Evaluation of the colorectal cancer screening Programme in the Basque Country (Spain) and its effectiveness based on the Miscan-colon model

Colorectal cancer (CRC) is the third leading cancer-related cause of death in developed countries. The European Union (EU) has the highest incident rate and ranks second in mortality of both sexs, with 446,000 newly-diagnosed cases each year and a mortality rate estimated in 214,000 cases annually [1].

In the Basque Country, one of the 17 autonomous regions of Spain, it is also the most frequent type of cancer. In 2008, 642 new cases and 286 deaths in women and 1227 new cases and 504 deaths in men were registered [2].

Different screening strategies have been proposed to reduce the CRC incidence and mortality, by means of different diagnostic tests. Previously, evidence of the reduction in mortality using the guaiac test (gFOBT) for population-based screening, showed a reduction in mortality of 10–16% [3‐5].

Although there are few studies demonstrating the impact on mortality of a CRC screening programme using immunochemical quantitative tests (FIT), several clinical trials show that these tests achieve a higher neoplasia detection rate and higher positive predictive values (PPV) than the gFOBT [6‐8]. In fact, the European guidelines of screening for CRC (2010) [9] recommended these tests for population-based screening programmes.

A recent study published by Zorzi et al. [10] established that the screening programmes based on FIT were associated with a reduction of up to 22% in CRC mortality.

In accordance with the European recommendation (2003) [11] and the National Health System’s strategy against cancer (NHS) [12, 13], in 2008 the Basque Government approved the implementation of a regional population-based screening programme for CRC. The programme was aimed at men and women between 50 and 69 years old, using one sample biennially of FIT and a colonoscopy under sedation as a diagnostic confirmation in positive cases. The programme started in 2009, reaching almost the whole target population (approximately 586,700 people) at the beginning of 2014. The main results found in the first period showed a high participation rate, as well as high adenoma and CRC detection rates [14, 15].

In order to measure the effectiveness of the Programme and its current strategy in comparison to no-screening, the MISCAN-colon tool [16], widely and internationally validated, was chosen.

The objectives of this study were to predict future scenarios and outcomes for the Basque population and to determine the epidemiological benefits of the screening programme in terms of incidence, mortality and years of life lost (L-y-L).

This kind of evaluation could be useful to those countries rolling out screening programmes in order to implement actions and guarantee their continuation.

The Basque Country CRC Screening Programme is population-based and its main strategy was based on: A) a coordinating office, including clinical epidemiologists and statisticians, to plan, organize, manage and evaluate the Programme; B) all residents from 50 to 69 years were invited biennially, taking into account the health centers and referral hospitals, in order to adjust the positivity expected and colonoscopy capacity; C) prior to the invitation, the coordinating office selected the target population and linked the database to the Basque population cancer and medical procedures registries to exclude individuals with a previously-diagnosed CRC, terminal illness and reported colonoscopy in the past five years; D) training and involvement of the primary care staff; E) individualized posted invitations providing information about the Programme. After 4–6 weeks from the initial invitation, the kit (FIT) was sent along with instructions and individualized bar code. This bar code allowed the sample and person to be identified when processing the result. Samples were collected at primary health centers and processed in centralized public laboratories under strict and total quality management systems; F) all results were reviewed by primary care physicians and introduced in “ad hoc” CRC prevention software. Letters were posted with results. In positive cases, participants were recommended to visit their general gractitioner, who referred them to the hospital for colonoscopy. G) Colonoscopies (diagnostic and therapeutic if needed) were performed in public hospitals under deep sedation by specialists. H) All cases were followed-up with close coordination between primary care and specialized units; I) every case was coded by the coordinating office staff following standard EU guidelines and Spanish network consensus [17]; J) interval cancer and complications were identified and monitored by registries linkage before invitation and after colonoscopy performance. The programme is identified in (Fig. 1).

This study was approved by the Basque Country’s Ethics Committee.

The FIT used was OC-Sensor Micro (Eiken Chemical Co. Ltd., Toyo, Japan) (from 2009 until now) and FOB-Gold (Sentinel CH. SpA, Milan, Italy) 2009–2010 in 15,000 invitations). The faecal-Haemoglobin (f-Hb) cut-off was 20 μg Hb/g faeces for both sexs. The decision to use one single sample of FIT and the biennial period between invitations followed the recommendations of Levis and van Rossum [18, 19], in order to reach the highest participation rate with the best balance between sensitivity and specificity.

