Factors related to longitudinal adherence in colorectal cancer screening: qualitative research findings

The effectiveness of screening in colorectal cancer prevention depends on sustained participation rates. The objective of this study was to explore factors related to the longitudinal adherence of screening behavior in the context of a biennial population-based cancer screening program. Eight focus groups were conducted with individuals who were invited two or three consecutive times to a population-based colorectal cancer screening program using a fecal occult blood test and who agreed to participate in the program at least once (n = 45). The criteria used to select the study members included adherence to fecal occult blood test maintenance, factors regarding their initial participation in the colorectal cancer screening, sex, and contextual educational level. The participants expressed a high level of satisfaction with the program; however, they showed a low level of understanding with respect to cancer screening. Consulting a general practitioner was cited by all participants as an important factor that mediated their final decision or influenced their behavior as a whole with regard to the program. Fear played a different role in the screening behavior for regular and irregular adherent participants. In the adherent participants, fear facilitated their continued participation in the screening program, whereas for the irregular participants, fear led them to avoid or refuse further screening. Having a close person diagnosed with colorectal cancer was a facilitator for the regular adherent participants. The irregular adherent participants showed some relaxation with respect to screening after a negative result and considered that further screening was no longer necessary. Considering the importance of primary healthcare professionals in the decision regarding sustained participation, it is important to better engage them with cancer screening programs, as well as improve the communication channels to provide accurate and balanced information for both health professionals and individuals.


Introduction
Colorectal cancer (CRC) is one of the most common tumors in the population and the third leading cause of cancer death among both women and men in developed countries (Ferlay et al., 2013).
However, a debate exists regarding which approach to implement. Benefits should be weighed against the costs, discomfort, complication rates, capacities needed, and potential differences in compliance. The Council of the European Union recommended the implementation of population-based screening programs for CRC using the gFOBT every 2 years in men and women between the ages of 50 and 74 years (CEC, 2003). Consequently, many countries, such as the United Kingdom, Spain, Finland, and France, have implemented population-based screening programs based on the gFOBT (Peris et al., 2007;von Wagner et al., 2011;Leuraud et al., 2013), whereas in regions of Italy, screening programs based on a fecal immunochemical test (FIT) have been adopted (Grazzini et al., 2004;Crotta et al., 2012). In several studies conducted in average-risk populations, a higher detection rate of advanced adenomas and CRC of the FIT, as well as a higher uptake rate have been identified compared with gFOBT screening (van Rossum et al., 2008;Hol et al., 2010;García et al., 2011;Tinmouth et al, 2015).
Consequently, the FIT is becoming a widely favored option for replacing the gFOBT.
Decreases in mortality rates for cancer in the population as a whole predominately depend on the percentage of participation in the screening programs (Parkin et al., 2008). The European Commission considers 45% an acceptable participation rate, whereas it recommends a participation rate of 65-70% (von Karsa et al., 2008). Most European programs achieve this accepted minimum; however, according to the European Commission, only Finland and the Netherlands reach the recommended rates (Ponti et al., 2017). Reductions in mortality may only be attained if uptake is adequate and sustained over time (Weller et al., 2009). Although high rates of adherence to repeat gFOBT screening have been reported in randomized trials (38-60%), longitudinal adherence to CRC screening in the population is expected to be substantially lower. A high level of ongoing and timely participation in screening is necessary to determine its effectiveness in reducing mortality from CRC (Calazel-Benque et al., 2011).
Screening invitations are typically repeated every 2 years, and the effectiveness of the gFOBT or FIT screening program is highly dependent on participation in multiple rounds, i.e., the longitudinal adherence. Ideally, eligible invitees accept the invitation to be screened at every screening round (Gellad et al, 2011;Steele et al., 2014;van der Vlugt et al., 2017).
A high rate of consistent participation increases the program sensitivity of CRC screening (Winawer et al., 1993;Nishihara et al., 2013;van der Vlugt et al., 2017). However, the success of a biennial screening program may be overestimated if there is a low willingness to participate in multiple rounds. To our knowledge, limited studies have examined longitudinal adherence to the FOBT over several years (Denis et al., 2015;Gellad et al, 2011;Kapidzic et al., 2014;Myers et al., 1993;O' Malley et al., 2002;Milà et al., 2012;Steele et al, 2014;van der Vlugt M, 2017).
Most studies have focused on one-time screening rather than longitudinal adherence. Myers et al. evaluated compliance rates with the gFOBT over 2 years among adult members of a health maintenance organization. They determined that only 23% of subjects completed two rounds of screening, with predictors of adherence including initial adherence and an age >65 years (Myers et al., 1993). O'Malley et al. used a targeted household telephone survey to evaluate adherence to an annual gFOBT in women over a 2-year period. They determined that only 29% of women completed two gFOBTs during the study period (O'Malley et al., 2002). Gellad et al. concluded that the proportion of individuals who received an adequate gFOBT screening was 14.1% for men and 13.7% for women over a 5-year period (Gellad et al., 2011). Therefore, these results show that longitudinal adherence is an important aspect of colorectal cancer screening, as participation is currently between 40-60%; thus, the loss of this percentage of participating individuals is an important aspect to consider. Moreover, there is limited knowledge regarding the determinants of longitudinal adherence to fecal testing in populationbased CRC screening programs.
The identification of potential determinants of inconsistent participation could aid in targeting the information to specific groups. Several studies and systematic reviews have been conducted to analyze reasons for participation in colorectal cancer screening (Honein-AbouHaiar et al., 2016;Khalid-de Bakker et al., 2011;von Euler-Chelpin et al., 2010); however, limited studies have assessed the reasons for longitudinal adherence. These limited studies indicate the factors that determine initial participation are different from the factors that determine longitudinal adherence to cancer screening (Lo et al., 2014;Lo et al., 2015;Palmer et al., 2014).
To develop interventions to encourage routine screening for colorectal cancer, it may be important to understand the differences among individuals who do and do not undergo repeat screening. Thus, the objective of this study was to explore factors related to the longitudinal adherence of screening behavior in the context of a biennial population-based CRC screening program using the FOBT in Catalonia, Spain.

