.

Ancillary