Responses to procedural information about colorectal cancer screening using faecal occult blood testing: the role of consideration of future consequences

Abstract

Background  Colorectal cancer (CRC) screening participation is low despite its effectiveness in reducing CRC mortality. Identifying benefits and barriers requires consideration of specific characteristics of screening modalities.

Aims and Research Questions  To monitor the impact of providing information about CRC screening via faecal occult blood testing (FOBt) on intentions to participate. To investigate moderation by individual differences in consideration of future consequences (CFC).

Design, setting and participants  A total of 211 healthy adults (aged 45–59) with no experience of CRC screening were presented with eight consecutive statements about FOBt‐based screening in a web survey. Participants completed measures of i) intention (after each statement), ii) CFC and iii) the importance of screening practicalities (e.g. unpleasantness of completing the test) and benefits (e.g. early detection of cancer).

Results  An 8 (information) × 2 (CFC) mixed ancova showed that intentions varied across the eight statements. (P < 0.001): increasing after information about FOBt being completed at home (P < 0.001) before subsequently decreasing after information about the requirement to collect faecal samples (P < 0.001) in a plastic tub (P < 0.01) on three occasions (P < 0.01) with the low CFC group generally being less inclined to complete the test (P < 0.01). Two between‐group anovas demonstrated that the low CFC group attributed greater importance to practicalities of screening than the high CFC group while the opposite was found for the importance of benefits (both P’s < 0.001).

Conclusion  Deconstructing FOBt‐based screening pointed to specific benefits and barriers which can advance research into public preferences of screening and educational materials.

Introduction

After lung cancer, colorectal cancer (CRC) is the second most common cause of cancer death in the United States, ¹ despite the fact that a substantial number of these could be preventable through early detection and intervention. ² The faecal occult blood test (FOBt) is a method of screening for cancer and precancerous adenomas that randomized trials have demonstrated to reduce CRC mortality among users by 27%. ³ It is one of the screening options recommended in the United States ² and is used as the primary screening tool in a wide range of European countries. ⁴

The UK National Health Service (NHS) has recently rolled out an organized, population‐based colorectal cancer screening programme which sends FOBt kits biennially to men and women aged 60–74 years. Test kits are delivered by mail with written instructions about how to collect and return three separate stool samples in a freepost envelope within 14 days of sample collection, a format similar to that utilized by other nationalized CRC screening programmes. ⁴ Abnormal results are followed up by colonoscopy.

Uptake of CRC screening is low. In the United States, the main source of information denoting participation relies on self‐reported data from the National Health Interview Survey (NHIS), a survey of non‐institutionalized American civilians. Data from the 2005 survey indicated that the age‐standardized proportion of screening attendees for FOBt and/or endoscopy within the last 10 years was 50%. ⁵ In the UK, participation was measured objectively as part of the bowel cancer screening pilot where it was found to be 58% in the first round and 52% second round. ⁶ In other countries offering organized CRC screening programmes (Czech Republic, Israel, Poland, Japan and Finland), uptake ranged between 20 and 71%. ⁷ , ⁸ As a consequence, a large proportion of individuals remain unscreened and are therefore at risk of late diagnosis where treatment is less effective.

As there is often no initial contact between screening participants and health professionals, the quality of information provided to individuals by national programmes is of particular importance. ⁹ A great deal of effort has been made to produce the current information in the English screening programme and to ensure that it is ‘comprehensive, informative and balanced’, ¹⁰ yet there are still areas which require improvement such as strategies to optimize the accessibility and cultural sensitivity of written materials. ¹¹ , ¹²

Another area of on‐going concern is the unpleasantness of the test procedure. Evidence has shown that the administration of a FOBt is a major barrier ¹³ , ¹⁴ , ¹⁵ , ¹⁶ , ¹⁷ , ¹⁸ and that simplifying the procedure by using less arduous tests can improve uptake. ¹⁹ , ²⁰ , ²¹ For example, Hoffman et al. demonstrated in a US sample that individuals randomized to receive a faecal immunochemical test by mail (a more analytically specific test, which detects the human globin moiety of haemoglobin and requires fewer stool samples) were over 50% more likely to complete the test than the group sent a FOBt. ²¹ However, using alternative testing technology provides its own organizational challenges to health systems (particularly issues around degradability of haemoglobin ²² ) which make it difficult to implement in the current English Bowel Cancer Screening Programme. ²³