Methods
We conducted a qualitative study using focus groups (FGs) that consisted of men and women between the ages of 50 and 69 years who had been invited to participate in the CRC screening program. FGs were chosen as the primary method of data collection because of their emphasis on participant interactions and potential to encourage greater candor, which make them particularly well-suited to investigate decision-making processes.

Sample selection
Our sample was derived from the population that had been invited to participant in CRC screening in 2010 in an industrial city of approximately 260,288 inhabitants, where the screening program was initially launched. From this population, we selected individuals who were invited at least twice, and the FGs were conducted after one year.
The screening program comprised a free, public, biennial, population-based screening program for colorectal cancer using the fecal occult blood test, which was provided free of charge to men and women aged 50 to 69 years. In the first four rounds, the guaiac test was used. An immunological test was subsequently applied. Eligible subjects were mailed a personal invitation letter, which was signed by the individual in charge of the screening program. Subjects with negative test results were informed by mail. All screened individuals with a positive FOBT were contacted by phone to provide information regarding the screening result and advise them that they would be referred for a colonoscopy examination. A more detailed description of the screening procedure is provided elsewhere (Peris et al., 2007).
Based on the available data from the CRC screening program, we devised a purposeful sampling strategy using a combination of intensity and maximum variation sampling (Patton, 2002) based on three criteria: 'prior screening behavior', 'sex' and 'educational level'.
The 'prior screening behavior' criterion was defined as colorectal cancer screening adherence in at least two consecutive screening rounds. We classified individuals as 'regular adherent participants' if they participated as many times as invited and as 'irregular adherent participants' if they participated fewer times than invited.
The 'educational level' criterion was defined using aggregate data, obtained from the census data (L'Hospitalet de Llobregat City Council, 2010), given that we did not have information regarding individual educational levels. Thus, we selected individuals who lived in a neighborhood (Area 1) with a poor educational level (26.64% of individuals with a level lower than primary studies) and individuals from a neighborhood (Area 2) with a better educational level (18.73% of individuals with a level lower than primary studies).

