Objectives To assess the population coverage and diagnostic yield of offering an immunochemical faecal occult blood test (FIT) to non-responders to a flexible sigmoidoscopy (FS) invitation.
Design A cohort study conducted in a population-based colorectal cancer (CRC) screening programme. In this programme, eligible men and women aged 58 (Turin; 43 748 subjects) or 60 (Verona; 19 970 subjects) are invited, with a personal letter signed by their general practitioner, to undergo an FS. Bowel preparation is limited to a single enema self-administered at home. Subjects in whom one distal polyp >5 mm (≥10 mm in Turin) or at least one adenoma (one advanced adenoma or more than two adenomas in Turin) is detected at FS are referred for colonoscopy. People who do not respond to the invitation to undergo an FS are invited to have an FIT (OC-Sensor; Eiken, Tokyo, Japan; single sample, cut-off 100 ng/ml). Attendance rate and neoplasia yield were analysed in four consecutive birth cohorts.
Results Overall participation rate for the FS invitation was 39.3% in Verona and 29.9% in Turin. Of the eligible non-responders to the FS invitation, 19.3% (95% CI 18.9% to 19.7%) underwent an FIT. As a result, the proportion of people undergoing screening by FS or FIT was 55.2% in Verona and 39.3% in Turin, with no gender differences in either centre. FIT detected 8.3% of all advanced adenomas and 20.4% of all CRCs diagnosed at screening.
Conclusions A strategy involving the sequential offer of FS and FIT is a feasible and efficient approach. FIT in people not attending for FS increases screening uptake and detection of advanced adenomas and CRCs.
- Colorectal cancer
- mass screening
- faecal occult blood test (FIT)
- flexible sigmoidoscopy (FS)
- cancer prevention
- cancer epidemiology
- colorectal neoplasia
- endoscopic ultrasonography
- colorectal cancer screening
- endoscopic polypectomy
Statistics from Altmetric.com
- Colorectal cancer
- mass screening
- faecal occult blood test (FIT)
- flexible sigmoidoscopy (FS)
- cancer prevention
- cancer epidemiology
- colorectal neoplasia
- endoscopic ultrasonography
- colorectal cancer screening
- endoscopic polypectomy
Significance of this study
What is already known on this subject?
Of the available methods for colorectal cancer (CRC) screening, an immunochemical faecal occult blood test (FIT) and sigmoidoscopy (FS) are supported by evidence from experimental studies.
Available evidence suggests that subjects targeted for CRC screening have clear preferences for specific methods, determined by test characteristics, supporting the hypothesis that subjects' preferences and attitudes can influence the uptake achievable with different strategies.
However, no increase in screening uptake has been reported from studies in which people were offered a choice from among effective screening methods.
What are the new findings?
The sequential offer of FIT to non-attenders for FS screening can increase both participation rate and neoplasia yield.
As men show a higher rate of attendance for FS, whereas women show a higher rate of attendance for FIT, this strategy reduces gender-related differences in participation.
How might it impact on clinical practice in the foreseeable future?
Stepwise screening strategies combining the offer of FS and FIT is a feasible and efficient option for enhancing screening uptake and addressing patients’ preferences in a population-based programme.
Among available methods for colorectal cancer (CRC) screening, determination of faecal occult blood using an immunochemical test (FIT) and flexible sigmoidoscopy (FS) are currently recommended for population-based programmes in Italy.1 FIT was recently recommended as the faecal test of choice for population screening in the European Guidelines on quality assurance for CRC screening,2 based on the results of several studies showing a higher attendance and detection rate for FIT compared with a guaiac faecal occult blood test (FOBT).3–5 The adoption of FS is supported by experimental evidence from two studies showing a long-lasting and substantial reduction in CRC incidence and mortality.6 ,7
The availability of different tests represents a new scenario for mass screening programmes, as subjects' preferences and attitudes will probably influence the uptake achievable with different strategies. There is a growing body of literature suggesting that subjects targeted for CRC screening have clear preferences for specific methods, determined by test characteristics such as comfort, accuracy, invasiveness, embarrassment, and anxiety about preparation and procedure. Several reports have suggested that the choice of FIT might be driven by the simple and non-invasive character of the test, while people placing higher value on the performance of the test would tend to prefer endoscopy.8–13 These trends have been confirmed in the context of a large pilot project conducted in Italy.14 Non-attenders for FS screening mentioned worries about pain, discomfort and injury associated with the test as the main reason for their refusal. In contrast, people undergoing FS describe the accuracy of the test as the main reason to support their choice of screening method.