As a result, it is important to learn more about how people respond to information about the FOB test kit, especially as a previous study offered some qualitative evidence about how a negative response to FOBt‐specific information may prevent people from engaging with information about the screening programme as a whole (‘You read it and you say O’ Blimey. It is going to be unpleasant, embarrassing and unpleasant. I am not going to do that’ and that can be as far as they get in the leaflet.’ ¹⁰ , p. 11). In addition, a quote in a similar qualitative study suggests that there are specific procedural aspects of the test information which may be difficult to understand (‘I didn’t realise until I was being interviewed and it was pointed out to me that I had 14 days to do it when it was convenient to myself. I thought it had got to be 1, 2, 3 [days] right off...it isn’t everyone that can do it [defecate] three consecutive days’ ¹⁸ , p. 2431–2432).

Negative responses to information about the FOBt self‐sampling procedure might be particularly strong among individuals who have a general disposition to evaluate health behaviours according to their short‐ rather than long‐term consequences, a tendency which has been commonly referred to as being low in consideration of future consequences (CFC). ²⁴ People low in CFC have been found to be significantly less likely to engage in a variety of health behaviours including physical activity, diabetes screening, and HPV vaccine acceptance and more likely to engage in illness precipitating behaviours such as smoking. ²⁵ , ²⁶ , ²⁷ , ²⁸ , ²⁹ , ³⁰ Specifically, in the context of CRC screening, Orbell et al. have demonstrated that individuals low in CFC are more persuaded when positive consequences are framed as short‐term and negative consequences framed as long‐term. The opposite is true for individuals that are high in CFC. ²⁹ , ³¹ The findings of these studies demonstrate the impact of CFC on a person’s readiness (from here on referred to as intention) to participate in CRC screening. However, as participants were explicitly directed to long‐ and short‐term consequences (e.g. ‘some people find that taking part in screening gives them immediate peace of mind’), it is difficult to infer how people high and low in CFC would naturally interpret information about CRC screening programmes.

The present study focused on responses to information about procedural aspects of the FOBt in a sample of healthy English adults who were approaching eligibility age for screening but had not yet been invited and were therefore naïve to the intricacies of the FOBt stool‐sampling process. We presented information about CRC screening and the FOBt self‐sampling procedure in the same order in which information is presented to screening invitees to find out more about the impact of procedural information on the intention to participate in the programme. Secondary aims were to investigate how intentions may be moderated by an individual’s disposition to evaluate health actions according to their short‐ vs. long‐term consequences and whether this disposition affects consideration of benefits and practicalities of screening.

Methods

Sample and data collection

With approval from the local ethics committee, we approached healthy individuals aged between 45 and 59 (who had not yet invited in the English Bowel Cancer Screening Programme) to participate by completing a web‐based questionnaire. The majority of participants were recruited by sending en email to academic and non‐academic staff at University College London (UCL) that contained a web link to the study. Additional recruitment strategies included posting the web‐address on ‘Online Psychology Research UK’ ³² and snowballing (i.e. asking participants to pass on the web address to family and friends in the target age range).

Measures

Once participants were logged into the website, they received eight consecutive statements about the English Bowel Cancer Screening Programme and how to self‐administer the home‐based FOBt kit. The first introductory statement (S1, see Table 1) instructed participants to consider a new screening test for men and women of a similar age group introduced by the National Health Service. This was followed by six additional statements which specified the aim of the screening test, i.e. early detection of colorectal cancer (S2) and how the test is administered: the test is administered at home (S3), the test involves self‐sampling of bowel motions (S4), the test involves catching the bowel motion (S5), the test involves transferring a sample of the bowel motion on a test card (S6), the test requires repeating this process three times within 14 days (S7). Each of the statements introduced new information that was not directly inferable from the previous ones. After the seventh statement, participants were presented with a series of pictures illustrating the steps required to complete a FOBt kit (S8). After each of the eight statements participants were asked to indicate their intention to complete a FOB test kit on a 7‐point Likert scale by the question ‘How likely would you be to take part if you were offered the test immediately?’ Responses were made on a scale ranging from 1 (very unlikely) to 7 (very likely).

