CEBP CTRC-AACR San Antonio Breast Cancer Symposium Translational Cancer Medicine 2008: Cancer Clinical Trials and Personalized Medicine
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
Cancer Research Clinical Cancer Research
Cancer Epidemiology Biomarkers & Prevention Molecular Cancer Therapeutics
Molecular Cancer Research Cancer Prevention Research
Cancer Prevention Journals Portal Cancer Reviews Online
Annual Meeting Education Book Cell Growth & Differentiation
Cancer Epidemiology Biomarkers & Prevention 16, 500-509, March 1, 2007. doi: 10.1158/1055-9965.EPI-06-0757
© 2007 American Association for Cancer Research
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by McQueen, A.
Right arrow Articles by Tilley, B. C.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by McQueen, A.
Right arrow Articles by Tilley, B. C.
Related Collections
Right arrow Behavioral Science
Right arrow Behavioral Science: Interventions: Behavioral, Pharmacologic, Pharmacogenetic

Correlates and Predictors of Colorectal Cancer Screening among Male Automotive Workers

Amy McQueen1, Sally W. Vernon1, Ronald E. Myers4, Beatty G. Watts3, Eun Sul Lee2 and Barbara C. Tilley5

1 Center for Health Promotion and Prevention Research and 2 Department of Biometry, University of Texas School of Public Health; 3 Department of Behavioral Science, The University of Texas M. D. Anderson Cancer Center, Houston, Texas; 4 Division of Medical Oncology, Department of Medicine, Thomas Jefferson University, Philadelphia, Pennsylvania; and 5 Department of Biometery and Epidemiology, Medical University of South Carolina, Charleston, South Carolina

Requests for reprints: Amy McQueen, Center for Health Promotion and Prevention Research, University of Texas School of Public Health, 7000 Fannin, Suite 2568, Houston, TX 77030. Phone: 713-500-9782; Fax: 713-500-9750. E-mail: Amy.McQueen{at}uth.tmc.edu


   Abstract
Top
 Abstract
Introduction
 Materials and Methods
 Results
 Discussion
 References
 
Background: Most studies examining factors associated with colorectal cancer (CRC) screening (CRCS) are cross-sectional and thus temporal relationships cannot be determined. Furthermore, less attention has been paid to psychosocial predictors of CRCS. We examined both cross-sectional correlates of prior CRCS and predictors of prospective CRCS initiation and maintenance during The Next Step Trial, a 2-year worksite behavioral intervention to promote regular CRCS and dietary change.

Method: The sample included 2,693 White male automotive workers at increased occupational risk for, but no history of, CRC who completed a baseline survey. Stratified analyses were conducted for three dependent variables (prior CRCS, CRCS initiation, and CRCS maintenance). We also assessed prior CRCS as a moderator in prospective analyses. Multivariable logistic regression analyses with generalized linear mixed models were used to adjust for cluster sampling.

Results: Except for education, cross-sectional correlates of prior CRCS including older age, family history of CRC or polyps, personal history of polyps, self-efficacy, family support, and intention were also significant prospective predictors of increased CRCS during the trial. Despite differences in the patterns of association for CRCS initiation and maintenance in stratified analyses, the only associations with prospective CRCS that were significantly moderated by prior CRCS were family history and CRCS availability.

Conclusions: Correlates of prior CRCS that also were prospective predictors of CRCS may be suitable targets for intervention. Additionally, intervention messages addressing psychosocial constructs may be relevant for both CRCS initiation and maintenance. However, studies with more diverse samples are needed to replicate the results reported here. (Cancer Epidemiol Biomarkers Prev 2007;16(3):500–9)


   Introduction
Top
 Abstract
 Introduction
Materials and Methods
 Results
 Discussion
 References
 
Colorectal cancer (CRC) is the second leading cause of cancer deaths in the United States (1). The early detection and removal of precancerous polyps may decrease incidence of CRC (2). Therefore, several authoritative groups recommend regular CRC screening (CRCS) beginning at age 50 years for average-risk adults and earlier for people at higher risk for CRC (1, 3). However, screening rates remain <50% (4, 5).

Previous studies have identified numerous cross-sectional correlates, but few studies have examined prospective predictors of CRCS. Although demographic variables have been associated with CRCS (6-8) and may be useful for identifying population subgroups for intervention, they are less amenable to change and are not as useful in interventions designed to motivate screening use. Only a few studies have systematically used theories or models of behavior change to predict CRCS (9). Psychosocial constructs positively associated with CRCS include a preventive health orientation, perceived benefits (pros) and fewer barriers (cons) to screening, self-efficacy, and intention (10). Very few psychosocial constructs have been examined in prospective studies, and fewer have been examined across studies (10). Even when the same constructs were used, they were often defined and operationalized differently, and most measures of psychosocial constructs have not been validated. Therefore, very little is known about the importance of these factors for CRCS.

Predictors may differ depending on whether the outcome of interest is CRCS initiation or maintenance; however, few prospective studies have examined CRCS stratified by past screening behavior (11). Identifying the similarities and differences in factors associated with initiation and maintenance will provide a stronger foundation for developing effective interventions. We examined correlates of prior CRCS (cross-sectional analysis) and predictors of CRCS initiation and maintenance (stratified prospective analyses). Additionally, we tested whether prior CRCS moderated the association between predictors and any future CRCS use.

