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
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 HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Watts, B. G.
Right arrow Articles by Tilley, B. C.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Watts, B. G.
Right arrow Articles by Tilley, B. C.
Cancer Epidemiology Biomarkers & Prevention Vol. 12, 339-349, April 2003
© 2003 American Association for Cancer Research

Intention to be Screened Over Time for Colorectal Cancer in Male Automotive Workers1

Beatty G. Watts, Sally W. Vernon2, Ronald E. Myers and Barbara C. Tilley

Department of Behavioral Science, The University of Texas M. D. Anderson Cancer Center, Houston, Texas 77030 [B. G. W.]; The Center for Health Promotion and Prevention Research, The University of Texas Health Science Center at Houston School of Public Health, Houston, Texas 77030 [S. W. V.]; Division of Medical Oncology, Department of Medicine, Thomas Jefferson University, Philadelphia, Pennsylvania [R. E. M.]; and Department of Biometry and Epidemiology, Medical University of South Carolina, Charleston, South Carolina 29425 [B. C. T.]


   Abstract
Top
 Abstract
Introduction
 Materials and Methods
 Results
 Discussion
 References
 
Intention is an important construct in health promotion research, yet very little is known about whether cross-sectional correlates of intention to be screened for colorectal cancer (CRC) also predict intention over time or intention change. We used survey data from The Next Step Trial, a worksite health promotion trial, to address the following questions: (1) What is the consistency over time of intention to be screened for CRC? (2) Are the patterns and magnitude of associations between intention to be screened and the Preventive Health Model variables consistent over time? (3) What are the predictors of improving weaker intention to be screened, i.e., changing to strong intention? (4) What are the predictors of no change in strong intention to be screened, i.e., maintaining strong intention? and (5) What is the predictive ability of the models to predict intention to be screened for CRC? The study population consisted of white male automotive employees who responded to baseline (1993) and follow-up (1994 and 1995) surveys and did not have CRC at baseline or develop it during the study period. Of 5042 eligible workers, 2903 (58%) returned a baseline survey, and 2556 (88% of survey responders) met eligibility criteria; 75% (1929 of 2556) returned the year 1 survey, and 74% (1892 of 2556) returned the year 2 survey. We fit logistic regression models separately for the Preventive Health Model variables measured at baseline and each outcome (intention at year 1, intention at year 2, improving weaker intention, and no change in strong intention). The prevalence of strong intention to be screened for CRC was approximately 60% on all three surveys. Overall, 66% maintained their baseline intention over time. The most consistent predictors of strong intention, improving weaker intention, and no change in strong intention were family support, belief in the salience and coherence of screening, prior screening, and lack of concern about screening-related discomfort. Intention measured at baseline predicted intention measured 1 and 2 years later. Perceived susceptibility and lack of fear and worry about a CRC diagnosis predicted improving weaker intention. Having a family history of CRC or polyps predicted maintaining strong intention. Plant factors, self-efficacy, and beliefs about polyp removal were not predictors beyond the baseline year. Basing intervention development on cross-sectional associations may miss important factors or may incorrectly assume that cross-sectional associations are stable over time. A more focused, tailored intervention may be developed using factors that consistently predict intention.


   Introduction
Top
 Abstract
 Introduction
Materials and Methods
 Results
 Discussion
 References
 
CRC3 ranks third among cancers in incidence and mortality for both men and women (1) . At present, the most promising methods of reducing incidence and mortality from CRC are periodic screening using FOBT and SIG or colonoscopy (2, 3, 4, 5, 6, 7, 8, 9, 10, 11) . However, a low percentage of the population adheres to the recommended screening guidelines (e.g., Refs. 12, 13, 14, 15, 16 ). Data from the Behavioral Risk Factor Surveillance System (16) show only modest increases in screening, from 19.6% to 20.6% between 1997 and 1999 for FOBT within the past year and from 30.3% to 33.6% between 1997 and 1999 for SIG. Data from the 1998 National Health Interview Survey also show low usage of CRC screening (17) .

Because CRC screening is low, there is a need for behavioral interventions to increase participation (18 , 19) . According to several behavioral science theories, intention is an immediate and necessary, although not sufficient, precursor to individual action (20 , 21) . Intention also may moderate or be moderated by other variables (22, 23, 24, 25, 26) . Empirically, intention to be screened is one of the strongest and most consistent factors associated with future cancer screening including breast (23 , 25 , 27, 28, 29) , prostate (30) , and colorectal4 (31, 32, 33) cancer screening.

Cross-sectional studies of CRC screening show that persons who intend to be screened are more likely to be older, have risk factors for CRC, have been screened in the past, hold certain beliefs about health, and respond to social norms (22 , 24 , 34, 35, 36) . Because associations were based on cross-sectional data, it is uncertain whether the same characteristics are associated with intention to be screened over time. In addition, little is known about the consistency or stability of a person’s intention over time. If intention changes over time, it would be useful, for intervention purposes, to know what the predictors of change in intention are. Such information could guide intervention development to reinforce a person’s strong intention or to strengthen weaker intention.

We examined a subset of data from a large cohort of current and former automotive employees participating in The Next Step Trial (37) to investigate consistency of intention over time, predictors of intention over time, and predictors of change in intention. The Next Step Trial was a theory-based, worksite health promotion trial of interventions to encourage CRC screening and dietary change (37 , 38) . The theoretic framework for the study drew constructs from the health belief model (39) , social cognitive theory (40) , theory of reasoned action (20) , and previous work by Antonovsky (41) on the sense of coherence in everyday health behavior. Constructs from these theories and models were combined to form the PHM (24) , a model that has been used to study intention and behavior for colorectal (24 , 35) and prostate cancer screening (30 , 35 , 42) . The PHM constructs measure background, cognitive, psychological, social support, intention, and program factors.

In this report, we used survey data from The Next Step Trial collected at baseline (1993), year 1 (1994), and year 2 (1995) to address the following questions: (1) What is the consistency over time of intention to be screened for CRC? (2) Are the patterns and magnitude of associations between intention to be screened and PHM variables consistent over time? (3) What are the predictors of improving weaker intention to be screened, i.e., changing to strong intention? (4) What are the predictors of no change in strong intention to be screened, i.e., maintaining strong intention? and (5) What is the predictive ability, e.g., sensitivity and specificity, of the models of intention to be screened for CRC?


   Materials and Methods
Top
 Abstract
 Introduction
 Materials and Methods
Results
 Discussion
 References
 
Study Population.
A number of epidemiological studies (43, 44, 45, 46) , including one using the cohort studied here (46) , found that pattern and model makers were at increased risk for CRC incidence or mortality. However, no studies identified a specific exposure. Employees in The Next Step Trial had been offered periodic screening since 1980 through a company-sponsored program. At the time the trial was initiated in 1993, fewer than 35% were being screened regularly (47) .