Gender, education and age influence uptake according to the reports from several studies showing a higher response rate among women invited to undergo FIT14–18 and a higher rate of attendance for FS screening among men and more educated or younger people.14 ,18 ,19 Therefore the tests cannot be considered as direct substitutes for each other.
The heterogeneity of patients' preferences on how to be screened therefore supports the adoption of strategies favouring their implementation as a possible means to improve participation in CRC screening. Provided that the available options are effective, the availability of a more acceptable, although less sensitive, test may favour equity of access, and it may also improve the health impact at population level, as a larger proportion of the target population would be screened. However, available evidence indicates that uptake does not increase when people are offered a choice from among effective screening methods.19 ,20
The aim of this study is to assess in two large population-based programmes the effect of a screening strategy involving the offer of FIT to non-responders to FS invitation, in terms of feasibility and population coverage, and also to determine the diagnostic yield of FIT in this setting.
We included in the analysis all subjects invited in the regional CRC screening programmes in Verona and Turin. In these centres, since 2003 (Turin) and 2004 (Verona), all men and women, aged 58 (in Turin) or 60 (in Verona) years of age, at average risk of CRC, included in the NHS lists are invited for FS screening. Every year a new birth cohort is targeted for recruitment in the programme.
General practitioners (GPs) are asked to exclude from the invitation subjects with a family (two first-degree relatives) or personal history of CRC, colorectal polyps or inflammatory bowel disease, those who have had a colonoscopy within the previous 5 years or an FOBT within the previous 2 years, or who have a medical condition that would preclude benefit from screening.
Eligible subjects are mailed a personal invitation letter, signed by their GP, with a pre-fixed appointment for an FS. They are asked to call the screening centre to confirm, modify or cancel their appointment. Those who agree to fix a test date are advised to visit a pharmacy to obtain the enema for bowel preparation. A leaflet containing a brief description of the screening procedure and mentioning its possible side effects is included with the letter. A reminder letter is mailed to all non-responders about 6 weeks after the initial invitation.
All people who do not respond to the invitation or the reminder, or who do not call the screening centre to report some condition precluding inclusion in the program, are invited after 6 months to undergo an FIT. The personal invitation letter for FIT, signed by the GP, includes a leaflet containing a brief description of the screening procedure and the advantages and drawbacks of the test. People who accept the invitation are required to visit a pharmacy to obtain the kit, including an information sheet with instructions for collecting, storing and returning the sample. No reminder is mailed to non-responders. In Turin, attenders for FIT screening are invited every 2 years until age 69, while non-attenders are offered two additional appointments for FS at 12 and 24 months after the initial invitation. No additional appointment for FS is scheduled in Verona, where both responders and non-responders to the FIT invitation are offered biennial FIT until age 69.
All the screening procedures, including assessments when indicated, are free of charge for the patient.
In both centres, FS examinations are performed in hospital endoscopy units, by experienced gastroenterologists, using 140 cm colonsoscopes. Bowel preparation is limited to a single enema self-administered at home within 2 h of the scheduled appointment.
The FIT system (OC-Sensor; Eiken, Tokyo, Japan) is an automated quantitative immunoturbidimetric assay with an average faecal sample mass of 10 mg and a preservative buffer volume of 2 ml. The local protocol stipulates a single-sampling screening procedure, with a cut-off concentration of 100 ng haemoglobin/ml buffer and no requirement for dietary restrictions.
In Verona, subjects who have one adenoma (any size) or a polyp >5 mm detected at FS are referred for colonoscopy (TC); in Turin, TC referral is restricted to those who have one polyp ≥10 mm, one advanced adenoma, or more than two adenomas detected at FS.