This report presents a secondary analysis of data from The Next Step Trial, a 2-year behavioral intervention trial to promote regular CRCS and dietary change in automotive workers (12). The Preventive Health Model (PHM; ref. 13) was the conceptual framework used in this study. The PHM has been used to study intention and behavior for colorectal (13-15) and prostate cancer screening (16, 17). We used this model to select variables that have previously been associated with cancer screening to answer our first research question: Are cross-sectional correlates of CRCS also significant predictors of future CRCS? Because of the scant attention to psychosocial variables in the literature, we were specifically interested in their association with CRCS after accounting for background predictors. Previous prospective associations have been found between CRCS and perceived benefits and barriers, self-efficacy, and intention; therefore, we expected that they would be significant predictors of CRCS (11, 13, 18). Few theories or models distinguish between screening initiation and maintenance or specify whether determinants differ (19, 20). Thus, our second research question was: Do the predictors of CRCS initiation and CRCS maintenance significantly differ? Because there was little basis for hypothesis testing, exploratory analyses examined whether these variables were similarly associated with prior CRCS, CRCS initiation, and CRCS maintenance. Additionally, the predictive ability (e.g., sensitivity and specificity) of the regression models predicting screening was examined.


   Materials and Methods
Top
 Abstract
 Introduction
 Materials and Methods
Results
 Discussion
 References
 
The Next Step Trial
Epidemiologic studies found that automotive pattern- and model-makers were at increased risk for CRC incidence and mortality (21, 22). As a result, employees were offered CRCS through a company-sponsored program beginning in 1980; however, by 1989, a large majority of eligible employees were not being screened regularly (23). Low screening rates prompted the development and implementation of The Next Step Trial, a National Cancer Institute–funded screening and nutrition intervention (12, 23, 24). All 28 worksites developed their own process for offering screening; 15 worksites were randomized to also receive an educational intervention. Employees at intervention worksites received a mailed invitation to the screening program, an educational booklet tailored to the employee's screening history, and a telephone call to reinforce messages from the booklet (23). Screening recommendations were based on American Cancer Society guidelines in effect at that time. Depending on their screening history and on previous examination findings, employees at all worksites were offered three types of CRCS [i.e., digital rectal examinations, fecal occult blood tests (FOBT), and/or sigmoidoscopy] or referral to their physician for colonoscopy or barium enema.

Study Population
There were 5,042 current and former automotive workers who were eligible for the trial (i.e., worked in pattern- and model-making for at least 2 years at a minimum of 20% time). Annual surveys were sent to eligible workers in 1993, 1994, and 1995. Nonrespondents received a reminder postcard after 2 weeks and a telephone call after 4 weeks. If a nonrespondent did not receive a survey or did not remember receiving it, another was mailed. In 1993, 2,903 (58%) workers responded to the baseline questionnaire. Survey responders were more likely to be older, married, retired, nonsmokers, have more formal education, have a personal history of colorectal polyps, and have had CRCS before and during the trial (23). After excluding women (n = 53), non-White males (n = 114), and men with a history of CRC (n = 43), 2,693 non-Hispanic White males were retained for analysis.

Measures
Dependent Variables
Each year, data on screening tests completed were provided by worksite staff and self-report surveys (12, 23). Eligible employees completing at least one screening test during the 2-year "pretrial" period were considered covered for that period (12). Likewise, eligible employees completing at least one test during the 2-year trial were considered covered for that period. Employees not recommended for screening because they were not due also were considered covered.

Three dependent variables assessing screening coverage were examined. In the cross-sectional analysis, the dependent variable was "screened prior" to the trial; men who were covered in the 2-year period before the trial were compared with men who were not. In the prospective analyses, men who met the definition for coverage during the 2-year trial were classified into two groups (i.e., CRCS initiation and maintenance) and compared with men who were not covered during the trial. "CRCS initiation" included men who were not covered during the pretrial period, but who were covered during the trial. Initiation during the trial included men who had never been screened previously, as well as men overdue for screening. It was not possible to distinguish between men who had never been screened and men who were overdue because CRCS histories were not available beyond the 2 years before the trial. "CRCS maintenance" included men who were covered during the 2-year pretrial period and during the trial.

Independent Variables
The PHM consists of background, cognitive and psychological, social influence, intention, and program variables. PHM variables were abstracted from employee records or the baseline survey. Program variables were collected at baseline by surveying plant medical staff at each worksite to ascertain characteristics related to delivery of the screening program (23).

PHM background variables included demographic characteristics (age, education, marital status, and retirement status), risk factors (family history of CRC or polyps), and medical history data (smoking status, personal history of colorectal polyps).

PHM cognitive and psychological, social influence, and intention variables were measured by scales and single items using a four-point Likert format (25). Scales measured belief in the salience and coherence of CRCS (four items), perceived CRCS self-efficacy (four items), perceived susceptibility to colorectal polyps or cancer (three items), worries or fears about being diagnosed with CRC (two items), and intention to be screened for CRC (two items). Single items were used to measure belief in the effectiveness of CRCS in detecting CRC, belief that polyp removal can prevent CRC, belief that CRC can be cured, concern about CRCS-related discomfort, and support for CRCS from family members. Responses were skewed and some response categories were sparse; therefore, all items and mean scale scores were dichotomized at <3 (disagree or strongly disagree) or ≥3 (strongly agree or agree).

PHM program variables indicated whether the worksite was assigned to an intervention or control study condition; CRCS was administered onsite (yes, no) and for what period of time (all year, not all year long); how FOBT kits were distributed (mail only, mail and/or at appointment, employee pickup, other); and the number of options for scheduling CRCS appointments (1, ≥2). Scheduling options varied by worksite and included employees who called to make an appointment, staff who called employees to make an appointment, and education/information sessions that included CRCS scheduling.