There were 5042 automotive workers in 28 worksites who were eligible for the trial by working in pattern- and model-making for at least 2 years at a minimum of 20% time (47) . In 1993, 2903 workers (58% of those eligible) responded to a baseline survey, and 2556 (88% of survey responders) met study criteria for a cross-sectional analysis of intention to be screened for CRC and correlates from the PHM (35) , the theoretic framework for the study. The cross-sectional study included white males who were free of CRC and remained so during the study period. These 2556 men were eligible for the prospective analyses reported here if they answered questions measuring intention on either follow-up survey. On the year 1 survey, 75% (1929 of 2556) did so, and in year 2, 74% (1892 of 2556) did so. To gain a broader understanding of normal fluctuations in the population, analyses were not restricted to individuals who responded to all three surveys. However, we repeated the analyses for the men who answered all surveys (n = 1602).

Details about The Next Step Trial including the primary outcome results have been published elsewhere (37 , 38) . Briefly, 28 worksites were randomized to receive an educational intervention or usual care. Employees at all worksites were offered CRC screening (i.e., DREs, FOBT, and/or SIG) during work time. Each worksite developed its own process for offering screening. In addition, employees at the intervention worksites received a mailed invitation to the screening program and an educational booklet tailored to the employee’s screening history followed by a telephone call to reinforce messages from the booklet. The booklet was based on behavior change theories and models (20 , 40 , 48 , 49) .

Data Collection.
Surveys were sent to all eligible workers at 1-year intervals in 1993, 1994, and 1995. Nonrespondents to each survey 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.

The PHM consists of five categories of variables: (a) background variables include demographics, risk factors, and medical history; (b) cognitive and psychological variables measure knowledge, perceptions, and beliefs; (c) social support and influence variables pertain to the family; (d) intention to be screened for CRC measures intention; and (e) program variables describe activities related to delivery of the worksite screening programs.

Background variables were abstracted from employment records or from the baseline survey. Cognitive and psychological, social support and influence, and intention variables were measured on the baseline survey. Program factors were collected at baseline by surveying plant medical staff at each of the 28 worksites to ascertain characteristics related to delivery of the screening program. Study records provided information about study worksite status (i.e., intervention or control worksite), medical history variables, and screening status.

Dependent Variables.
Intention to be screened for CRC measured on the year 1 and year 2 surveys was the dependent variable for the first two study questions: (a) What is the consistency over time of intention to be screened for CRC? and (b) Are the patterns and magnitude of associations between intention to be screened and PHM variables consistent over time? For these analyses, an intention scale score was computed using the average of responses to two items: I intend to undergo colorectal screening; and I do not intend to go through colorectal screening (reverse coded). To be included in the analysis for a given year, a respondent had to answer both items. These items, and the scale score, were measured using a 4-point Likert format from strongly disagree (1) to strongly agree (4) . Respondents with a scale score of 4 were classified as having a strong intention to be screened and compared with respondents scoring <4, who were classified as having weaker intention. This categorization is consistent with Manski’s theory that intention should describe the "best case" probability of completion of a behavior (50) .

Change in intention from baseline was the dependent variable for question 3 (What are the predictors of improving weaker intention to be screened, i.e., changing to strong intention?) and question 4 (What are the predictors of no change from strong intention to be screened, i.e., maintaining strong intention?). Specifically, we wanted to know whether the results from the baseline cross-sectional and prospective analyses were similar to the predictors of strong intention to be screened for the subgroup of men who changed from weaker to stronger intention compared with men who reported weaker intention in all 3 years. Likewise, we examined whether predictors were similar for the subgroup of men who had strong intention in all 3 years compared with men who changed to weaker intention. For these analyses, we divided men who responded to the baseline survey into two groups: (a) men who indicated a weaker intention (intention scale score of <4); and (b) those who indicated a strong intention to be screened (intention scale score of 4). Change in intention was defined as a change in baseline intention scale score (i.e., 4 or <4) at least once over the two follow-up measurement periods (i.e., year 1, year 2, or both). To answer question 3 about improving weaker intention, men whose baseline score changed to 4 ("strong") from <4 ("weaker") in year 1, year 2, or both were compared with men whose intention score was <4 on all three surveys. We labeled this analysis "changing to strong." To answer question 4 about no change in intention, we compared the characteristics of men whose "strong" intention did not change from baseline (a score of 4) with those of men whose intention changed to "weaker" (a score of <4) in year 1, year 2, or both. We labeled this analysis "maintaining strong intention."

We chose this method to define the comparison groups for the change in intention analyses because we wanted a set of predictors that would be useful in the presence of "noise" from fluctuating intention. That is, we wanted to model a process that was more reflective of what would be encountered in the "real world." We also wanted a clear and simple way to classify all potential participants in an intervention study. By including men whose scores were inconsistent over the two follow-up surveys, we avoided selection bias that might have occurred had we restricted the study population to the "best case" scenario.

Independent Variables.
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, screening history in the 2 years prior to baseline, and screening examination results between the baseline and year 2 surveys).

PHM cognitive, psychological, social support and influence, and intention variables were measured by scales and single items (51) . Scales measured belief in the salience and coherence of CRC screening (four items), perceived self-efficacy related to screening (four items), perceived susceptibility to colorectal polyps or cancer (three items), and worries or fears about being diagnosed with CRC (two items). Single items were used to measure belief in the efficacy of CRC screening, belief that polyp removal can prevent CRC, belief that CRC can be cured, and concern about CRC screening-related discomfort. Support for CRC screening from family members and receptivity to family member support for CRC screening were also measured using single items. Scale development and validation are discussed in detail elsewhere (51) .

All items and scales were measured using the same 4-point Likert format as the intention outcome variable. High scores corresponded to more of the variable being measured, e.g., more worry, greater intention, higher self-efficacy. The item measuring concern about screening-related discomfort was coded such that persons who agreed with the item were less concerned about discomfort. Because the response scores were skewed, and some cells were sparse when the entire range of the scale was used, all items and scales were dichotomized at <3 (disagree or strongly disagree) or >=3 (strongly agree or agree). To maintain a conceptually consistent definition of intention, intention was dichotomized at <4 (agree, disagree, or strongly disagree) or 4 (strongly agree) when analyzed as an independent variable, e.g., baseline intention used as a predictor of intention measured at year 1 or year 2.

PHM program variables were coded yes or no to indicate whether screening was offered all year or only during certain months, whether screening examinations (DRE and SIG) were offered at the worksite or off-site, whether FOBTs were distributed by mail or by other means, and whether the plant medical staff notified employees by mail or telephone to schedule an appointment for screening or whether responsibility to schedule an appointment was left to the employee.