Table 1.

Intentions to participate after introducing information about the test

Statements	Description of the test
1	The NHS has introduced a screening test for men and women of a similar age group
2	This test can detect colorectal cancer and pre‐cancerous signs of colorectal cancer
3	This test is self‐administered in your own home
4	This test provides a simple way for you to collect small samples of your bowel motions
5	This test involves you collecting your stools in a plastic tub and sampling them for tiny amounts of blood
6	This test involves smearing a sample of faeces onto the test kit using a cardboard stick
7	The test involves sampling three separate bowel movements within 14 days
8	Pictorial description of the test

We used an abbreviated version of the CFC scale ²⁴ because previous research demonstrated compromised readability in the full scale. ²⁸ The abbreviated version consisted of four items: ‘I think about the future and this influences my behaviour today’; ‘I leave the future to take care of itself’’; ‘I’m prepared to make sacrifices now for benefit in the long run’; ‘I prefer to think of the ‘here and now’ rather than the future.’ The mean readability score of these items was 7.91 (as assessed by the Gunning Fog Index, GFI), indicating that participants had to have 8 years of education to understand these items at first reading which met readability recommendations ²⁸ and was considerably lower than the level of education required to understand the full scale (mean GFI = 14.49). Responses were made on a 7‐point Likert scale ranging from 1 (strongly disagree) to 7 (strongly agree), summed (range 4–28) and dichotomized at the median (20) to reflect low (4–20) or high (21–28) CFC. The adapted version had already been used previously and was found to be reliable. ³³ For the present study, we also found acceptable internal reliability as indicated by a Cronbach’s alpha of 0.82.

Participants were asked to indicate the extent to which ‘unpleasantness’, ‘time constraints’, ‘difficulty in remembering’ and ‘how complicated the test is’ were an important factor in making a decision to complete the colorectal cancer screening test on 7‐point Likert scales. Responses ranged from 1 (strongly disagree) to 7 (strongly agree). Internal reliability of the importance of practicalities scale was acceptable (α = 0.79). The importance of the benefits was assessed using two items on a 7‐point Likert scale: ‘the ability of the test to detect (i) cancer, and (ii) pre‐cancerous lumps is an important factor for me in making a decision to complete the colorectal cancer screening test.’ Responses ranged from 1 (strongly disagree) to 7 (strongly agree). Internal reliability in our sample was acceptable (α = 0.74).

Statistical analysis

We performed an 8 (information statements: S1–S8) × 2 (CFC: high vs. low) mixed‐model ancova, to compare differences in intention to take part in CRC screening across the eight information stages and by CFC while controlling for age group (45–49; 50–54; 55–59). Lack of sphericity was dealt with by using the Greenhouse‐Geisser statistic. We also performed two one‐way anovas to test for CFC differences in the perceived importance of benefits and practicalities in intention formation. spss 17.0 was used for all analysis. A priori power calculations determined that a minimum of 211 individuals would be required to detect a medium effect size of 0.25 ³⁴ with type I error set at 0.05 and power at 0.80 in the ancova analysis. A smaller sample size was needed for the same parameters in the anova analysis.

Results

Participants

A total of 211 participants agreed to participate and complete the survey. The mean age of participants was 51.4 years (SD = 4.2). A total of 61.9% of the participants were women, and the sample predominantly comprised of individuals who were white (91.4%), married or cohabiting (72.6%) and well‐educated (81.6% had a university degree).