Data collection
The composition of the FGs was stratified by 'prior screening behavior', 'educational level' and 'sex' to ensure homogeneity in terms of the background and enable candid discussions regarding colorectal cancer screening procedures. In addition, we considered factors related to their initial participation in the colorectal cancer screening, such as the ease of recruitment (acceptance to participate immediately after receiving the screening invitation or six weeks after issue of the first invitation), the number of kits used and the FOBT result (negative or inconclusive FOBT) to ensure heterogeneity in terms of attitudes and experiences.
Eight FGs were conducted with 45 participants who had been invited two or three times to a population-based colorectal cancer screening program using the FOBT and who agreed to participate in the program at least once (Table 1). The sample size was determined by data saturation using concurrent data analysis. Prior to starting the FGs, the study team developed a topic guide that covered the key objectives of the study (Table 2). Written informed consent was obtained from all individuals who attended a focus group session. The study protocol was approved by the Clinical Research Ethics Committee of Bellvitge University Hospital (230/05).
After each FG was completed, the facilitators participated in a structured, self-administered debriefing session and completed their field notes. All FGs were audio or video recorded and transcribed verbatim. The transcripts were anonymized and reviewed by two members of the research team for accuracy. The FGs were conducted and transcribed in the native language of the participants (Catalan and Spanish). Selected data excerpts were then professionally translated into English for reporting.

Data analysis
Transcripts and field notes were subjected to thematic analysis (Braun & Clarke, 2006) with the assistance of ATLAS.ti software for data management (Muhr, 2004). Debriefing discussions and field notes served as the basis to refine the initial coding, which was subsequently discussed and refined by the research team, resulting in an initial set of 18 descriptive themes (Table 3).
We subsequently generated analytical themes by further interrogating the data set drawing on the constant comparative method (Glaser, 1965). Codes and emerging themes were then discussed, revised, refined and agreed upon by the research team through critique and consensus. As a result, three overarching analytical themes were established: (1) common factors underpinning the screening experiences and decisions of regular and irregular adherent participants; (2) common factors interpreted in opposing ways by regular and irregular adherent participants; and (3) differential factors across regular and irregular adherent participants.

Common factors underpinning the screening experiences and decisions of regular and irregular adherent participants
These factors were identified in both the regular and irregular adherent participants when explaining their views regarding the program and their screening experience.

Perceived benefit of prevention
One key overarching finding was that the system of beliefs concerning generic preventive health issues of both the regular and irregular adherent participants did not appear to significantly differ (Table 4, quotes 1-3).
With respect to the specific convergences between the regular and irregular adherent participants in relation to their screening experience and decisions, they were particularly significant because they were expected to play a decisive role in the participants' decision-making process during the screening period. However, this expectation was not supported because they were shared by the regular and irregular adherent participants. These factors were as follows: a lack of comprehension, a lack of media information and a high level of satisfaction with the program.

Lack of comprehension
The lack of comprehension referred to all manifestations of incomprehension that were more or less explicitly expressed by the participants, which mainly comprised difficulties in comprehension concerning the process to be followed to participate in the program, the periods established between rounds, the age limits established by the program and, in general, the rationale underlying a population-based screening program or the preventive health actions/policies that tend to collide with everyday life views (Table 4, quotes 4-6).
Initially, the possibility of incomprehension was valued as a handicap or a barrier to participation; however, these manifestations of incomprehension were identified in both the regular and irregular adherent participants. Therefore, it cannot be concluded that they influenced adherence to the program.