Statistical Analysis.
Bivariate tests for statistical significance of the association between each of the independent variables, all measured at baseline, and the dependent variables including intention to be screened measured at baseline (35) , year 1, year 2, and change in intention (i.e., improving weaker and maintaining strong) were conducted using {chi}2 tests adjusted for clustering (52) .

Intention to be screened measured at baseline was also included as a predictor variable when intention to be screened at year 1 and year 2 were used as dependent variables. Likewise, intention to be screened measured at year 1 was examined as a predictor variable when intention to be screened at year 2 was used as a dependent variable. In addition, results of the screening exams during the study period (i.e., between baseline and year 2) were included as an independent variable for the year 2 analysis and categorized as no exam, normal exam, or abnormal exam.

Multivariable analyses were conducted to assess the relative importance of the independent variables measured at baseline in relation to intention measured at year 1 and year 2 as dependent variables and to identify the predictors of change in intention over time. We fit logistic regression models separately for each outcome: intention at year 1; intention at year 2; improving weaker intention (change in intention from weaker to strong); and maintaining strong intention (not changing intention from strong to weaker). Independent variables with a P <= 0.20 in bivariate analyses were included in step one of the logistic regression analyses as recommended by Hosmer and Lemeshow (53) . In subsequent steps, covariates with a P > 0.05 in the prior step were removed until all covariates in the model were statistically significant at P < 0.05. A variable had to be removed from all models being compared to be excluded in the final step so that the models were comparable and to adjust for variables significant at any measurement. Age and education were included in all multivariable analyses as continuous variables. Because this was a randomized controlled trial, intervention or control worksite status was also included in all multivariable models, even though it was not statistically significant in bivariate analyses. By including the intervention term, we adjusted for even modest intervention effects before drawing conclusions about other factors associated with intention.

All multivariable analyses were conducted using SUDAAN statistical software to adjust for the effects of cluster sampling (i.e., worksite), taking correlations within a worksite into account (52) . Individuals who were missing data for any variable in each logistic regression model were not included in the analyses for that model. From 6% to 9% were excluded due to missing data in all models except for year 2, where 22% had some missing data. The year 2 analysis was limited to men who completed all three surveys because intention measured at baseline and intention measured at year 1 were included as predictor variables. Odds ratios and 95% confidence intervals were used to summarize the analyses for the final multivariable models. The reference groups and categorization of the independent variables for the multivariable analyses are shown in Table 3Citation . As noted above, the analysis for question 2 was repeated for the men who responded to all three surveys (n = 1602).


View this table:
[in this window]
[in a new window]
  		Table 3 Odds ratiosa (95% confidence interval) for intention to have CRC screening among white male automotive workers who responded to the baseline (1993) and to the year 1 (1994) and year 2 (1995) surveys, The Next Step Trial

 
Finally, we wanted to assess the predictive ability of the models to correctly classify persons scoring <4 or 4. Predictive ability was measured by sensitivity, specificity, and, where appropriate, positive or negative predictive value. These calculations are an important step in model validation because showing association is not sufficient evidence of prediction. Models with high sensitivity and specificity or models with high positive or negative predictive value indicate that there is little misclassification for the questions of interest. To calculate the sensitivity and specificity, we used the coefficients from these models of strong intention to calculate the predicted probability of having a value of 4 (strong intention) for each study participant for each model. We then categorized the individuals as <4 or 4 using a cutoff value for the probability equal to the observed prevalence of strong intention (scale score value of 4). We calculated the sensitivity and specificity of our models of strong intention by comparing the classification based on our models with the observed value for each study participant. To assess our ability to identify at baseline those participants who had weak intention but who changed to strong intention over the study period, we calculated the predictive value negative using the same approach as for sensitivity and specificity but using a score of <4 for all 3 years as the cutoff. To assess our ability to identify at baseline those participants who had strong intention at baseline and who maintained strong intention over time, we calculated the predictive value positive using the same approach as for sensitivity and specificity but using a score of 4 for all 3 years as the cutoff.


   Results
Top
 Abstract
 Introduction
 Materials and Methods
 Results
Discussion
 References
 
As reported previously (35 , 47) , responders were more likely to be older, married, retired, nonsmokers, more highly educated, to have participated in prior screening, and to have a personal history of polyps as well as a family history of CRC or polyps. Table 1Citation describes the baseline characteristics of the study population at baseline, year 1, and year 2 as well as the baseline characteristics of those who responded to all three surveys. Although there was attrition, baseline characteristics showed a similar response pattern over all measurement periods as well as for the analyses of those who responded to all three surveys. Overall, the prevalence of strong intention to be screened measured at baseline was 57% (95% confidence interval, 56–58%), 60% (95% confidence interval, 59–61%) in year 1 and year 2, and 62% (95% confidence interval, 60–63%) for those responding to all three surveys.


View this table:
[in this window]
[in a new window]
  		Table 1 Baseline characteristics of the subset of male automotive workers who responded to the baseline survey (1993), and to surveys in year 1 (1994), year 2 (1995), and to all 3 years in The Next Step Triala

 
Question 1: What Is the Consistency Over Time of Intention to be Screened for CRC?
Compared with their baseline intention score, 77% of men in year 1 and 75% of men in year 2 stayed in the same category of intention to be screened (Table 2)Citation . Among men who completed all three surveys, 66% maintained their baseline intention over the three measurement periods. A higher percentage of men who stated a strong intention at baseline restated a strong intention both at year 1 (82%) and year 2 (79%) compared with men who stated a weaker intention at baseline. Of those with weaker intention, 71% in year 1 and 69% in year 2 reported intention consistent with their baseline intention (Table 2)Citation . Among men who completed all three surveys, 71% of those who reported strong intention at baseline and 58% of those who reported weaker intention were consistent over the three measurement periods.


View this table:
[in this window]
[in a new window]
  		Table 2 Consistency of intention to have a CRC screening exam among male automotive workers who responded to the baseline survey (1993) and to surveys in year 1 (1994), year 2 (1995), and all 3 years in The Next Step Trial

 
Question 2: Are the Patterns and Magnitude of Associations between Intention to be Screened and PHM Variables Consistent Over Time?
Bivariate analyses of the associations between the PHM variables measured at baseline and intention to be screened measured at year 1 and year 2 were similar to results for the previously reported cross-sectional analyses (35) . PHM background factors of increasing age, more formal education, being married, having a family history of CRC or polyps, having a personal history of colorectal polyps, participating in the screening program during the 2 years prior to baseline, and not being a current smoker were associated with having a strong intention to be screened (P <= 0.20). All PHM cognitive, psychological, and social support and influence variables met the P <= 0.20 criterion for inclusion in the multivariable analyses; however, only one program factor did so [i.e., plant notification of employees that they were due for screening was positively associated with intention (35) ]. In addition, having an abnormal exam result during the 2-year study period was associated with intention to be screened at year 2. Intention measured at baseline was associated with intention measured at year 1 and year 2, and intention measured at year 1 was associated with intention measured at year 2. Retirement status and study worksite status, i.e., intervention or control worksite, were not associated with intention at baseline, year 1, or year 2.