Data Analysis
Univariate analyses were conducted using {chi}2 tests to examine associations between baseline independent variables and prior CRCS, CRCS initiation, and CRCS maintenance. To test the statistical significance of predictors found to be differentially associated with CRCS initiation and maintenance in the stratified prospective analyses, we assessed whether prior CRCS coverage modified any of the associations between those predictors and CRCS coverage during the trial. Following Hosmer and Lemeshow (26), P < 0.10 was selected for model building, which decreases the likelihood of making type II errors. However, in the multivariable logistic regression models, P < 0.05 was used to minimize type I errors. To maintain comparability, independent variables were retained in all multivariable models unless the univariate associations were P > 0.10 for all models. Study group also was retained in all models to account for any intervention effects. Generalized linear mixed models were conducted with Proc Glimmix in SAS 9.1 to adjust for the effects of cluster sampling and account for correlations within a worksite. We report odds ratios and 95% confidence intervals and we used least squares means to interpret significant interactions.

For all analyses, participants with any missing data were excluded. The percentage of missing responses for independent variables ranged from 0 for employment status (retired versus active) and program variables to 8.2% for family support. {chi}2 tests with Bonferroni correction for multiple tests indicated that respondents with complete data were more likely (P < 0.001) to be younger, married, have more formal education, have a family history of CRC or polyps, and had had CRCS in the 2 years before the trial. However, all effect sizes were very small (ESr < 0.10; ref. 27).

Only the prospective models allowed direct assessment of the abilities of our models to predict CRCS initiation and maintenance. We compared the classifications based on our prospective models to the observed behavior for each study participant. To make these comparisons, we calculated sensitivity, specificity, and positive and negative predictive values. Sensitivity measures the proportion of participants classified as being screened (i.e., CRCS initiators or maintainers) by the set of variables in the regression model who were observed to be screened during the trial. Specificity measures the proportion of participants classified as not being screened during the trial who, in fact, were not screened. Positive predictive value indicates the proportion of participants who were actually screened during the trial among the participants who were predicted to be screened by the regression model. Negative predictive value refers to the proportion of participants who actually did not get screened among the participants who were predicted not to be screened. All model variables were included in the classification plots with a cutoff of 0.5. Cases with predicted values that exceed the classification cutoff are classified as positive (screened), whereas those with predicted values smaller than the cutoff are classified as negative (not screened). These calculations are an important step in model validation because association is not sufficient evidence of prediction. Models with high sensitivity, specificity, and predictive values indicate that there is little misclassification.


   Results
Top
 Abstract
 Introduction
 Materials and Methods
 Results
Discussion
 References
 
Preliminary Analyses
Table 1 describes the baseline characteristics of the sample (column 1) and the distribution of CRCS for the three dependent variables by PHM variables (columns 2-4). Overall, 76.5% of baseline participants had CRCS in the 2 years before the trial, and 74.7% were screened during the trial. Of those not screened in the 2 years before the trial, 44.2% had CRCS during the trial, whereas of those screened before the trial, 84.1% maintained screening coverage during the trial.


View this table:
[in this window]
[in a new window]

  		Table 1. Frequency distribution of respondents to The Next Step Trial baseline survey and the percent screened based on univariate cross-sectional and prospective analyses

 
Univariate results, stratified by dependent variable, showed that all but four variables (employment status, smoking, fear and worry about screening, and CRCS appointment scheduling) were associated (p ≤ 0.10) with at least one dependent variable (Table 1, columns 2-4) and were included in all three multivariable analyses.

Stratified Multivariable Analyses
For CRCS before the trial, most background variables and three psychosocial variables were significant correlates. Older age, more years of education, a family history of CRC or polyps, a personal history of polyps, high perceived self-efficacy, support for CRCS from family members, and strong intention to be screened were positively associated with prior CRCS.

For CRCS initiation during the trial, few variables remained statistically significant in multivariable analysis (Table 2 ). Being married, having a family history of CRC or polyps, having any personal history of polyps, and reporting a strong intention to be screened were positively associated with CRCS initiation.


View this table:
[in this window]
[in a new window]

  		Table 2. Stratified and moderated prospective associations with CRCS during the trial

 
CRCS maintenance during the trial was positively associated with older age, having a personal history of polyps, believing in the effectiveness of screening for detecting CRC, high self-efficacy, greater family support for CRCS, a strong intention to be screened, exposure to the intervention, and having CRCS offered onsite throughout the year (Table 2). Univariate analyses suggested that the intervention was not significantly associated with CRCS; however, in the presence of other program variables, working at an intervention worksite was positively associated with CRCS maintenance.

Moderated Analyses
Univariate analyses identified several factors that significantly interacted with prior CRCS when predicting any CRCS coverage during the trial (i.e., initiation or maintenance): age, marital status, family history of CRC or polyps, belief that CRC can be cured, CRCS availability, and FOBT distribution method. However, in the multivariable analysis predicting CRCS coverage during the trial that added all main effects, only the effects of family history and CRCS availability were significantly moderated by prior CRCS (Table 2). Men with a family history of CRC or polyps were significantly more likely to initiate screening during the trial compared with men without a family history (Fig. 1 ). Having CRCS available onsite throughout the year was associated with increased CRCS maintenance, but CRCS availability had no effect on CRCS initiation (Fig. 2 ). Significant main effects that were maintained in the full model with interactions included age, family history, personal polyp history, self-efficacy, family support, intention, study condition, prior CRCS, CRCS availability, and FOBT distribution (Table 2). Being in a worksite that distributed FOBT kits by other methods (e.g., at an education/information session) was associated with increased CRCS during the trial.


Figure 1
View larger version (8K):
[in this window]
[in a new window]

  		Figure 1. Least squares means for the association between family history of colorectal cancer or polyps and screening coverage during the trial significantly moderated by screening before The Next Step Trial.

 

Figure 2
View larger version (10K):
[in this window]
[in a new window]

  		Figure 2. Least squares means for the association between colorectal cancer screening availability and screening coverage during the trial significantly moderated by screening before The Next Step Trial.