The effect of information provision

An 8 (CRC information statements) × 2 (CFC: high vs. low) ancova revealed a main effect of information statement on intentions; F _{(3.33, 691.90)} = 8.15, P < 0.001, η ² = 0.04, demonstrating that providing information about the behavioural requirements of FOBt completion would affect intentions to participate. Fig. 1 plots the changes in intentions across the eight information stages. One important feature of this graph is that intentions did not decline immediately as a result of specifying to participants that the English Bowel Cancer Screening Programme was designed to detect colorectal cancer and pre‐cancerous signs of colorectal cancer [S2– S1 = −0.02 (95% CI = −0.11 to 0.14), P = 1]. Furthermore, there was a significant increase in intentions once we further specified that the test would be self‐administered at home [S3–S2 = 0.29 (95% CI = −0.42 to −0.16), P < 0.001]. Intentions subsequently decreased significantly when we introduced information about collecting faecal samples from a bowel motion [S4–S3 = −0.23 (95% CI = −0.32 to −0.13), P < 0.001] and collecting a bowel motion in a plastic tub [S5–S4 = −0.23 (95% CI = −0.35 to −0.11), P < 0.01]. A further explanation that the procedure would involve smearing traces of faeces onto a test kit using a cardboard stick did not significantly alter intentions [S6–S5 = 0.02 (95% CI = −0.09 to 0.12), P = 1]. However, the requirement of repeating the procedure three times within 14 days led to a further significant decline in intention ratings [S7–S6 = −0.24 (95% CI = −0.35 to −0.13), P < 0.01]. The pictorial overview of the stool‐sampling procedure did not significantly reduce intention ratings over and above the impact previously recorded in response to written information [S8 –S7 = −0.04 (95% CI = −0.16 to 0.09), P = 1].

Figure 1.

 Intention across the eight consecutive colorectal cancer screening statements by consideration of future consequences.

Moderation by CFC

There was a significant information stage x CFC interaction; F _{(3.33, 691.90)} = 2.98, P < 0.05, η ² = 0.02. Specifically, there was a steeper decline in intention following information about stool sampling (see Fig. 1), in the low compared with the high CFC group. Furthermore, the low CFC group was less inclined to participate across each of the eight information stages [M =5.33 (95% CI = 5.03 to 5.63)] compared with the high CFC group [M = 5.87 (95%CI = 5.60 to 6.15)], F _(1,208) = 6.92, P < 0.01, η ² = 0.03.

Finally, Fig. 2 shows that the low CFC group weighted practicalities as significantly more important than the high CFC group [M = 3.30 (95% CI = 3.05 to 3.54) vs. M = 2.64 (95%CI = 2.34 to 2.95)], F _(1,209) = 11.20, P < 0.001, η ² = 0.05. Conversely, the high CFC group rated the benefits of screening as significantly more important than the low CFC group [M = 6.38 (95%CI = 6.17 to 6.60) vs. M = 5.85 (95%CI = 5.62 to 6.08)], F _(1,209) = 10.69, P < 0.001, η ² = 0.05.

Figure 2.

 Consideration of future consequences and the importance of practicalities and benefits in forming a screening intention.

Discussion

This study improves our understanding of how the introduction of information about procedural aspects of FOB test completion affects the intentions to participate in CRC screening. Similar to findings reported by Ellis et al., ¹⁷ participants responded positively to information that the test is self‐administered at home. While the incremental introduction of details about how to complete the FOB test had a negative impact on screening intentions, it became evident that intentions to participate remained moderately high. Perhaps more interesting was the observation that not every information statement led to a significant reduction of intentions, further emphasizing those aspects of the FOBt self‐completion process which did lead to significant decline, such as the initial mention of stool sampling and the need to collect stools in a plastic tub.

Despite the cumulative impact of the initial details of sample collection, we recorded another significant intentional decline when informing participants about the requirement to repeat this process three times in 14 days demonstrating that participants remained sensitive to further details and that the prospect of repetition had an impact on their motivation over and above the general concept of stool sampling. The provision of photographs at the end of the information sequence resulted in only a small and non‐significant reduction in intentions.

A secondary aim was to examine whether individual differences in future orientation moderated responses to the information provided. We found a steeper decline in intentions and lower overall intentions among individuals in the group classified as low in CFC. In comparison with those high in CFC, these individuals expressed greater agreement that the practicalities of completing the test were important when deciding whether to complete a CRC screening test. In contrast, those classified as high in CFC rated the benefits of screening as more important. These findings confirm that people who generally emphasize the short‐term outcomes of health behaviours are prone to be dissuaded by the behavioural requirements of FOBt‐based screening.