Multivariable analyses identified only one factor that was positively associated with strong intention at all three measurement periods: family member support for screening (Table 3)Citation . The magnitude of association for this factor remained stable over time. In contrast, the magnitude of associations at baseline decreased for all other variables.

Three factors previously reported to be associated with baseline intention (35) predicted year 1 intention: (a) perceived salience and coherence of CRC screening; (b) participation in the screening program during the 2 years before the baseline survey; and (c) perceived self-efficacy related to screening. High perceived salience and coherence of screening showed a strong positive association with intention in baseline cross-sectional analysis, but the strength of the association decreased substantially at year 1 and was not statistically significant at year 2. In addition, men who at baseline had a strong intention to be screened had 6 times and 3 times the odds, respectively, of having a strong intention 1 year and 2 years later, compared with men who had a weaker intention at baseline. Similarly, men who at year 1 had a strong intention to be screened had 6 times the odds of having a strong intention 1 year later, compared with men who had a weaker intention at year 1. Thus, odds ratios for intention as an outcome were greater when intention used as a predictor was measured closer in time to the outcome.

Four factors that were significantly associated with intention in the previously published cross-sectional analyses of the baseline data (35) were not predictors of intention measured beyond the baseline year: (a) belief in the efficacy of CRC screening; (b) belief that polyp removal can prevent CRC; (c) perceived susceptibility to CRC and polyps; and (d) having the plant schedule DRE and SIG examinations (Table 3)Citation . In contrast, four additional variables that were not significant in the previously published cross-sectional analyses of the baseline data (35) were predictors of strong intention in either year 1 or year 2: (a) having a family history of CRC of polyps; (b) having fear and worry about being diagnosed with CRC; (c) lack of concern about screening-related discomfort; and (d) receptivity to family member support for CRC screening (an unexpected negative association).

In addition, having an abnormal screening exam during the study period predicted intention to be screened in year 2.

As described above, we performed an additional analysis for the men who answered all surveys (n = 1602). We observed the same pattern of results when we restricted the multivariable analysis to men who answered the intention questions in all 3 years of the study (n = 1602), although the magnitude of the odds ratios was reduced.

Question 3 (What Are the Predictors of Improving Weaker Intention to be Screened, i.e., Changing to Strong Intention?) and Question 4 (What Are the Predictors of No Change in Strong Intention to be Screened, i.e., Maintaining Strong Intention?).
Bivariate analyses of the associations between PHM variables measured at baseline and both measures of change in intention, i.e., "changing to strong," and "maintaining strong," were similar to the published cross-sectional analysis (35) , with only a few exceptions. Age was not associated with either measure of change in intention. Contrary to expectation, men who worked in a plant that distributed FOBTs by mail were less likely to maintain a strong intention compared with men who worked in plants that did not mail FOBT kits. Similarly, men who worked in a plant that offered screening examinations on-site were less likely to change to strong intention compared with men who worked in plants that did not offer on-site exams. Having examinations scheduled by the plant was not associated with either measure of change in intention.

In multivariable analyses of both measures of change in intention, four variables were associated with maintaining or changing to strong intention (Table 3)Citation . Men who previously participated in the screening program, who perceived the salience and coherence of CRC screening, who lacked concern about screening-related discomfort, and who reported family member support for screening were more likely to change to strong intention and to maintain strong intention. Among men who indicated at baseline that they had strong intention to be screened, only one man had a score of <3 on the salience and coherence scale and also maintained a strong intention to be screened in all 3 years. Due to this colinearity, the salience and coherence scale was not included in the multivariable analysis for that model.

Men who changed to strong intention also had more formal education and reported increased perceived susceptibility to CRC, whereas men who had fear and worry about being diagnosed with CRC were less likely to change to strong intention. In contrast to expectation, men who were receptive to family member support for CRC screening were less likely to maintain a strong intention, that is, they were more likely to change from strong to weaker intention.

Question 5: What is the Predictive Ability, e.g., Sensitivity and Specificity, of the Models of Intention to be Screened for CRC?
The models of strong intention achieved a moderate level of predictive ability (Table 4)Citation . Sensitivity was moderate when predicting strong intention in the true positives, i.e., those with a score of 4. The models also had a moderate level of specificity when predicting weaker intention in the true negatives, those with a score of <4. The model was more successful in predicting strong intention (sensitivity, 86%) than weaker intention (specificity, 65%) at baseline. For the year 1 and year 2 models, sensitivity (80%) and specificity (75%), respectively, were similar. The models of change achieved a moderate predictive value negative score for those with weak intention at baseline, i.e., of those with a weak intention at baseline, 73% were correctly predicted to have weaker intention in the future (i.e., a scale score <4 in all 3 years). For those men with strong intention at baseline (score of 4), there was a similar predictive value positive score. Of those with strong baseline intention, 73% were correctly predicted to maintain strong intention (i.e., a scale score of 4 in all 3 years). The sensitivity and specificity for the model of change from weak to strong intention were similar to sensitivity for models of strong intention (65–76%). The model for maintaining strong intention yielded an extremely low specificity (8%).


View this table:
[in this window]
[in a new window]
  		Table 4 Validity measures with 95% confidence intervals (CI) of multivariable models of intention to be screened and change in intention to be screened, The Next Step Trial, 1993–1995

 

   Discussion
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
References
 
Compared with the results of the previously published cross-sectional analysis of the baseline data (35) , we found that PHM constructs associated with intention fluctuated in importance over time. Identifying the correlates or predictors of intention is important because studies have consistently found a positive association between intention and completion of a number of cancer screening behaviors (23 , 25 , 27, 28, 29, 30, 31, 32, 33) . The magnitude of those associations was 2.0 or greater when data were reported in that manner (25 , 27, 28, 29, 30 , 32 , 33) . In multivariable analyses of unpublished data from The Next Step Trial compliance (completion of all recommended screening examinations during the study period) and coverage (completion of at least one recommended screening examination during the study period), the odds ratios for strong intention were 5.6 (95% confidence interval = 2.89–10.83) for compliance and 3.8 (95% confidence interval = 1.78–8.10) for coverage.4 Thus, data from a number of studies are consistent with the view that positively influencing intention will increase the chances of screening completion.

We found no published studies that prospectively examined predictors of intention to be screened for any type of cancer or that examined predictors of change in intention. Our results show that basing intervention development on cross-sectional associations, as is frequently done, may miss important factors or may incorrectly assume that cross-sectional associations are stable over time. This possibility should be evaluated further in intervention trials. In our analyses, the most consistent predictors of strong intention to be screened, of changing to strong intention, and of maintaining strong intention were support from family members, the salience and coherence of CRC screening, prior screening, and lack of concern about screening-related discomfort. Results for social support and for the effects of prior testing are similar to the preponderance of the evidence reported from cross-sectional studies of intention to be screened for various types of cancer (Table 5)Citation .