 
Predictive Validity Analyses
The predictive ability of the models also differed for CRCS initiation and maintenance (Table 3 ). The model for CRCS initiation had a moderate level of sensitivity and specificity, suggesting that it was similarly able to identify men who were never screened and men who initiated CRCS during the trial. The model for maintenance displayed a high level of sensitivity when predicting maintenance, but it was not successful in identifying men who had been screened before the trial but were not screened during the trial (specificity, 6.4%). In other words, the model overestimated the number of men screened before the study that would maintain their screening during the study. In the full model that included the significant interactions between age and family history with prior screening predicting CRCS coverage, specificity (38%) improved, but the model still overestimated the number of men that would be screened. The positive predictive values suggest that the classification of participants is more accurate when predicting maintenance than initiation. However, the negative predictive values suggest that the classification of participants as nonscreeners during the trial was accurate ~70% of the time.


View this table:
[in this window]
[in a new window]

  		Table 3. Predictive validity measures for CRCS initiation and maintenance with 95% confidence intervals

 

   Discussion
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
References
 
Relying on cross-sectional results when designing interventions to change behavior may overlook important factors influencing cancer screening decisions. This is the first study that we are aware of that compared cross-sectional correlates and prospective predictors of CRCS. In another study, Bastani et al. (28) examined correlates versus predictors of mammography screening and found that, although the results of cross-sectional and prospective analyses were very similar, a larger number of variables were related to future behavior compared with past behavior. They also found that the additional factors were more often attitudinal than demographic. Several variables examined in this study were directly associated with CRCS in both cross-sectional and prospective analyses, which may suggest that for this study population, these factors are stable targets for intervention. We found consistent associations for both background (age, family history, personal polyp history) and psychosocial variables (self-efficacy, family influence, intention). Although demographic variables may identify subgroups in greater need for intervention, cognitive and psychosocial predictors of screening behaviors are important targets of interventions because they are more amenable to change.

Significant associations in this study are generally consistent with observations from cross-sectional studies of prior CRCS, and they extend the literature on predictors of future screening behavior (Table 4 ). Cumulative evidence from our findings and other reports in the literature involving prospective studies and multivariable analyses suggests consistent associations between CRCS and perceived benefits and barriers, self-efficacy, physician recommendation, and intention (Table 4). Future interventions should address the variables that are consistently associated with CRCS; however, additional psychosocial variables that have not been as thoroughly examined in previous research also may be important targets for intervention. Specifically, variables measuring affect and social influence have seldom been studied in relation to CRCS (Table 4). The positive associations we observed between family members' support and CRCS suggest that involving family members may be a strategy for increasing both CRCS initiation and maintenance. The inconsistent findings about the effect of knowledge on CRCS (Table 4) may be due to low variance in knowledge measures or to a bidirectional association with screening behavior. Additionally, the inconsistent findings in the literature about perceived susceptibility to CRC on CRCS may be due to the different measures used (29) and the likelihood that the effect of perceived susceptibility on behavior is mediated by other psychosocial variables (30).6 Our conceptual framework, the PHM, posits only direct effects between predictors and intention and behavior, which may underestimate the total effects of CRCS determinants through other mediating variables (30). Future studies that explore more complex causal models of CRCS are needed to extend previous research, which has largely focused on demographic correlates of CRCS in direct-effects-only models.


View this table:
[in this window]
[in a new window]

  		Table 4. Psychosocial correlates and predictors associated with FOBT, sigmoidoscopy, and colonoscopy use

 
In addition to psychosocial variables, this study also included program variables that have not previously been examined in CRCS studies. Univariate analyses suggested that the intervention was not significantly associated with CRCS; however, in the presence of other program variables, the intervention was associated with screening during the trial. The intervention may have facilitated the use of worksite screening programs, as well as influenced participants' reactions to the characteristics of specific worksite programs and procedures. Although the intervention literature shows greater screening among those mailed FOBT kits compared with those not provided test kits (9, 10), being in a worksite that distributed FOBT kits by other methods (e.g., at an education/information session) was associated with higher screening during the trial. The effects of specific program characteristics on CRCS warrant further investigation because such factors may be important for understanding how to increase CRCS in different settings (i.e., worksites and healthcare practices). Future research is needed to identify the complex mediating and moderating effects of interventions and program variables on CRCS over time.

Despite the differences in stratified analyses, most predictors were relevant for both CRCS initiation and maintenance. Prior screening only moderated the effect of family history and CRCS availability, suggesting that interventions addressing psychosocial variables may be equally effective for motivating individuals who have never been screened and those who are overdue for regular screening. However, future research should replicate these findings with more diverse samples. Because most health behavior theories do not identify specific determinants or causal mechanisms that distinguish between behavior initiation and maintenance, more research is needed to provide further support for the usefulness of conceptual models like the PHM for both initiation and maintenance of cancer screening behaviors.

Our observed interactions between prior CRCS and family history and CRCS availability may inform future interventions designed to increase CRCS initiation and/or maintenance. Family history of CRC has been consistently associated with CRCS in prior research (7, 31, 32); however, results in this report show that reporting a family history of CRC or polyps influenced men to initiate, but not maintain, CRCS. Screening may allay concerns about increased hereditary risk factors and no longer motivate CRCS maintenance, whereas a personal history of polyps reinforces the need for repeated CRCS. The availability of CRCS varied across worksite screening programs, and our results suggest that offering CRCS onsite throughout the year increased CRCS maintenance but had no effect on CRCS initiation. Although more convenient for men overdue for repeat CRCS, having CRCS available throughout the year may decrease any sense of urgency and allow nonscreeners to procrastinate about initiating the behavior. More research is needed to examine predictors of CRCS and their interaction with prior CRCS to better understand and influence screening behavior over time.