Histological classification of polyps and cancers is based on the WHO criteria.21 Advanced adenoma is defined as an adenoma with any of the following features: size 10 mm or larger, high-grade dysplasia, or a villous component >20%. Cancer is defined as the invasion of malignant cells beyond the muscularis mucosae. Patients with intramucosal carcinoma or carcinoma in situ are classified as having high-grade dysplasia.
Patients are classified on the basis of their most advanced lesion to determine the prevalence of pathological features.
We included in the analysis four consecutive birth cohorts in both centres (1945–1948 in Turin and 1944–1947 in Verona). Participation rate was calculated separately for FS, among eligible subjects receiving the initial invitation, and for FIT, considering only eligible non-responders to the FS invitation letter and reminder. Overall participation was calculated as the number of subjects having either FS or FIT in relation to all those invited for FS. In Turin, non-responders to the initial FS invitation and the subsequent offer of FIT screening who were screened after the additional appointments at 12 and 24 months were included among participants in the calculation of overall attendance rate.
CIs for the proportions were calculated using the exact method.22 RR and 95% CI were used as a measure of association between the outcome of interest and the variables under evaluation.
The number needed to screen (NNS) to detect one advanced adenoma or a CRC was derived from the inverse of the point estimates for the prevalence of the finding; 95% CIs were derived by inverting the values for the 95% CI for the prevalence risk proportions. We used SAS V.8.2 for all analyses. All statistical tests were two-sided and were considered significant at p<0.05.
Overall, 63 718 people (43 748 in Turin and 19 970 in Verona) in four birth cohorts were invited for FS. The attendance rate (figure 1) after the initial invitation and the mail reminder was 30.3% (39.3% in Verona and 26.0% in Turin). In Turin, after the additional invitations at 12 and 24 months, mailed to those people who had responded to neither the initial FS nor the subsequent FIT invitation, the overall proportion of subjects screened with FS increased to 29.9% (table 1). The rate of attendance for FS screening was higher among men than women (RR 1.24, 95% CI 1.21 to 1.26), and the trend was consistent in both centres.
Among non-responders to the FS invitation, about 8% (934/12113) were excluded in Verona and 18% (5862/32274) in Turin (figure 1), as they could no longer be traced (Turin: 6.0%, N=1956; Verona: 0.7%, N=87), or they had called to refuse any further contact (Turin: 6.3%, N=2041; Verona: 4.2%, N=508), or to report conditions precluding inclusion (Turin: 5.8%, N=1865; Verona: 2.8%, N=339). Among eligible non-responders, 19.3% (95% CI 18.9% to 19.7%) underwent an FIT, the attendance rate being significantly lower among men than women (RR: 0.71, 95% CI 0.68 to 0.74), with a similar trend in both centres. People responding to the FIT invitation made up 15.9% (95% CI 15.4 to 16.4) and 9.4% (95% CI 9.1 to 9.7) of those initially invited for screening in Verona and Turin, respectively. As a result, the proportion of people undergoing a screening test (either FS or FIT) increased from 39.3% to 55.2% in Verona and from 29.9% to 39.3% in Turin (table 1). To achieve the observed increase in the programme uptake, about 60% of the target population was mailed one additional letter, over and above the routinely scheduled invitation and reminders (ie, on average 0.6 additional letters per targeted person).
Of 20 944 people who attended for FS screening, 20 296 (96.9%) could be examined, while the remaining 648 were referred for a second examination because of inadequate bowel preparation and did not attend. Of the 1243 (10.0%) subjects referred for TC in Turin, 92.6% (1151/1243) had the test; the corresponding values in Verona were 1370 (17.1%) and 92.1% (1251/1370). As a result of FS screening (table 2), CRC was detected in 59 men (0.5%; 95% CI 0.4 to 0.7) and 27 women (0.3%; 95% CI 0.2 to 0.4); 834 men (7.6%; 95% CI 7.2 to 8.2) and 357 women (3.8%; 95% CI 3.4 to 4.2) were found to have advanced adenomas.