A satisfactory colonoscopy was considered if the caecum was reached and the quality of colonic cleansing was coded higher than 6 in all segments measured by the Boston Bowel Preparation Scale (BBPS). The American Association’s classification was used for CRC and stages [20]. Accordingly, the results of the colonoscopy were coded and follow-up recommendations assigned to each one as: 1) Normal/No adenomatous pathology and will be invited to perform a screening test within 10 years; 2) Hyperplastic polyps and will be invited to perform a screening test within 10 years; 3) Low risk adenomas and will be invited to perform a test within 5 years 4) Intermediate risk adenomas and remain on colonoscopy surveillance within 3 years; 5) High risk adenomas and remain on colonoscopy surveillance within 1 year; 7) Cancer, neoplasia which infiltrates the submucosa layer ≥pT1) followed by the hospital specialists.

The main results from 2009 to 2014 were used in order to describe the main benefits of the Programme. For the simulation model, the result of the period of 2009–2012 was used for the Basque Country’s inhabitants, and the results obtained from the invitation during 2013–2014 were used to check and contrast the results obtained by the simulation on the MISCAN-colon.

The strategy of the CRC Screening Programme in the Basque Country has been implemented according to the recommendations of the EU [9], taking into account the target group and professionals when considering its implementation.

The main results of the Programme showed a high participation rate in both sexs in the three rounds from 2009 to 2014, possibly related to the implemented strategy, according to McGregor et al. [29], who demonstrated a relation to participation in both sexs (men OR 5.0; 95% CI 2.9 to 8.3 and women OR 3.8; 95% CI 2.3 to 6.5). Tinmouth et al. [30] also showed the importance of the family physician when providing information about the programme’s role after Programme invitation. However, Van Roosbroeck et al. [31] demonstrated a higher participation rate related to the type of invitation, higher in shipping kits to a participant’s home than when delivered by the rimary care physician (OR 2.96 95% CI 2.78 to 3.14). Combined strategies could be efficient to achieve a higher participation rate. Also, quality assurance plays an important role (Von Karsa et al., 2013) [32].

The f-Hb cut-off point chosen in FIT has generated a lot of discussion in terms of the number of colonoscopies to be performed, which was an initial limitation to the total extension of the Programme. However, it has not been modified in terms of cut-off age to deal with the management of positive cases, but that should be taken into account in successive rounds, according to van Rossum 2009 [33].

The lesion detection rate analysis reported a high trend in the first round with a significant decrease in successive rounds, following the same pattern as the positive FIT test. The largest decline occurred primarily in men and in AA. Denters et al. [34] found a significant decrease in PPV for AN (Advanced Neoplasia) between the first and the second round of 55% (132/239) to 44% (112/252), (p = 0.017). The PPV for CRC was 8% (20/239) in the first round vs 4% (9/252) in the second round (p = 0.024).

CRC detected by screening were in early stages (I-II) in 66.4%, contrasting with previous data (45.8%) (Departamento de Sanidad y Consumo et al., 2010) [2].

In the Basque Country Programme, considering its rapid extension and its high participation rate and lesions detected, a positive medium-to -term impact could be expected. This impact was suggested by Zorzi et al. [10] who found a better impact related to geographic locality and the implementation of screening, with higher reductions in mortality in women (RR=0.64; 95% CI = 0.51–0.80) than in men (RR 0.87 95% CI 0.73–1.04), but with significant results in all cases.

The choice of the MISCAN-colon model to simulate the impact of the Programme, both mid and long term, has given us the opportunity to establish a future scenario based on real data, regarding the incidence and mortality before screening as well as the Programme’s results after its start in 2009. One outstanding feature of this method was being able to count on internationally-renowned cancer registers, which make the study of the effectiveness in screening feasible (Anttila et al., 2015) [35].

In this respect, the incidence and mortality rates in the Basque Country are different than in other European regions [36]. When compared with European Population Registers, the Basque Country showed a higher incidence rate in men and an average rate in women compared to the Netherlands, Italy, and Scotland and North Thames in the UK. The mortality rate in men was also higher. However, these incidence and mortality rates showed an intermediate position for women [37].