Predictive Validity
The predictive validity analyses may inform decisions about what variables to target in intervention studies by indicating the percentage of study participants the model correctly identifies. Based on these analyses, there may be greater confidence in the predictors of maintenance. However, among men who were screened before the trial, the model was unable to identify those who were not screened during the trial ("true negatives") and misclassified many participants as repeat screeners ("false positives"). Only 15% of men who were screened before the trial were not screened during the trial, which may make it difficult to identify true negatives. These results suggest that other, unmeasured variables are important for predicting relapse among prior screeners. Likewise, future work is needed to identify other important factors that affect CRCS initiation.

Limitations and Conclusions
Limitations of The Next Step Trial have been discussed elsewhere (12, 23, 24). The survey response rate of 58% was similar to other worksite health promotion studies that used similar methods (33-35); however, basing inferences only on data from survey respondents may introduce bias. Survey nonresponse increased the CRCS prevalence estimates reported here because screening data were available for 75% of survey respondents but only 43% of nonrespondents. In a previous report of the intervention trial results, an intention-to-treat analysis was done that included participants without screening data who were assumed to be nonscreeners (12). Because survey respondents were more likely to be older, married, and to have been screened in the 2 years before the trial, all factors positively associated with CRCS during the trial, the observed associations between CRCS, and the predictor variables may have been attenuated.

Item nonresponse reduced the analysis sample for some comparisons and may have decreased the power to detect some significant associations. Although the sample was composed of White, male automotive workers at increased risk for CRC, the patterns of association observed here were generally consistent with those reported in the literature (Table 4). The data were collected before 1997 when guidelines for CRCS were published (36). However, the study population was informed that they were at increased risk for CRC and were offered free screening through their employer. Thus, their awareness of and belief in the efficacy of CRCS may be similar to the general population after 1997.

Because CRCS histories were not available beyond the 2 years before the trial, we could not determine the number of men who initiated screening during the trial who had never been screened versus those who were overdue for screening, which may have caused some misclassification. Additional studies are needed to examine whether predictors differ across measures of first-time screening, overdue screening, and repeated, on-schedule screening (maintenance). Additionally, the effect of prior screening experience on future screening behavior is likely to be influenced by a number of factors including physician recommendation and preferences for test characteristics such as invasiveness, cost, and intervals for repeat screening.

Strengths of the study include a prospective design with multiple data collection periods, a well-defined study population, a large sample size, a validated survey instrument (25, 37), and the comparison of factors associated with prior CRCS, CRCS initiation, and CRCS maintenance. In this study, if only one definition of CRCS had been examined, different conclusions would have been drawn. More research is needed to investigate similarities or differences in operational definitions of CRCS to better inform intervention efforts. Our findings and others in the literature suggest that perceived benefits and barriers, family influence, self-efficacy, and intention are significant cross-sectional correlates as well as prospective predictors of CRCS. Only two factors significantly interacted with prior screening to suggest that different strategies may be needed to motivate CRCS initiation and maintenance for this study population. However, additional prospective research with more diverse samples is needed to confirm the results reported here and to better understand the mediated and moderated pathways linking background, psychosocial, and program factors with CRCS.


   Footnotes
 
Grant support: National Cancer Institute grant 5R01CA052605-04 Colorectal Cancer Screening and Nutrition Intervention; National Cancer Institute training grant R25CA57712-11 (A. McQueen); National Cancer Institute grants R01CA076330 Women Veterans and Breast Cancer Screening, R01CA097263 Tailored Interactive Intervention to Increase CRC Screening, and R03CA103512 A Cancer Study Among Female Veterans in Texas, 1979-2001 (S.W. Vernon); National Cancer Institute grants 5R01CA084140-04 Increasing Colon Cancer Screening in Primary Care, 5R21CA102418-02 Tailored Messaging in Colorectal Cancer Screening, and 5U01CA086084-05 Increasing Access to Clinical and Educational Studies (R.E. Myers); and Department of Defense grant SP0001 Occupational and Nutritional Risk Factors for Colon Cancer and Adenomatous Polyps in the Tri-Services (B.C. Tilley).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

6 A. McQueen, S.W. Vernon, R.E. Myers, B.C. Tilley. Examining mediators of perceived susceptibility of colorectal cancer on screening intention and behavior. Paper presented at the 27th annual meeting for Society of Behavioral Medicine, San Francisco, CA. Submitted for publication, 2006. Back

Received 9/ 7/06; revised 12/ 6/06; accepted 12/15/06.