Implications for health communication strategies to improve uptake

This study supports the growing body of evidence demonstrating the procedural barriers relating to FOBt‐based CRC screening and the extent to which these are further compounded by individual differences such as CFC. First and foremost, our observation that respondents were highly sensitive to specific aspects of the FOBt sampling procedure demonstrates that there is something to be gained by trying to make even subtle modifications to the procedure. This therefore strengthens calls for further refinement and eventual implementation of less arduous CRC screening modalities such as the faecal immunochemical test which minimizes the number of samples required at any given screening. In this respect, this study has advanced qualitative research ¹⁰ , ¹⁸ in pointing out more specific barriers to participation (such as the requirement to repeat the sample three times). It is equally important to note, however, that the findings of this study do not deride the general concept of home‐based sampling. Instead, by disentangling different aspects of the screening process, we were able to show that people generally appreciate the opportunity to complete this test at home.

Overall, the current study was based on a methodology that appears to be useful in pinpointing specific aspects of complex behaviours that are troublesome to users. Further research may wish to explore this methodology in other CRC test modalities (particularly in the wake of the proposed introduction of Flexible Sigmoidoscopy into the English CRC screening programme), ³⁵ thus enabling qualitative research on patient preferences to be developed in a quantitative framework. One such method that enables specific test attributes to be examined for their impact on intentions in a statistically efficient manner is that of discrete choice experiments (DCEs). ³⁶ Respondents are presented with a sequence of hypothetical choice scenarios that are composed of several attributes (e.g. test location) which can vary according to levels (e.g. home vs. hospital). Participants are then asked for their preference, thus enabling the utility of each attribute to be observed. ³⁷ As such, the study reported here constitutes a useful demonstration of the specific test attributes that could be tested as part of a larger DCE.

This methodology need not be limited to cancer screening and could be applied to other medical interventions where adherence levels could be increased. For example, it would be possible to use similar designs to the one used in this study to advance our understanding of the public acceptability of the self‐sampling tests used to detect sexually transmitted infections, as these are likely to invoke similar psychological barriers. ³⁸

Our findings with respect to the impact of time orientations also have unique implications for the advancement of health promotion strategies and for policy makers wishing to optimize screening participation in national programmes and within countries implementing opportunistic screening. For example, when communicating the aims of screening, it is important not to concentrate solely on the long‐term benefits but to also increase awareness of positive short‐term consequences of the test, e.g. that the test does not require hospital attendance. Providing messages affirming that the degree of unpleasantness can be overcome (as evidenced by more than 50% of the invited population returning the FOB test kit) could also minimize attrition among screening invitees. Using these messages as part of implementation intentions that act as specific instructions to overcome emotional barriers in individuals who are otherwise highly motivated to perform the test is one example of how this might be achieved. ³⁹

Highlighting more immediate benefits (e.g. peace of mind) should have a positive impact on those low in CFC and no detrimental effects for those who are more interested in long‐term gain. There are also potential applications in tailored communications and more individual‐based promotion strategies, which are becoming increasingly feasible as internet‐based communications become more popular. The introduction of a choice of test modalities, including those which may be more suitable to individuals with low CFC (i.e. less accurate but also less unpleasant), might also increase uptake. While this would be difficult to implement in standardized programmes such as that offered in England, ²³ it could be explored and would be of definite interest to health systems that promote user‐led choices. More feasible in the context of organized FOBt programmes would be attempts to offer additional practical assistance to directly tackle short‐term barriers, for example, by supplying gloves or other materials that reduce disgust‐related barriers.

The findings of this study may also have wider implications for socio‐economic inequalities in CRC screening ⁴⁰ and cancer screening in general. ⁴¹ As individuals with lower socio‐economic status preferentially consider short‐ rather than long‐term aspects of health behaviour, ⁴² strategies addressing short‐term negative consequences of screening participation may improve CRC screening participation and concomitantly reduce existing inequalities.