The simulation applied to sexs offered a wider vision of CRC, which was not reflected in a majority of research, and which was, however, important to calculate the impact of screening programmes. In the current study, the impact of dealing with different population groups was evident, not only regarding the incidence and mortality of CRC, but also how both sexs behaved in participation, positive test rates and the rate of detected lesions. Hence, the programme’s impact was shown to be greater in men than in women, but unfortunately men participated less than women.

After a 30-year projection, and with participation rates adjusted to the results of the Programme, the decrease in incidence and mortality found seems compatible with what is reflected in current literature, although it is difficult to compare results, due to the dissimilarities in context, including simulations of 100% participation and short or indeterminate follow-up periods. However, the quality of simulation and the adaptation of parameters proved successful according to the real data provided by the Programme.

The reduction in incidence would start in the first ten years of the Programme’s implementation with significant increases over time. As other authors have stated, CRC screening not only decreases mortality, but it prevents new cases Ventura et al. [38], which contributes to minimizing the burden of the disease in the future. On the basis of higher incidence rates, the mid-to-long term impact could represent an important reduction in both the number of cases and death, according to Parente et al. [39] who found a significantly lower mortality rate in screening in 5 years compared to non-screening or pre-screening colorectal cancer patients (19% vs 37% and 41%; p < 0.001).

In this sense, the L-y-L for both sexs is very high and provided an important tool for regional and national authorities, as well as policy makers, to invest and support these types of programmes, taking into account organization and quality indicators. That recommendation was suggested by van Hees et al., [40] from the Netherlands.

Comparisons between programmes are difficult, as was suggested by Klabunde et al. [41], who found a range of invitation coverage from 30 to 100% and coverage by the screening Programme from 7 to 67.7%, overall participation rate from 7 to 67.7%, and first invitation participation from 7 to 64.3%. These differences could be minimized by implementing different measures to increase coverage and catchment in order to maximize the equality of access and the impact on public health recommended by Senore et al. [42]. One of the limitations of this study is that the classification of the risk of those with removed adenoma, due to the use of MISCAN model, had to be done based only on the size of the lesion, so those identified with other characteristics such as number of adenomas or the grade of dysplasia, had to be proportionally distributed [43].

Another limitation is the uncertainty in estimated adenoma prevalence, which was considerably higher than previously observed in other studies included, to build the MISCAN-colon model. This is, however, consistent with the fact that the study programme has a high participation rate that has been maintained throughout the study period and has not declined in new participants. Based on the robustness of the model, this maintained rate supports the prediction.

An important strength of our study is a well validated model based on several years of data from a high participation-rate population-based programme, directly reported by The Basque Country data and the concordant results observed.

In the future, some findings in FIT performance characteristics, with respect to repeating screening rounds, would be taken into account in order to increase efficiency (van der Meulen et al., 2016) [44]. NowaCurrently, we still have difficulty comparing data from different models, relatedd to a lack of randomized control trials on the effectiveness of FIT, and a lack of data on participation in surveillance and uncertainty in adenoma prevalence. Consequently, there is a need to carry out prospective cohort studies to evaluate the impact of the effectiveness of these programmes within the context of implementation and considering all the possible parameters and their influence.

The results obtained within this research are in line with previously published studies on cost-effectiveness analysis [45].

Nevertheless, the results of the projections offer a rather modest reduction of the main parameters measured. These projections indicate a need to consider how to improve the efficiency of currently implemented strategies. This incudes analyzing the possibility of implementing complementary or improved strategies such as the introduction of algorithms of risks, differentiating among men and women, familiar susceptibility (detected lesions subgroup analysis) or adjusting the cut off levels of the current test. Primary care physicians and authorities are key in maintaining the programme as it is described here. Primary care physicians are central to informing the population about the benefits of being screened and, thus, maintaining the high participation rates. Authorities are important in ensuring the level of investment in order to guarantee that no delays in the subsequent diagnostics and managing processes are generated. The latter is crucial not just from the perspective of the programme itself and its intermediate results (detected lesions as early as possible), but to improve the final outcomes on life expectancy and quality of life.

The Basque Country CRC Programme results are aligned to its strategy and comparable to other programmes. MISCAN model was found to be a useful tool to predict the benefits of the programme in the future. According to the parameters of simulation of MISCAN-colon and by means of the early obtained data of the Programme, the screening seems to be an effective strategy in order to reduce the incidence, mortality and L-y-L. These results provide further evidence on the efficiency of population-based CRC programmes. These data support the continuity of the programme and show the need for further improvements in the selected strategy to increase its efficiency.