Lack of information regarding CRC in the media
The lack of information regarding CRC in the media was a recurrent complaint that both the regular and irregular adherent participants highlighted at various points during the FG sessions.
They tended to refer mainly to the information provided on TV and the written press regarding CRC. They also identified differential treatment by the media, for example, in relation to other cancers, such as breast or lung cancer, or in relation to other diseases that tend to be considered thematically relevant to the public in general (Table 4, quotes 7-9).

Satisfaction with the program
A high level of satisfaction with the program is another factor that was identified in the regular and irregular adherent participants. This satisfaction was expressed through three main considerations: 1. The elements of comfort provided by the program procedure, which enabled them to protect their health and act preventively without going to a hospital or consultation (Table 4, quote 10).
2. The clarity by which the instructions are given to the patient during every step of the screening process (Table 4, quote 11).
3. The health benefit that this type of service introduces to the population in general and to them in particular (Table 4, quote 12). Therefore, the views and positions regarding preventive health activities in general and the screening program in particular were convergent and followed similar patterns. However, the narratives and arguments of the regular and irregular adherent participants also indicated key factors that helped explain the differences in their actual decisions regarding CRC screening.
These factors were divided into two types: (a) common factors interpreted in opposing ways, which were shared between the regular and irregular adherent participants but interpreted in opposing ways in each case's decision-making process; and (b) differential factors, which were identified exclusively in the adherent or non-adherent participants.

Common factors interpreted in opposing ways by regular and irregular adherent participants
These factors were identified in both the regular and irregular adherent participants as important factors that played a significant role in the decisions of the participants regarding the screening.
The particularity of these factors is the fact that they were interpreted in opposing ways by the regular and irregular adherent participants: the same factors opposed implications. These two factors were fear and consulting a general practitioner.

The role of fear in participants' decision-making
Fear was recursively based on the regular and irregular adherent participants' views and explanations. These fears were related to the illness, colorectal cancer, and several questions that typically surround it: fear of suffering, fear of the way of life of sick individuals, and other questions. Nevertheless, these fears were interpreted in two clearly opposed ways by the regular adherent and irregular adherent participants in terms of the decisions they made during the screening period.
In contrast, the regular adherent participants identified this factor to explain and argue why they decided to participate every time they were invited (Table 4, quotes 13-14). The irregular adherent participants also used this factor to explain and argue why they sometimes decided to stop participating (Table 4, quotes 15-16). Therefore, the possibility of being in fear acted as an encouragement to participate in every round for the regular adherent participants, whereas for the irregular adherent participants, fear was sufficiently paralyzing to modify their screening behavior over time.

Consulting a GP as part of the decision-making process
Consulting a general practitioner was a common practice between the regular and irregular adherent participants that was conducted after receiving the invitation to participate and before making the decision to participate and maintain this decision throughout the screening period. This factor was cited by all participants as an important factor that mediated their final decision or influenced their behavior as a whole with respect to the program. However, this influence was again exerted in two opposing ways for the regular and irregular adherent participants.
The regular adherent participants experienced an action of reinforcement (in the form of additional explanations, showing support, making the patient aware of the importance of the preventive action, and other factors) as feedback from the GPs when they told them about their invitation to the screening program (Table 4, quotes 17-18). The irregular adherent participants experienced a lack of reinforcement (not sufficient importance attached by the GP, a lack of needed explanations, and other factors) as feedback from the GP, which they tended to link to incorrect medical attention (Table 4, quotes 19-20).

Differential factors across regular and irregular adherent participants
These factors were identified solely in the regular or irregular adherent participants. Therefore, they appeared to be key factors in the decisions made during the screening period.

Regular adherent participants: having a close person diagnosed with CRC
Having a close person diagnosed with CRC was a distinct element of the regular adherent participants' experiences (Table 4, quotes 21-24). The effects of both the closeness of the experience and the level of affectation were of a relative nature, which indicates that the experience could refer to nuclear or extended relatives, as well as friends, work mates or neighbors. Furthermore, the participants' experiences could refer to fatal cases of CRC affectation or a wide range of cases that were perceived as difficult by the participants.
Although this factor was significant in all adherent groups, it was not used by the participants to attribute relevancy to their arguments when explaining their adherent behavior.