   References
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
 References
 

  1. American Cancer Society. Cancer facts and figures, 2006. Atlanta (GA): American Cancer Society; 2006.
  2. Winawer SJ, Zauber AG, Ho MN, et al.; The National Polyp Study Workgroup. Prevention of colorectal cancer by colonoscopic polypectomy. N Engl J Med 1993;329:1977–81.[Abstract/Free Full Text]
  3. U S. Preventive Services Task Force Screening for colorectal cancer: recommendations and rationale. Ann Intern Med 2002;137:129–31.[Abstract/Free Full Text]
  4. Meissner HI, Breen NL, Klabunde CN, Vernon SW. Patterns of colorectal cancer screening uptake among men and women in the US. Cancer Epidemiol Biomarkers Prev 2006;15:389–94.[Abstract/Free Full Text]
  5. Swan J, Breen NL, Coates RJ, Rimer BK, Lee NC. Progress in cancer screening practices in the United States: results from the 2000 National Health Interview Survey. Cancer 2003;97:1528–40.[CrossRef][Medline]
  6. Cokkinides VE, Chao A, Smith RA, Vernon SW, Thun MJ. Correlates of underutilization of colorectal cancer screening among U.S. adults, age 50 years and older. Prev Med 2003;36:85–91.[CrossRef][Medline]
  7. Seeff LC, Nadel MR, Klabunde CN, et al. Patterns and predictors of colorectal cancer test use in the adult US population: results from the 2000 National Health Interview Survey. Cancer 2004;100:2093–103.[CrossRef][Medline]
  8. Shapiro JA, Seeff LC, Nadel MR. Colorectal cancer-screening tests and associated health behaviors. Am J Prev Med 2001;21:132–7.[CrossRef][Medline]
  9. Peterson SK, Vernon SW. A review of patient and physician adherence to colorectal cancer screening guidelines. Semin Colon Rectal Surg 2000;11:58–72.
  10. Vernon SW. Participation in colorectal cancer screening: a review. J Natl Cancer Inst 1997;89:1406–22.[Abstract/Free Full Text]
  11. Myers RE, Trock BJ, Lerman CL, Wolf TA, Ross EA, Engstrom PF. Adherence to colorectal cancer screening in an HMO population. Prev Med 1990;19:502–14.[CrossRef][Medline]
  12. Tilley BC, Vernon SW, Myers RE, et al. The Next Step Trial: impact of a worksite colorectal cancer screening promotion program. Prev Med 1999;28:276–83.[CrossRef][Medline]
  13. Myers RE, Ross EA, Jepson C, et al. Modeling adherence to colorectal cancer screening. Prev Med 1994;23:142–51.[CrossRef][Medline]
  14. Myers RE, Vernon SW, Tilley BC, Lu M, Watts BG. Intention to screen for colorectal cancer among white male employees. Prev Med 1998;27:279–87.[CrossRef][Medline]
  15. Watts BG, Vernon SW, Myers RE, Tilley BC. Intention to be screened over time for colorectal cancer in male automotive workers. Cancer Epidemiol Biomarkers Prev 2003;12:339–49.[Abstract/Free Full Text]
  16. Myers RE, Wolf TA, McKee L, et al. Factors associated with intention to undergo annual prostate cancer screening among African American men in Philadelphia. Cancer 1996;78:471–9.[CrossRef][Medline]
  17. Myers RE, Chodak GW, Wolf TA, et al. Adherence by African American men to prostate cancer education and early detection. Cancer 1999;86:88–104.[CrossRef][Medline]
  18. Sutton SR, Wardle J, Taylor T, et al. Predictors of attendance in the United Kingdom flexible sigmoidoscopy screening trial. J Med Screen 2000;7:99–104.[Abstract/Free Full Text]
  19. Gritz ER, Bastani R. Cancer prevention—behavior changes: the short and long of it. Prev Med 1993;22:676–88.[CrossRef][Medline]
  20. Rothman AJ. Toward a theory-based analysis of behavioral medicine. Health Psychol 2000;19:64–9.[CrossRef][Medline]
  21. Robinson CR, Waxweiler RJ, McCammon CS. Pattern and model makers, proportional mortality 1972-1978. Am J Ind Med 1980;1:159–65.[Medline]
  22. Swanson FM, Belle SH, Burrows RW. Colon cancer incidence among modelmakers and patternmakers in the automobile manufacturing industry. J Occup Med 1985;27:567–9.[CrossRef][Medline]
  23. Tilley BC, Vernon SW, Glanz K, et al. Worksite cancer screening and nutrition intervention for high-risk auto workers: design and baseline findings of the Next Step Trial. Prev Med 1997;26:227–35.[CrossRef][Medline]
  24. Tilley BC, Glanz K, Kristal AR, et al. Nutrition intervention for high-risk auto workers: results of the Next Step Trial. Prev Med 1999;28:284–92.[CrossRef][Medline]
  25. Vernon SW, Myers RE, Tilley BC. Development and validation of an instrument to measure factors related to colorectal cancer screening adherence. Cancer Epidemiol Biomarkers Prev 1997;6:825–32.[Abstract]
  26. Hosmer DW, Lemeshow S. Applied logistic regression. New York: John Wiley & Sons, Inc.; 2000.
  27. Cohen J. A power primer. Psychological Bulletin 1992;112:155–9.
  28. Bastani R, Maxwell AE, Bradford C. Cross-sectional versus prospective predictors of screening mammography. Cancer Epidemiol Biomarkers Prev 1996;5:845–8.[Abstract]
  29. Vernon SW. Risk perception and risk communication for cancer screening behaviors: a review. J Natl Cancer Inst Monogr 1999;(25):101–19.[Abstract/Free Full Text]
  30. Aiken LS, West SG, Woodward CK, Reno RR, Reynolds KD. Increasing screening mammography in asymptomatic women: evaluation of a second-generation, theory-based program. Health Psychol 1994;13:526–38.[CrossRef][Medline]
  31. Codori A-M, Petersen GM, Miglioretti DL, Boyd PA. Health beliefs and endoscopic screening for colorectal cancer: potential for cancer prevention. Prev Med 2001;33:128–36.[Medline]