View this table:
[in this window]
[in a new window]
  		Table 5 Multivariable cross-sectional studiesa of psychosocial constructs associated with intentionb to be screened for colorectal, prostate, breast, and cervical cancers in studies published 1989–2000 (N = 18)

 
The factor of salience and coherence was positively associated with baseline and year 1 intention and with both measures of change in intention, indicating that this factor may be more important in reinforcing strong intention among those who already have one and may also influence persons to change their intention from weaker to strong. Approximately 30% of men with strong intention at baseline and 40% of men with weaker intention at baseline had inconsistent intention over time. Future intervention trials might evaluate the effectiveness of emphasizing the salience and coherence of screening in these subgroups.

Few studies of intention to be screened have included screening-related discomfort as a predictor variable, and most studies, including ours, found no association with intention in cross-sectional analyses (Table 5)Citation . Results of our prospective analyses were consistent with the view that this factor may be important in influencing intention over time. In addition, a lack of concern about screening-related discomfort was positively associated with changing one’s intention from weaker to strong and with intention staying strong over time. Because screening-related discomfort may be an important barrier to initiating or maintaining screening, its role should be studied further in terms of factors that affect one’s perception, for example, anticipation of discomfort based on others’ reports or on one’s prior experience with screening.

Self-efficacy has been infrequently studied in relation to intention for any type of cancer screening, and the results are mixed (Table 5)Citation . In our study, it was strongly associated with intention in cross-sectional analysis and associated with predicted intention in year 1 (although the strength of the association was reduced), but it did not predict intention in year 2 or change in intention. These findings suggest that self-efficacy may be important in initiating screening intention but less important in maintaining it. The lack of consistency across models in the association between self-efficacy and behavior is not consistent with what would be predicted by social cognitive theory (40) , the theory of planned behavior (54) , or the transtheoretical model (55) , in which self-efficacy is hypothesized to be an important determinant of both intention and behavior. Because of its theoretical importance, this factor should be studied for its potential mediating or moderating effects on both intention and behavior.

The association between intention and future intention was stronger than the other major predictors, although, like other predictors, the magnitude of the association between baseline intention and intention in year 1 and year 2 decreased over time. Although not examined in other studies, it was expected that prior intention would predict future intention, just as prior screening is a predictor of future screening. Given the well-established association between intention and cancer screening behaviors (23 , 25 , 27, 28, 29, 30, 31, 32, 33) , this finding underscores the importance of intervention efforts to change a person’s weak intention or to reinforce strong intention.

Of the four factors not statistically significant beyond the baseline year, only perceived susceptibility predicted a change in intention, i.e., changing to strong from weaker intention. The preponderance of studies, including our baseline data (35) , showed that perceived susceptibility was positively associated with intention in cross-sectional analyses (Table 5)Citation . In our data, belief in the efficacy of the screening test and having the plant schedule exams were also associated with intention only in cross-sectional analysis. Other studies of the association between test efficacy and intention have yielded mixed results (Table 5)Citation . It may be that these factors are important in initiating screening intention but not in maintaining it.

The benefits of polyp removal as a predictor of intention have not been studied previously. It has only been in the past decade that evidence for the benefits of polyp removal has been documented (2, 3, 4, 5, 6 , 8 , 9 , 11) . The men in our study, who completed the baseline interview in 1993 and who were at increased risk for CRC, may have been aware of ongoing studies such as the Minnesota Trial (10) . It is possible that an awareness of the benefits of polyp removal may have influenced initiation of screening intention, but over time, its influence decreased in importance relative to other beliefs, attitudes, and behaviors.

Three factors were not significant in our previously published cross-sectional analyses (35) but were significant in year 1 and also predicted a change in intention. Consistent with other cross-sectional studies (Table 5)Citation , family history of CRC or polyps was not associated with intention in cross-sectional analysis; however, in our study, it predicted intention in year 1 and maintaining strong intention. Although family history is not modifiable, calling attention to its importance may change a person’s intention from weak to strong or may reinforce strong intention. Receptivity to family member support, in contrast to expectation, was negatively associated with strong intention in the year 1 prospective analysis and also with maintaining strong intention. We have no explanation for the apparent paradox that having a strong intention or maintaining it is consistently positively associated with reporting family member support for CRC screening, but that being receptive to such support was, in two of the models, negatively associated with intention. This finding may be due to chance and should be considered tentative until confirmed in other studies. The factor of fear and worry about being diagnosed with CRC was negatively associated with intention at year 1 and with change in intention from weaker to strong. Cross-sectional studies have reported inconsistent results for fear and worry (Table 5)Citation . An interesting finding in our study was that men who changed to strong intention reported increased perceived susceptibility and decreased fear and worry about being diagnosed with CRC. Future studies should examine the potential mediating or moderating effects of these two variables on intention and subsequent behavior.

Our previously published cross-sectional findings (35) regarding the demographic variables were consistent with the psychosocial literature on cancer screening behaviors in that demographic characteristics are overshadowed by other variables in multivariable analysis (Table 5)Citation . Only education showed a statistically significant association with intention, and its effect was modest.

The predictive ability analyses, e.g., sensitivity and specificity (Table 4)Citation , of the models may inform decision making about whom to target in intervention studies by indicating the percentage of study participants the model correctly identifies. The moderate scores indicate that intention may be influenced by factors not included in the models. Because the measures of predictive ability were tested on the data sets used to derive the models, the resulting values are likely to be higher than would be observed if the model’s predictive ability were tested in a new data set. The models of change provide a helpful beginning to guide the use of scant resources when developing interventions to target intention status, i.e., to improve weaker intention or to reinforce strong intention; however, these factors should be evaluated in a randomized trial.

There are several limitations that should be considered in interpreting our results. Two considerations that relate to external validity and that may affect the generalizability of study findings are our response rate and characteristics of the study population. Our overall response rate was similar to other worksite health promotion studies that used similar survey methods (56, 57, 58) , and there were few differences in baseline characteristics between respondents and nonrespondents to the year 1 and year 2 surveys. Although there may have been selection bias in the subset of men who responded to the baseline survey, additional selection bias, if present initially, does not appear to have been introduced during the course of the study. Keeter et al. (59) found that despite two different methods of conducting telephone surveys (labeled "standard" and "rigorous") yielding response rates of 36% and 61%, both surveys produced similar results. Thus, it is possible that a low response rate does not necessarily mean results cannot be generalized to the target population. Second, the study population was composed of white, male, employed or retired automotive workers who were identified as being at increased risk of CRC and who had been offered a company-sponsored program of free CRC screening. Nevertheless, the patterns of association between intention and predictor variables were generally consistent with those reported in the literature for other study populations shown in Table 5Citation .