Irregular adherent participants: prioritization issues and being relaxed about screening after a negative result
A distinct factor that characterized the experiences of the irregular adherent participants was the prioritization of everyday tasks and activities over the actions needed to participate in the screening program.
As a result, all actions with regard to program participation (e.g., answer the letter, ask for the FOBT kit, collect the samples, and send them back) were not regarded as preferential in the context of the participants' everyday life. Therefore, these actions were postponed until or beyond the deadline established by the program to participate in each round (Table 4, quotes 25-26).
A key argument to explain these prioritization issues was the emergence of a relaxation effect after a negative screening result. The irregular adherent participants highlighted that after they obtained a negative result in a previous round, they tended to feel safe and reassured, thus assuming that 'everything is OK' with regard to their health as a two-year period was not viewed as sufficient time for anything to have changed in this respect (Table 4, quotes 27-28).

Discussion
Our study identified factors related to the longitudinal adherence of screening behavior for colorectal cancer in Catalonia, Spain. Facilitating factors and barrier factors were identified by irregular and regular participants in a CRC screening program through FGs. The factors identified by this study include a lack of comprehension, fear of the consequences of screening, inconsistent or inadequate support for screening from providers and the media, and a relaxation effect after a negative result in the FOBT, which coincided with the results of other studies that analyzed facilitators and barriers for participation in colorectal cancer screening using qualitative methodology and the FOBT, colonoscopy or flexible sigmoidoscopy as screening methods (Honein-AbouHaidar et al., 2016;Jones et al., 2010;Wools et al., 2016).
Fear: The presence of fear as an important influence has been documented in other screening studies (Amonkar et al., 1999). Cancer fear may be a facilitator or a deterrent, depending on the specific aspect of the fear. The presence of cancer as the greatest health fear or substantially worrying about cancer facilitated intentions to attend, whereas uncomfortable thoughts of cancer did not affect the intention and were a deterrent to actual participation (Vrinten et al., 2015). Consistent with the results of other studies, cancer worry facilitated screening by enhancing the intention to attend, which may be motivated by a desire for reassurance, whereas a more visceral negative response to thinking about cancer acted as a deterrent in the action stage (Clemow et al., 2000;Jandorf et al., 2010;Wong et al., 2013). This deterrent effect is referred to as the "ostrich effect", in which an individual prefers not to obtain information regarding her state of affairs because of the fear that she may receive bad news, despite the prospect of making better decisions based on this information (Panidi, 2014).

Lack of comprehension:
It has been observed a lack of comprehension of cancer screening in both regular and irregular adherence participants, but it could be minimized by primary healthcare professionals.
Inconsistent or inadequate support from providers: Consistent with other screening tests, the population clearly expected to receive information regarding colorectal screening from their physicians. Our findings reinforce the importance of shared decision-making between providers and consumers. Decision aids to support doctors in their discussions with patients may be valuable, given the need to balance potential risks and benefits, as well as the different perspectives on test quality and acceptability that were expressed (Goel et al., 2004). Primary healthcare professionals can facilitate informed choices by patients who participate in CRC screening, and this role requires health care to have access to relevant, accurate and complete information.
Considering the lack of comprehension and the importance of primary healthcare professional in the decision on sustained participation, it is important to better engaged them with cancer screening programs, and also improve the communication channels. Possible methods for facilitating communication could include continuous briefings, regular message reminders or educational websites. Screening information is currently given to the individuals by a brochure sent by mail. This communication channel is certainly quick and economical but does not guarantee the individuals' good understanding of the benefits and risks of screening. It would be preferable that primary health care professional to have a private interview and to discuss such documents when their patients come in for a medical consultation. General practitioner could increase their involvement in CRC screening if they were more associated with patient information at different stages of screening. Their privileged mode of communication remains the face-to-face consultation with the patient (Papin-Lefebre et al.2017).
Inconsistent or inadequate support from media: The participants also manifested the importance of information related to screening to encourage individuals to take the test regularly. They suggested that it is important to publicize the colorectal cancer screening program in the media. However, most studies highlight the lack of information and media interest in CRC screening, with more focus on the difficulty of discussing CRC screening with other individuals because it is considered a shameful subject (Beeker et al., 2000;Bridou et al., 2013).