  32. Lemon SC, Zapka JG, Puleo E, Luckmann R, Chasan-Taber L. Colorectal cancer screening participation: comparisons with mammography and prostate-specific antigen screening. Am J Public Health 2001;91:1264–72.[Abstract/Free Full Text]
  33. Abrams DB, Boutwell WB, Grizzle J, Heimendinger J, Sorensen G, Varnes J. Cancer control at the workplace: the Working Well Trial. Prev Med 1994;23:15–27.[CrossRef][Medline]
  34. Heimendinger J, Feng Z, Emmons KM, et al.; The Working Well research group. The Working Well Trial: baseline dietary and smoking behaviors of employees and related worksite characteristics. Prev Med 1995;24:180–93.[CrossRef][Medline]
  35. Levy MC. Correlates of cancer-preventive nutrition practices in a worksite population. Los Angeles (CA): University of California Los Angeles; 1994.
  36. Winawer SJ, Fletcher RH, Miller L, et al. Colorectal cancer screening: clinical guidelines and rationale. Gastroenterology 1997;112:594–642.[CrossRef][Medline]
  37. Tiro JA, Vernon SW, Hyslop T, Myers RE. Factorial validity and invariance of a survey measuring psychosocial correlates of colorectal cancer screening among African Americans and Caucasians. Cancer Epidemiol Biomarkers Prev 2005;14:2855–61.[Abstract/Free Full Text]
  38. Farrands PA, Hardcastle JD, Chamberlain JO, Moss SM. Factors affecting compliance with screening for colorectal cancer. Community Med 1984;6:12–9.[Medline]
  39. Ioannou GN, Chapko MK, Dominitz JA. Predictors of colorectal cancer screening participation in the United States. Am J Gastroenterol 2003;98:2082–91.[CrossRef][Medline]
  40. Chao A, Connell CJ, Cokkinides V, Jacobs EJ, Calle EE, Thun MJ. Underuse of screening sigmoidoscopy and colonoscopy in a large cohort of US adults. Am J Public Health 2004;94:1775–81.[Abstract/Free Full Text]
  41. Zapka JG, Puleo E, Vickers-Lahti M, Luckmann R. Healthcare system factors and colorectal cancer screening. Am J Prev Med 2002;23:28–35.[CrossRef][Medline]
  42. Brawarsky P, Brooks DR, Mucci LA. Correlates of colorectal cancer testing in Massachusetts men and women. Prev Med 2003;36:659–68.[CrossRef][Medline]
  43. Walsh JME, Posner SF, Perez-Stable EJ. Colon cancer screening in the ambulatory setting. Prev Med 2002;35:209–18.[CrossRef][Medline]
  44. Straus WL, Mansley EC, Gold KF, Wang Q, Reddy P, Pashos C. Colorectal cancer screening attitudes and practices in the general population: a risk-adjusted survey. J Public Health Manag Pract 2005;11:244–51.[Medline]
  45. Etzioni DR, Ponce NA, Babey SH, et al. A population-based study of colorectal cancer test use: results from the 2001 California health interview survey. Cancer 2004;101:2523–32.[CrossRef][Medline]
  46. Gorin SS, Heck JE. Cancer screening among Latino subgroups in the United States. Prev Med 2005;40:515–26.[Medline]
  47. Brawarsky P, Brooks DR, Mucci LA, Wood PA. Effect of physician recommendation and patient adherence on rates of colorectal cancer testing. Cancer Detect Prev 2004;28:260–8.[Medline]
  48. MacRae FA, Hill DJ, St. John DJB, Ambikapathy A, Garner JF. Predicting colon cancer screening behavior from health beliefs. Prev Med 1984;13:115–26.[CrossRef][Medline]
  49. Neilson AR, Whynes DK. Determinants of persistent compliance with screening for colorectal cancer. Soc Sci Med 1995;41:365–74.[CrossRef][Medline]
  50. Janz NK, Wren PA, Schottenfeld D, Guire KE. Colorectal cancer screening attitudes and behavior: a population-based study. Prev Med 2003;37:627–34.[CrossRef][Medline]
  51. Juon H, Han W, Shin H, Kim KB, Kim MT. Predictors of older Korean Americans' participation in colorectal cancer screening. J Cancer Educ 2003;18:37–42.[Medline]
  52. Brown ML, Potosky AL, Thompson GB, Kessler LG. The knowledge and use of screening tests for colorectal and prostate cancer: data from the 1987 National Health Interview Survey. Prev Med 1990;19:562–74.[CrossRef][Medline]
  53. Greiner KA, James AS, Born WK, et al. Colorectal cancer screening preferences and predictors of Fecal Occult Blood Test (FOBT) return among low income adults. Prev Med 2005;41:676–84.[Medline]
  54. Paskett ED, D'Agostino RB, Jr., Tatum CM, Velez R, Brenes GA. Colorectal cancer screening practices among low-income women. Clin J Women's Health 2000;1:3–9.
  55. Menon U, Champion VL, Larkin GN, Zollinger TW, Gerde PM, Vernon SW. Beliefs associated with fecal occult blood test and colonoscopy use at a work-site colon cancer screening program. J Occup Environ Med 2003;45:891.[Medline]
  56. Weinberg DS, Turner BJ, Wang H, Myers RE, Miller SM. A survey of women regarding factors affecting colorectal cancer screening compliance. Prev Med 2004;38:669–75.[Medline]
  57. Price JH. Perceptions of colorectal cancer in a socioeconomically disadvantaged population. J Community Health 1993;18:347–62.[CrossRef][Medline]
  58. Gorin SS. Correlates of colorectal cancer screening compliance among urban Hispanics. J Behav Med 2005;28:125–37.[CrossRef][Medline]
  59. Brenes GA, Paskett ED. Predictors of stage of adoption for colorectal cancer screening. Prev Med 2000;31:410–6.[CrossRef][Medline]
  60. Paskett ED, Rushing J, D'Agostino RB, Jr., Tatum CM. Cancer screening behaviors of low-income women: the impact of race. Women Health 1997;3:203–26.
  61. Sun WY, Basch CE, Wolf RL, Li XJ. Factors associated with colorectal cancer screening among Chinese-Americans. Prev Med 2004;39:323–9.[Medline]