Another consideration in interpreting our findings is that we measured intention to be screened for CRC in general rather than for each specific test that was offered, e.g., FOBT or SIG. Because this was a program in which different combinations of CRC screening tests were offered depending on a person’s risk status (e.g., age, history of polyps) and prior screening history, our outcome measure is more reflective of "real world" circumstances, where the CRC screening tests offered will vary depending on factors such as physician or patient preference and availability. A methodologic issue pertains to our use of the intention scale score of 4 to measure the outcome variable. Using only "strongly agree" for the intention outcome (i.e., a score of 4) may be subject to fluctuation in intention strength between questionnaires because individuals may not be able to consistently distinguish between "agree" and "strongly agree." Some of the inconsistency over time may be due to the tendency of extreme values to regress to the mean. However, we wanted to measure the strongest intention because it carries the greatest likelihood of performing the screening test recommended.

One predictor variable we did not include was physician recommendation. Because the program was delivered through the medical department at each worksite and received union endorsement, it may be possible that some men assumed that plant physicians recommended screening, but we could not evaluate this possibility. Physician recommendation combined with the support of family and friends has been shown to be an important correlate of intention to pursue other cancer screening behaviors (Table 5)Citation and should be included in future studies where applicable.

Methodologic strengths of the study include a prospective study design with multiple measurement periods, a well-defined study population, a large sample size, and verification through company records of prior screening status. A strength of our study is the use of a number of psychosocial constructs from different conceptual domains, many of which have not been simultaneously examined in relation to intention to pursue any type of cancer screening. In addition, by examining the beliefs of men with weaker intention at baseline, we were able to glean some understanding of a potentially nonadherent group. Most of these factors are modifiable and therefore should be evaluated in intervention studies to change intention and behavior. Likewise, we identified factors that predicted maintaining strong intention over time and may be useful in reinforcing initial strong intention to be screened. In conclusion, our study contributes to the literature by examining prospective predictors of intention and change in intention in a defined population of men at increased for CRC.


   Acknowledgments
 
We thank Carl de Moor for helpful comments on an earlier version of the manuscript. We thank Paul Callen and Eun Sul Lee for assistance with analysis and Arada Halder for assistance with the reference list.


   Footnotes
 
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.

1 Supported by National Cancer Institute Grant CA52605. Back

2 To whom requests for reprints should be addressed, at University of Texas-Houston School of Public Health, Center for Health Promotion and Prevention Research, 7000 Fannin, Suite 2560, Houston, TX 77030. Phone: (713) 500-9760; Fax: (713) 500-9750; E-mail: svernon{at}sph.uth.tmc.edu Back

3 The abbreviations used are: CRC, colorectal cancer; PHM, Preventive Health Model; FOBT, fecal occult blood test; SIG, sigmoidoscopy; DRE, digital rectal examination. Back

4 R. E. Myers, S. W. Vernon, and B. C. Tilley, unpublished data. Back

Received 2/ 4/02; revised 12/30/02; accepted 1/17/03.