The FIT positivity rate was 5.6% (5.9% in Turin; 5.3% in Verona), and compliance with the recommended TC assessment was 78.8%: 74.3% (179/241) in Turin and 85.2% (144/169) in Verona. The neoplasia yield (table 2) was the same in both centres, with a higher prevalence among men of both advanced adenomas (RR 2.74, 95% CI 1.86 to 4.03) and CRC (RR 2.42, 95% CI 1.04 to 5.66). FIT detected 8.3% of all advanced adenomas and 20.4% of all CRCs in the study cohorts.
The NNS to detect one advanced adenoma with FS was 13 (95% CI 12 to 14) among men and 27 (95% CI 24 to 29) among women; the corresponding NNS to detect one CRC was 184 (95% CI 143 to 244) and 351 (95% CI 238 to 526). Among FS refusers screened with FIT, 41 (95% CI 32 to 52) men and 111 (95% CI 81 to 154) women had to be screened to detect one advanced adenoma; one CRC was detected for every 209 (95% CI 119 to 385) examinations among men and every 507 (95% CI 256 to 1042) examinations among women. To detect one advanced neoplasm (CRC or advanced adenoma), 51 people had to be invited for FS; the corresponding value when FIT was offered to FS refusers was 290.
A strategy aimed at addressing preferences of subjects targeted for CRC screening is feasible in the context of a population-based programme and can improve screening uptake. About 20% of those who refuse an invitation for FS attend the subsequent invitation to undergo FIT, resulting in a substantial increase in the proportion of people having a screening test. As men show a higher attendance rate for FS, while women show a higher attendance rate for FIT, this strategy also allows a similar screening coverage to be achieved in both genders. This trend was consistent in both centres, although the response rate was lower in Turin than in Verona for both tests.
FIT screening of non-attenders for FS also increases the yield of neoplasia, by detecting an additional number of neoplasms, making up about 8% of all advanced adenomas and 20% of all CRCs detected in the screened cohorts.
Similar trends in the rate of response to FIT invitation among FS refusers have been reported by a recent study23 performed in a larger trial comparing FIT (same test as in our study) and FS screening in an average-risk population aged 50–74. The inclusion of people older than 64, who show a higher prevalence of right-sided lesions, together with the adoption of a lower positivity cut-off (50 ng/ml) may explain the higher relative yield of advanced adenomas achieved by FIT screening in that study. Our study, conducted in the context of two large ongoing population-based programmes, in different geographical areas, was more specifically focused on the target population of FS screening, including only people around age 60. Our results should therefore be directly relevant to the implementation of a stepwise approach to mass screening.
The adoption of a sequential strategy based on once-in-a-lifetime FS screening followed by the invitation for FIT to non-attenders can reduce the organisational burden for the health system related to the need to issue regular invitations and to implement measures to sustain participation over time. Indeed, only non-attenders for FS are targeted for regular FIT screening.
Also, even if the test adopted in the first stage is more invasive than FIT, screening-related inconvenience may be reduced, as there is no need to undergo repeated tests to achieve the expected protective effect, and the probability of undergoing TC assessment is the same with repeated FIT as with FS screening, or even higher.
When considering the patient's perspective, this approach is questionable, as the invitees were not informed of the subsequent FIT invitation. In fact, some subjects screened with FS may have preferred to undergo FIT, if they had been told that this would be an opportunity in the case of FS refusal. However, it should be considered that those who do not demonstrate a distinct preference are oriented to follow, as is usually the case for regional or nation-wide programmes implementing single-modality testing, the screening policy adopted in our region, which identified FS offered once in a lifetime as the protocol of choice. People who do not respond to an FS invitation because of a clear preference for FIT screening are offered the chance to take up their preferred method in the second step. However, we are planning to modify the information given with the invitation, to present, consistent with the adopted approach, FS as the first choice test for CRC screening and FIT as the second option.
We cannot exclude the possibility that, as the knowledge of this approach becomes widespread, a higher proportion of people will decline the offer of FS to wait for the subsequent FIT invitation. This choice would diminish the efficiency and possibly the clinical impact of this strategy. However, even though the protocol was presented at several public meetings and all GPs have been fully informed about the screening options since the beginning of the programme, the preliminary data on screening participation among recently enrolled birth cohorts do not show any change in the trends presented in this analysis.