  62. Rawl SR, Champion VL, Menon U, Skinner CS. Validation of scales to measure benefits and barriers to colorectal cancer screening: scale development. J Psychosocial Oncol 2001;19:47–63.
  63. Lewis SF, Jensen NM. Screening sigmoidoscopy: factors associated with utilization. J Gen Intern Med 1996;11:542–4.[Medline]
  64. Madlensky L, Esplen MJ, Gallinger S, McLaughlin JR, Goel V. Relatives of colorectal cancer patients. Am J Prev Med 2003;25:187–94.[CrossRef][Medline]
  65. James AS, Campbell MK, Hudson MA. Perceived barriers and benefits to colon cancer screening among African Americans in North Carolina: how does perception relate to screening behavior? Cancer Epidemiol Biomarkers Prev 2002;11:529–34.[Abstract/Free Full Text]
  66. Manne SL, Markowitz A, Winawer SJ, et al. Correlates of colorectal cancer screening compliance and stage of adoption among siblings of individuals with early onset colorectal cancer. Health Psychol 2002;21:3–15.[CrossRef][Medline]
  67. Stockwell DH, Woo P, Jacobsen BC, et al. Determinants of colorectal cancer screening in women undergoing mammography. Am J Gastroenterol 2003;98:1875–80.[Medline]
  68. Hoogewerf PE, Hislop TG, Morrison BJ, Burns SD, Sizto R. Health belief and compliance with screening for fecal occult blood. Soc Sci Med 1990;30:721–6.[CrossRef][Medline]
  69. Halper MS, Winawer SJ, Brody RS, Andrews M, Roth D, Burton G. Issues of patient compliance. In: Winawer SJ, Schottenfeld D, editors. Colorectal cancer: prevention, epidemiology, and screening. New York (NY): Raven Press; 1980. p. 299–310.
  70. Weller DP, Owen N, Hiller JE, Willson K, Wilson D. Colorectal cancer and its prevention: prevalence of beliefs, attitudes, intentions and behaviour. Aust J Public Health 1995;19:18–23.
  71. Kelly RB, Shank JC. Adherence to screening flexible sigmoidoscopy in asymptomatic patients. Med Care 1992;30:1029–42.[CrossRef][Medline]
  72. Farraye FA, Wong M, Hurwitz S, et al. Barriers to endoscopic colorectal cancer screening: are women different from men? Am J Gastroenterol 2004;99:341–9.[CrossRef][Medline]
  73. Harewood GC, Wiersema MJ, Melton LJ III. A prospective, controlled assessment of factors influencing acceptance of screening colonoscopy. Am J Gastroenterol 2002;97:3186–94.[CrossRef][Medline]
  74. DeVellis BM, Blalock SJ, Sandler RS. Predicting participation in cancer screening: the role of perceived behavioral control. J Applied Soc Psychology 1990;20:639–60.
  75. Kremers SPJ, Mesters I, Pladdet IE, Van den Borne B, Stockbrugger RW. Participation in a sigmoidoscopic colorectal cancer screening program: a pilot study. Cancer Epidemiol Biomarkers Prev 2000;9:1127–30.[Abstract/Free Full Text]
  76. Honda K. Factors associated with colorectal cancer screening among the U S. urban Japanese population. Am J Public Health 2004;94:815–22.[Abstract/Free Full Text]
  77. Costanza ME, Luckmann R, Stoddard AM, et al. Applying a stage model of behavior change to colon cancer screening. Prev Med 2005;41:707–19.[Medline]
  78. Friedman LC, Webb JA, Richards CS, Plon SE. Psychological and behavioral factors associated with colorectal cancer screening among Ashkenazim. Prev Med 1999;29:119–25.[CrossRef][Medline]
  79. Sandler RS, DeVellis BM, Blalock SJ, Holland KL. Participation of high-risk subjects in colon cancer screening. Cancer 1989;63:2211–5.[CrossRef][Medline]
  80. Powe BD. Fatalism among elderly African Americans: effects on colorectal cancer screening. Cancer Nurs 1995;18:385–92.[Medline]
  81. Sarna L, Chang BL. Colon cancer screening among older women caregivers. Cancer Nurs 2000;23:109–16.[Medline]
  82. Kinney AY, Bloor LE, Martin C, Sandler RS. Social ties and colorectal cancer screening among blacks and whites in North Carolina. Cancer Epidemiol Biomarkers Prev 2005;14:182–9.[Abstract/Free Full Text]
  83. Herbert C, Launoy G, Gignoux M. Factors affecting compliance with colorectal cancer screening in France: differences between intention to participate and actual participation. Eur J Cancer Prev 1997;6:44–52.[Medline]
  84. Tang TS, Solomon LJ, McCracken LM. Barriers to fecal occult blood testing and sigmoidoscopy among older Chinese-American women. Cancer Pract 2001;9:277–82.[CrossRef][Medline]
  85. Taylor VM, Lessler D, Mertens K, et al. Colorectal cancer screening among African Americans: the importance of physician recommendation. J Natl Med Assoc 2003;95:806–12.[Medline]
  86. Greiner KA, Engelman KK, Hall MA, Ellerbeck EF. Barriers to colorectal cancer screening in rural primary care. Prev Med 2004;38:269–75.[CrossRef][Medline]
  87. Morales LS, Rogowski J, Freedman VA, Wickstrom SL, Adams JL, Escarce JJ. Sociodemographic differences in use of preventive services by women enrolled in Meidcare+Choice plans. Prev Med 2004;39:738–45.[Medline]




This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by McQueen, A.
Right arrow Articles by Tilley, B. C.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by McQueen, A.
Right arrow Articles by Tilley, B. C.
Related Collections
Right arrow Behavioral Science
Right arrow Behavioral Science: Interventions: Behavioral, Pharmacologic, Pharmacogenetic

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
Cancer Research Clinical Cancer Research
Cancer Epidemiology Biomarkers & Prevention Molecular Cancer Therapeutics
Molecular Cancer Research Cancer Prevention Research
Cancer Prevention Journals Portal Cancer Reviews Online
Annual Meeting Education Book Cell Growth & Differentiation