   References
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
 References
 

  1. Jemal A., Thomas A., Murray T., Thun M. J. Cancer statistics, 2002. CA Cancer J. Clin., 52: 23-47, 2002.[Abstract/Free Full Text]
  2. Newcomb P. A., Norfleet R. G., Storer B. E., Surawicz T. S., Marcus P. M. Screening sigmoidoscopy and colorectal cancer mortality. J. Natl. Cancer Inst. (Bethesda), 84: 1572-1575, 1992.[Abstract/Free Full Text]
  3. Winawer S. J., Flehinger B. J., Schottenfeld D., Miller D. G. Screening for colorectal cancer with fecal occult blood testing and sigmoidoscopy. J. Natl. Cancer Inst. (Bethesda), 85: 1311-1318, 1993.[Abstract/Free Full Text]
  4. Atkin W. S., Cuzick J., Northover J. M., Whynes D. K. Prevention of colorectal cancer by once-only sigmoidoscopy. Lancet, 341: 736-740, 1993.[Medline]
  5. Winawer S. J., Zauber A. G., Ho M. N., O’Brien M. J., Gottlieb L. S., Sternberg S. S., Waye J. D., Schapiro M., Bond J. H., Panish J. F., Ackroyd F. W., Shike M., Kurtz R. C., Hornsby-Lewis L., Gerdes H., Stewart E. T., The National Polyp Study Workgroup. Prevention of colorectal cancer by colonoscopic polypectomy. N. Engl. J. Med., 329: 1977-1981, 1993.[Abstract/Free Full Text]
  6. Lieberman D. A., Weiss D. G., Bond J. H., Ahnen D., Garewal H., Chejfec G. Use of colonoscopy to screen asymptomatic adults for colorectal cancer. Veterans Affairs Cooperative Study Group 380. N. Engl. J. Med., 343: 162-168, 2000.[Abstract/Free Full Text]
  7. Selby J. V., Friedman G. D., Quesenberry C. P., Jr., Weiss N. S. A case-control study of screening sigmoidoscopy and mortality from colorectal cancer. N. Engl. J. Med., 326: 653-657, 1992.[Abstract]
  8. Hardcastle J. D., Chamberlain J. O., Robinson M. H. E., Moss S. M., Amar S. S., Balfour T. W., James P. D., Mangham C. M. Randomized controlled trial of fecal-occult-blood screening for colorectal cancer. Lancet, 348: 1472-1477, 1996.[Medline]
  9. Kronborg O., Fenger C., Olsen J., Jorgensen O. D., Sondergaard O. Randomised study of screening for colorectal cancer with faecal-occult blood test. Lancet, 348: 1467-1471, 1996.[Medline]
  10. Mandel J. S., Bond J. H., Church T. R., Snover D. C., Bradley G. M., Schuman L. M., Ederer F. Reducing mortality from colorectal cancer by screening for fecal occult blood. Minnesota Colon Cancer Control Study. N. Engl. J. Med., 328: 1365-1371, 1993.[Abstract/Free Full Text]
  11. Mandel J. S., Church T. R., Bond J. H., Ederer F., Geisser M. S., Mongin S. J., Snover D. C., Schuman L. M. The effect of fecal occult-blood screening on the incidence of colorectal cancer. N. Engl. J. Med., 343: 1603-1607, 2000.[Abstract/Free Full Text]
  12. Byers T., Levin B., Rothenberger D., Dodd G. D., Smith R. A. American Cancer Society guidelines for screening and surveillance for early detection of colorectal polyps and cancer: update 1997. CA Cancer J. Clin., 47: 154-160, 1997.[Abstract]
  13. Winawer S. J., Fletcher R. H., Miller L., Godlee F., Stolar M. H., Mulrow C. D., Woolf S. H., Glick S. N., Ganiats T. G., Bond J. H., Rosen L., Zapka J. G., Olsen S. J., Giardiello F. M., Sisk J. E., van Antwerp R., Brown-Davis C., Marciniak D. A., Mayer R. J. Colorectal cancer screening: clinical guidelines and rationale. Gastroenterology, 112: 594-642, 1997.[Medline]
  14. United States Preventive Services. . Task Force Guide to Clinical Preventive Services, Williams & Wilkins Baltimore, MD 1996.
  15. Solomon M. J., McLeod R. S. Periodic health examination, 1994 update. II. Screening strategies for colorectal cancer. Canadian Task Force on the Periodic Health Examination. Can. Psychol. Assoc. J., 150: 1961-1970, 1994.
  16. Centers for Disease Control and Prevention. Trends in screening for colorectal cancer-United States, 1997 and 1999. Morb. Mortal. Wkly. Rep., 50: 162-166, 2001.[Medline]
  17. Breen N., Wagener D. K., Brown M. L., Davis W. W., Ballard-Barbash R. Progress in cancer screening over a decade: results of cancer screening from the 1987, 1992, and 1998 National Health Interview Surveys. J. Natl. Cancer Inst. (Bethesda), 93: 1704-1713, 2001.[Abstract/Free Full Text]
  18. Vernon S. W. Participation in colorectal cancer screening: a review. J. Natl. Cancer Inst. (Bethesda), 89: 1406-1422, 1997.[Abstract/Free Full Text]
  19. Vernon S. W. Adherence to colorectal cancer screening: a brief overview. Ann. N. Y. Acad. Sci., 768: 292-295, 1995.[Medline]
  20. Fishbein M., Ajzen I. . Belief, Attitude, Intention and Behavior: An Introduction to Theory and Research, Addison-Wesley Reading, MA 1975.
  21. Prochaska J. O., DiClemente C. C. Transtheoretical therapy: toward a more integrative model of change. Psychotherapy: Theory, Research, and Practice, 20: 161-173, 1982.
  22. DeVellis B. M., Blalock S. J., Sandler R. S. Predicting participation in cancer screening: the role of perceived behavioral control. J. Appl. Soc. Psychol., 20: 639-660, 1990.
  23. Montaño D. E., Taplin S. H. A test of an expanded theory of reasoned action to predict mammography participation. Soc. Sci. Med., 32: 733-741, 1991.[Medline]
  24. Myers R. E., Ross E., Jepson C., Wolf T., Balshem A., Millner L., Leventhal H. Modeling adherence to colorectal cancer screening. Prev. Med., 23: 142-151, 1994.[Medline]
  25. Lechner L., de Vries H., Offermans N. Participation in a breast cancer screening program: influence of past behavior and determinants on future screening participation. Prev. Med., 26: 473-482, 1997.[Medline]
  26. Myers R. E., Hyslop T., Wolf T. A., Burgh D., Kunkel E. J., Oyesanmi O. A., Chodak G. African-American men and intention to adhere to recommended follow-up for an abnormal prostate cancer early detection examination result. Urology, 55: 716-720, 2000.[Medline]
  27. King E., Rimer B. K., Benincasa T., Harrop C., Amfoh K., Bonney G., Kornguth P., Demark-Wahnefried W., Strigo T., Engstrom P. Strategies to encourage mammography use among women in senior citizens’ housing facilities. J. Cancer Educ., 13: 108-115, 1998.[Medline]
  28. Cockburn J., Sutherland M., Cappiello M., Hevern M. Predictors of attendance at a relocatable mammography service for rural women. Aust. N. Z. J. Public Health, 21: 739-742, 1997.[Medline]
  29. Cockburn J., Schofield P., White V., Hill D., Russell I. Predictors of returning for second round screening at a population based mammographic screening programme in Melbourne. Australia. J. Epidemiol. Community Health, 51: 62-66, 1997.
  30. Myers R. E., Chodak G. W., Wolf T. A., Burgh D. Y., McGrory G. T., Marcus S. M., Diehl J. A., Williams M. Adherence by African American men to prostate cancer education and early detection. Cancer (Phila.), 86: 88-104, 1999.[Medline]
  31. Myers R. E., Balshem A. M., Wolf T. A., Ross E. A., Millner L. Adherence to continuous screening for colorectal neoplasia. Med. Care, 31: 508-519, 1993.[Medline]
  32. Sutton S. R., Wardle J., Taylor T., McCaffery K., Williamson S., Edwards R., Cuzick J., Hart A., Northover J., Atkin W. Predictors of attendance in the United Kingdom flexible sigmoidoscopy screening trial. J. Medical Screening, 7: 99-104, 2000.[Abstract/Free Full Text]
  33. Myers R. E., Trock B. J., Lerman C. L., Wolf T. A., Ross E. A., Engstrom P. F. Adherence to colorectal cancer screening in an HMO population. Prev. Med., 19: 502-514, 1990.[Medline]
  34. 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., 6: 44-52, 1997.[Medline]