This stepwise approach also seems more effective than the strategy of offering the invitees the option to choose their preferred test. Two population-based studies, conducted in different countries, have already shown that the response rate was not increased when a direct choice between FS and FIT was offered, compared with the offer of a single test.19 ,20 In the Italian study, the offer of a direct choice was associated with a higher response rate among subjects with a high educational level, and participants showed clear preferences14: pain-averse patients tended to undergo FIT, whereas those who valued improved accuracy opted for FS. These findings thus suggest that such an approach may reduce screening accessibility. The need to decide on both whether to participate in screening and which test to undergo may reduce attendance by those who do not have a clear preference. Also, the effect of this strategy may be reduced by the need to convey a large amount of information to guide invitees' decision making. As in most population-based programmes, information about the characteristics, benefits and risks of the tests is communicated mainly through written material, mailed together with the invitation letter. Written communication represents a barrier for subjects who are not familiar with the process of orienting their actions on the basis of written information only,24 ,25 and the positive association of higher educational level with the likelihood of attending in the group that was offered the choice between FS and FIT in the Italian study would support such an hypothesis.
Personal encounters, allowing patients to ask questions about the information, to express their views and to share the decision with the provider,26 ,27 are probably necessary, particularly for the less-well educated, to guide patients' choice. However, the feasibility and sustainability of a shared decision-making approach, developed in the clinical context, should be assessed in a screening setting, with respect to both its organisational impact and the training requirements. Primary care providers and GPs rarely elicit patients' preferences and do not often adequately discuss the pros and cons of screening.28 Also, physician counselling has been found to have a positive influence on knowledge and beliefs, but not on attitudes regarding screening.14
This difficulty in effectively addressing patients' anxiety about screening procedures probably explains the low compliance with the TC referral among FS refusers with a positive FIT, compared with the average participation rate (82%) achieved in the Italian programmes using FIT as the only screening test,17 even though the same procedure was used to refer these patients for TC. This finding suggests that the decision to refuse the invitation for FS screening reflected a strong negative attitude towards endoscopy, which could not even be influenced by the FIT result.
As expected, the number of people needed to be invited to detect an advanced neoplasm is low with FS, while nearly 300 people needed to be invited for FIT to detect an additional advanced neoplasm among those who refused FS. However, this ratio is consistent with data from FIT-based programmes17 with similar uptake, and it can be justified, to reach an additional proportion of the target population. The NNS with FIT to detect one advanced neoplasm (67 for advanced adenoma and 331 for CRC) in our setting is consistent (or even slightly lower considering the restricted age range of our population) with the estimates from population-based pilot projects,5 or mass screening programmes,29 adopting the same test with the same cut-off level. This further suggests that FS refusers who respond to the FIT invitation indeed represent a subgroup of people at average risk of CRC who can be covered if offered the preferred test.
Although both FS and FIT may represent cost-effective options, based on simulation models,30 in order to assess the cost-effectiveness of this combined approach, even compared with single-modality testing, we still need additional information about the proportion of regular attenders and the cumulative yield of neoplasia over several screening rounds using FIT, either as a primary test or within a sequential approach.
Repeated invitations and reminders enhance screening uptake, but the increase observed in our setting is probably related to the response of people who had a negative attitude to FS and a preference for the alternative test. Indeed, although we do not have a randomised comparison in this study, we have already shown that a second invitation for an FS is associated with ∼4% response rate.31 The rate of response to the additional FS invitations in Turin is consistent with this hypothesis. The observed differences in participation between the two centres, already reported in experimental trials32 and in the ongoing regional programmes,17 are probably explained by the different sociodemographic background and the different degree of involvement of community resources (media, administrations, non-profit organisations, volunteer groups).
In conclusion, a strategy involving sequential offers of FS and FIT may represent an effective and efficient approach to CRC screening, allowing implementation of subjects' preferences without undermining equity of access.
Funding This study was conducted in the context of the screening programme organised by the Regional Health Authority within the National Health Service.
Competing interests None.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.