The observation that the final pictorial summary of steps required to complete the FOBt did not reduce intentions was perhaps not surprising as this statement did not introduce any new information. However, the fact that illustrating the processes that had previously only been verbally described did not reduce intentions was encouraging because providing information in such an easy to understand format could increase the public understanding of cancer screening, ⁴³ thus improving participation rates. ¹³ , ⁴⁴ , ⁴⁵ , ⁴⁶ , ⁴⁷

Limitations and future research

There were a number of limitations to this study which primarily relate to the composition of the sample, lack of behavioural data and the information presented to participants. The focus on information dissemination in this study meant that participants should not have previously participated in CRC screening. We therefore deliberately targeted participants below the age of 60 and were unable to collect data about actual screening participation. Although intentions are the most reliable predictor of behaviour, ⁴⁸ they are subject to self‐report biases such as ‘the acquiescence bias’ ⁴⁹ and the moral satisfaction associated with supporting a ‘good cause’ both of which could have led respondents to be overly optimistic about their likely participation in CRC screening. ⁵⁰

The current sample was well‐educated and predominantly comprised of employees at University College London. Because most of our participants were white, we could not examine the role of ethnicity as a potential moderator of our findings. As a result, it is likely that respondents’ intentions to participate in CRC screening were higher than one would expect for respondents from black and ethnic minority groups or lower socio‐economic background, particularly as these groups have been demonstrated to show a stronger aversion to the unpleasantness of faecal sampling. ⁵¹ However, it is important to note that this study did not aim to be a population‐based survey of intentions to participate in CRC screening; instead, the principle aim was to monitor changes in intentions as a result of exposing participants to information about the aims of CRC screening and its procedural demands. Given what is known about the role socio‐economic status and ethnicity in screening participation, ⁴¹ , ⁵² , ⁵³ , ⁵⁴ we may expect that a more socially diverse population would demonstrate more dramatic reductions in intentions than those shown in our educated, white, middle‐class sample. In this respect it is also important to bear in mind that the prevalence of individuals low in CFC is significantly higher among socially deprived groups. ⁴²

All participants within our study were presented information in the order in which this information is presented to individuals receiving a postal screening invitation. As a result, each piece of information was to some extent confounded by the cumulative information presented previously. While the pattern of intention ratings illustrated in Fig. 1 highlights changes in intention ratings at specific points, it may nevertheless be interesting to determine the independent impact of specific aspects of the test. For this purpose, future research could randomize the order of presentation, which would be particularly informative to those wishing to develop interventions aimed at improving the acceptability of the testing procedure itself.

The information presented to participants as part of this study was focussed upon the completion of a single FOBt. However, optimal reductions in CRC mortality are only observed if the full FOBt biennial programme is adhered to and are not achieved via the completion of a single FOBt. While the purpose of this study was to investigate specific barriers to a single FOBt completion, one should bear in mind that FOBt‐based screening only conveys optimal protection if the test is offered at regular intervals. ⁵⁵ Repeat adherence to CRC screening recommendations is likely to bring up another set of unique issues which need to be explored in future research, especially because this caveat is shared by other screening programmes (e.g. breast and cervical) where optimal mortality reductions are contingent on continued participation.

Finally, there are additional short‐term consequences of screening which we did not consider in this study, (e.g. the potential short‐term impact of screening on general and cancer‐specific anxiety). ⁵⁶ Future studies should examine the extent to which additional adverse emotional responses that are identified by public experience act as a deterrent to screening participation and how these relate to CFC. In turn, this will allow health communication strategies to be developed that minimize adverse responses to screening information.

Conclusion

Overall, this was an important hypothesis‐generating study with regard to FOBt‐specific barriers such as the requirement to handle faecal matter and repeated sampling which were shown to reduce intentions to participate in CRC screening. These barriers were particularly problematic for individuals who have a tendency to focus on immediate rather than long‐term consequences. The current study therefore presents compelling evidence of the need to increase efforts to tackle the short‐term barriers associated with the stool sampling procedure which are likely to act as deterrents to participation in CRC screening. The study presents various avenues for future research including the development of communication strategies, which could improve uptake and the application of the methodology to other areas of health research. In addition, our results suggest that consideration should be given to less arduous testing modalities in health‐care systems which afford such opportunities. Tackling negative short‐term consequences of screening is also likely to promote screening behavior among socially deprived groups (who have a higher prevalence of low CFC ⁴² ) and therefore has the potential to reduce the social gradient in CRC screening uptake. ⁶ , ⁴⁰

Conflict of interest

None declared.

Source of funding

This study was supported by a programme grant from Cancer Research UK to JW.