  35. Myers R. E., Vernon S. W., Tilley B. C., Lu M., Watts B. G. Intention to screen for colorectal cancer among white male employees. Prev. Med., 27: 279-287, 1998.[Medline]
  36. Weller D. P., Owen N., Hiller J. E., Willson K., Wilson D. Colorectal cancer and its prevention: prevalence of beliefs, attitudes, intentions and behaviour. Aust. J. Public Health, 19: 18-23, 1995.
  37. Tilley B. C., Vernon S. W., Myers R., Glanz K., Lu M., Hirst K., Kristal A. R. The Next Step Trial: impact of a worksite colorectal cancer screening promotion program. Prev. Med., 28: 276-283, 1999.[Medline]
  38. Tilley B. C., Glanz K., Kristal A. R., Hirst K., Li S., Vernon S. W., Myers R. Nutrition intervention for high-risk auto workers: results of the Next Step Trial. Prev. Med., 28: 284-292, 1999.[Medline]
  39. Rosenstock I. M. The Health Belief Model: explaining health behavior through expectancies Glanz K. Lewis F. M. Rimer B. K. eds. . Health Behavior and Health Education: Theory, Research, and Practice, 39-62, Jossey-Bass Publishers San Francisco, CA 1990.
  40. Bandura A. . Social Foundations of Thought and Action: A Social Cognitive Theory, Prentice Hall Englewood Cliffs, NJ 1986.
  41. Antonovsky A. . Health Stress and Coping, Jossey-Bass San Francisco, CA 1979.
  42. Myers R. E., Wolf T. A., McKee L., McGrory G., Burgh D. Y., Nelson G., Nelson G. A. Factors associated with intention to undergo annual prostate cancer screening among African American men in Philadelphia. Cancer (Phila.), 78: 471-479, 1996.[Medline]
  43. Demers R. Y., Parsons K. C. Colorectal cancer incidence in pattern and model makers: evidence from a screening program. Am. J. Ind. Med., 26: 33-45, 1994.[Medline]
  44. Robinson C. R., Waxweiler R. J., McCammon C. S. Pattern and model makers, proportional mortality 1972–1978. Am. J. Ind. Med., 1: 159-165, 1980.[Medline]
  45. Swanson F. M., Belle S. H., Burrows R. W. Colon cancer incidence among modelmakers and patternmakers in the automobile manufacturing industry. J. Occup. Med., 27: 567-569, 1985.[Medline]
  46. Tilley B. C., Johnson C. C., Schultz L. R., Buffler P. A., Joseph C. L. Risk of colorectal cancer among automotive pattern and model makers. J. Occup. Med., 32: 541-546, 1990.[Medline]
  47. Tilley B. C., Vernon S. W., Glanz K., Myers R., Sanders K., Lu M., Hirst K., Kristal A. R., Smereka C., Sowers M. F. Worksite cancer screening and nutrition intervention for high-risk auto workers: design and baseline findings of the Next Step Trial. Prev. Med., 26: 227-235, 1997.[Medline]
  48. Janz N. K., Becker M. H. The health belief model: a decade later. Health Educ. Q., 11: 1-47, 1984.[Medline]
  49. Becker, M. H. The Health Belief Model and Personal Health Behavior. Health Education Monographs Vol. 2, No. 4, 1974.
  50. Manski C. F. The use of intentions data to predict behavior: a best-case analysis. J. Am. Stat. Assoc., 85: 934-940, 1990.
  51. Vernon S. W., Myers R. E., Tilley B. C. Development and validation of an instrument to measure factors related to colorectal cancer screening adherence. Cancer Epidemiol. Biomark. Prev., 6: 825-832, 1997.[Abstract]
  52. Shah B. V., Barnwell B. G., Bieler B. S. . SUDAAN User’s Manual 85 (Release 7.0), Research Triangle Institute Research Triangle Park, NC 1996.
  53. Hosmer D. W., Jr., Lemeshow S. . Applied Logistic Regression, John Wiley & Sons, Inc. New York 2000.
  54. Ajzen I. The Theory of Planned Behavior. Organizational Behavior and Human Decision Processes, 50: 179-211, 1991.
  55. Prochaska J. O., Velicer W. F. The Transtheoretical Model of health behavior change. Am. J. Health Promot., 12: 38-48, 1997.[Medline]
  56. Abrams D. B., Boutwell W. B., Grizzle J., Heimendinger J., Sorensen G., Varnes J. Cancer control at the workplace: The Working Well Trial. Prev. Med., 23: 15-27, 1994.[Medline]
  57. Heimendinger J., Feng Z., Emmons K., Stoddard A., Kinne S., Biener L., Sorensen G., Abrams D. B., Varnes J., Boutwell B. The Working Well Trial: baseline dietary and smoking behaviors of employees and related worksite characteristics. The Working Well Research Group. Prev. Med., 24: 180-193, 1995.[Medline]
  58. Levy M. C. . Correlates of Cancer-preventive Nutrition Practices in a Worksite Population, University of California Los Angeles Los Angeles 1994.
  59. Keeter S., Miller C., Kohut A., Groves R. M., Presser S. Consequences of reducing nonresponse in a national telephone suvey. Public Opinion Quarterly, 64: 125-148, 2000.[Abstract]
  60. Weller D., Pinnock C., Silagy C., Hiller J. E., Marshall V. R. Prostate cancer testing in SA men: influence of sociodemographic factors, health beliefs and LUTS. Aust. N. Z. J. Public Health, 22: 400-402, 1998.[Medline]
  61. Orbell S. Cognition and affect after cervical screening: the role of previous test outcome and personal obligation in future uptake expectations. Soc. Sci. Med., 43: 1237-1243, 1996.[Medline]
  62. Wardle J., Sutton S. R., Williamson S., Taylor T., McCaffery K., Cuzick J., Hart A. R., Atkin W. Psychosocial influences on older adults’ interest in participating in bowel cancer screening. Prev. Med., 31: 323-334, 2000.[Medline]
  63. Lauver D., Nabholz S., Scott K., Tak Y. Testing theoretical explanations of mammography use. Nurs. Res., 46: 32-39, 1997.[Medline]
  64. Burack R. C., Liang J. The acceptance and completion of mammography by older black women. Am. J. Public Health, 79: 721-726, 1989.[Abstract/Free Full Text]
  65. Savage S. A., Clark V. A. Factors associated with screening mammography and breast self-examination intentions. Health Educ. Res., 11: 409-421, 1996.[Free Full Text]
  66. Clemow L., Costanza M. E., Haddad W. P., Luckmann R., White M. J., Klaus D., Stoddard A. M. Underutilizers of mammography screening today: characteristics of women planning, undecided about, and not planning a mammogram. Ann. Behav. Med., 22: 80-88, 2000.[Medline]
  67. Baumann L. J., Fontana S. A., Brown R. L., Cameron L. Testing a model of mammography intention. J. Appl. Soc. Psychol., 23: 1733-1756, 1993.
  68. Stein J. A., Fox S. A., Murata P. J., Morisky D. E. Mammography usage and the Health Belief Model. Health Educ. Q., 19: 447-462, 1992.[Medline]
  69. McCaul K. D., Reid P. A., Rathge R. W., Martinson B. Does concern about breast cancer inhibit or promote breast cancer screening?. Basic Appl. Sociol., 18: 183-194, 1996.



This article has been cited by other articles:


Home page
 Cancer Epidemiol. Biomarkers Prev.Home page
A. McQueen, S. W. Vernon, R. E. Myers, B. G. Watts, E. S. Lee, and B. C. Tilley
Correlates and Predictors of Colorectal Cancer Screening among Male Automotive Workers
Cancer Epidemiol. Biomarkers Prev., March 1, 2007; 16(3): 500 - 509.
[Abstract] [Full Text] [PDF]

Home page
 Health Educ ResHome page
J. R. Stark, E. R. Bertone-Johnson, M. E. Costanza, and A. M. Stoddard
Factors associated with colorectal cancer risk perception: the role of polyps and family history
Health Educ. Res., October 1, 2006; 21(5): 740 - 749.
[Abstract] [Full Text] [PDF]

Home page
 Cancer Epidemiol. Biomarkers Prev.Home page
J. A. Tiro, S. W. Vernon, T. Hyslop, and R. E. Myers
Factorial Validity and Invariance of a Survey Measuring Psychosocial Correlates of Colorectal Cancer Screening among African Americans and Caucasians
Cancer Epidemiol. Biomarkers Prev., December 1, 2005; 14(12): 2855 - 2861.
[Abstract] [Full Text] [PDF]

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