When societies offer screening programs for the entire population, they are trying to identify a relatively small number of individuals with early indication of disease, while the majority of the population will not have much benefit. It is therefore important to target people expected to be at increased risk and to identify determinants for accepting the screening offer.For the overall study population, we identified a very high participation rate of 82,6%. The difference to the 65,3% participation rate reported by a Danish registry study for 2015/2016 [16] can be explained by the different criterion of “ever-use” (at least once) employed in the present study. In comparison to other countries [6,7], Danes in general appear to have a high level of acceptance of the national colon cancer screening program. The overall high participation rate might be explained by a postal reminder which is sent every second year to non-participants. This might be an even more active alert for participation in CRC screening than an e-mail which has also been shown to be effective [17]. Caution must, however, be raised regarding such comparisons as different countries may have different organization and payment schemes for their screening programs.In agreement with previous studies from Denmark and UK, we found that female gender was associated with increased participation in CRC screening [16,18]. Women’s generally higher health-consciousness and preventive orientation might thus also manifest itself with respect to colon cancer screening. However, it also needs to be noted that a review by Wools et al. [7], including studies worldwide, found female gender rather to be a barrier than a facilitator, so findings about gender might be country-specific. The higher responsivity among older participants is in line with results from a 12-country study by Klabunde et al. [6] as well as by a review by Wools et al. [7] and might reflect a stronger awareness of older people about the fact that colon cancer risk increases with age and/or less restrictive time schedules among those who have left the labor market.
That high income was a positive predictor for screening participation in the entire study population as well as in individual educational subgroups is a finding consistent with the literature [7,16,19], but might nevertheless be considered surprising since participation in Denmark is free of charge. It is therefore likely that the influence of income is not a directly enabling one but might be mediated by differing subcultural norms, concerns and benefit expectations. Alternatively or additionally, groups with lower income might have other than direct financial opportunity costs. Thus, they might rather spend their time and energy on more imminent seeming problems or might not want to lose income if they work on an hourly basis or are self-employed.
An unexpected effect occurred for education. Commonly, a higher level of education has been identified as a facilitator for screening participation [7,19], and this was also the case in a prior Danish study based on registry data [16]. In contrast to that, in our study we observed that among the total group as well as in all individual income groups, people with the longest education (more than appr. 13 years in school) participated to a lesser degree.
Reasons for this discrepancy remain speculative at this point. It is possible that the shift in the Danish program from FOBT to FIT in 2018 might have made a difference in terms of raising participation rates in the lower educated groups or else that critical media reporting in recent years [20] on a low predictive value of the test (too many false positives), unwarranted coloscopies, and a 1% risk of things going wrong during coloscopy, has specifically deterred higher educated population segments, who might reflect more on such information, not to say understand the numbers.
Further, a certain amount of selection bias may have played a role. Our study population showed some overrepresentation of the higher educated while the lower educated segment was underrepresented when compared to the reference population. In particular, immigrants have been largely excluded, since the questionnaire was in Danish only, and particularly non-Western immigrants are known to have lower average income and be less likely to attend screening programs (e.g. 16).
It certainly appears particularly contradictory that the higher educational groups participated less while higher income was associated with more participation. When stratified for gender, the association related to educational attainment was driven mainly by the female segment of the study population, whereas the positive association with income was driven by the male participants only. Whether our observation is a spurious statistical finding, or whether there is a gender-specific difference in the influence of education and income on screening uptake will await further studies.
Reasons for this association are most likely diverse. Obesity as well as non-participation in general health screenings might reflect some level of negligence towards health issues. Otherwise, the obese in the targeted age group are more likely to already suffer from other chronic diseases, such as diabetes and/or coronary artery/vascular diseases, which might make them less likely to participate in colon cancer screening [23] because they might focus on coping with their manifest disease instead of a new, “hypothetical” health risk. In any case, lower uptake rates in this particular group might be problematic, since obesity is a significant risk factor for colon cancer [24,25].
A parallel effect was observed for smoking. Thus, those at higher risk for colon cancer, i.e. smokers, participated less than non-smokers. This finding agrees with results reported by comparable public cancer screening programs [26] and might be due to a tendency in smokers to have more pessimistic and avoidant beliefs about cancer [27].
Among men, having a moderate to high willingness to engage in or accept health risks was found to be a barrier for CRC screening. This finding is in line with prior studies indicating that a personal disposition like sensation-seeking is associated with more risky health behavior [28].
A recurrent issue when discussing uptake of screening offers has been whether people with inadequate health literacy would profit from written information to a similar degree as those with high health literacy. In our study, health literacy did not influence participation in the screening program. This contrasts with a previous review indicating that low health literacy is generally associated with poor cancer screening uptake [8]. The most positive explanation is that information provided in the Danish setting is sufficiently good to enable also people with more limited health literacy to read and understand the messages. However, the non-effect might also be due to a limited variance in health literacy as an adequate health literacy was observed among 83% of our sample, which is relatively high compared to samples from other countries [29].
Within the subgroup of those who had not participated, 61% expressed that if the FIT were replaced by a blood test, they would participate. The obvious explanation is that they prefer a blood test to a fecal test, which agrees with studies demonstrating that an unwillingness to deal with the collection of fecal matter is an important subjective reason not to participate [30]. The observed opinion shift could, however, also involve some degree of regret and/or some degree of social desirability by providing what is perceived of as the ‘right’ response to an (as yet) hypothetical decision situation.
A strength of the present study was the large sample (n>6,000) allowing for subgroup analyses. Further, a participation rate of 45% for this kind of internet distributed questionnaire study, while surely not optimal, is relatively high. Also, non-responder analyses revealed few differences between participants and non-participants suggesting no major selection bias. Moreover, the present study was part of a larger study on health-related issues, so participants were not biased by knowing that CRC screening was addressed when they decided to enter the study, nor did they know this while responding to the initial questions on individual attitudes and risk behavior. Another strength is that we included only participants who had already made an actual decision about screening participation, so our study is not based on hypothetical deliberations. As for limitations, we cannot exclude the existence of confounders, e.g. family history of cancer, which have not been controlled for in the present analyses. Further, the free access to screening will limit comparability to settings with out-of-pocket payment, but on the other hand also eliminate a potentially strong determinant from obscuring other potential influence factors. Eventually, we also cannot exclude any social desirability bias among the self-reported lifestyle factors included as well as answers towards the question about participation in screening which might have led to an overestimation of participation rates.