Relaxation effect after a negative result in FOBT:
Individuals who had once participated in the screening and had obtained a negative result in the FOBT indicated that they did not participate again because they believed that it was no longer necessary. However, this response is inconsistent with the available evidence because a high rate of consistent participation increased the program sensitivity of the FIT screening (Winawer et al, 1993;Nishihara et al, 2013). Therefore, the effectiveness of a FIT screening program is highly dependent on participation in multiple rounds. Consequently, once-in-a-lifetime participation is not sufficient to prevent colorectal cancer. This false relaxation does not occur in other types of cancer screening, such as breast or cervical cancer, in which women continue to participate after a negative screening test.
Another study has shown that non-adherence is not caused by a difficulty related to performing the test, as the participants generally considered the test to be relatively simple to perform with no particular problems, at least compared with other similar types of medical procedures. This finding confirms the results of several authors who reported that this type of test is convenient and relatively simple to use (Bridou et al., 2013). Although most countries in Europe currently use the immunological test, the reasons for longitudinal adherence to colorectal cancer screening are common for both immunological and guaiac tests.
An individual's experience during the initial screen has been shown to influence her longitudinal adherence to screening (Bulliard et al., 2003;Maxwell et al., 1996). Several studies have shown that individuals who expressed dissatisfaction or negative views regarding their initial screen were more likely to not re-attend (Marshall, 1994;Orton et al., 1991;Peipins et al., 2006). Thus, longitudinal adherence may be a good indicator of satisfaction because individuals who have been satisfied with the process are likely to re-attend during the subsequent round.
Therefore, longitudinal adherence may be a proxy for measuring satisfaction in a feasible manner. Satisfaction with the cancer screening process should be an indicator that is evaluated on a regular basis. However, the determination of user satisfaction incurs significant costs.
Consequently, ascertaining longitudinal adherence as an indicator of satisfaction could reduce costs and increase the feasibility of this indicator. In contrast and according to the current findings, non-adherence may result from not only non-satisfaction with the screening process but also a lack of knowledge, which was one factor identified in this study as a barrier to longitudinal adherence. This study has several limitations. We attempted to minimize several forms of bias by recruiting participants from different areas of the city and including both men and women of different ages and with varying degrees of experience with CRC screening.
The qualitative nature of this study may limit the generalizability of our findings. However, qualitative research is concerned with generating insights that may be useful in different settings because the understanding that is generated is applicable to specific groups of individuals who share characteristics, engage in behaviors or live in circumstances relevant to the phenomenon investigated (Mays and Pope, 1995). This form of generalizability differs from that gained through statistical studies and was ensured by our sample strategy, which focused on reflecting the diversity within the population under study relevant to the research (including differences in prior screening behavior, socio-economic background, age, sex, ease of recruitment and initial FOBT results) rather than aspiring to recruit a representative sample.
We will use the findings from our detailed analysis in this qualitative study to generate a framework to better understand facilitators and barriers that affect decision-making to participate in CRC screening. The results from these types of qualitative studies may be used to develop interventions to increase participation in colorectal cancer screening programs and specifically increase the longitudinal adherence.   3 "I find this to be very good, because it is designed to prevent. And if I do not have anything, then it's great. And if I have something, and they detect it in time, then it is much better (...) you'll be a little afraid, it's a hard pill to swallow, but sometimes it's better to know as soon as possible... Well, sometimes, not always, but knowing the unpleasant news as soon as possible may make it pass better." (Irregular adherent, Area 1, Female)

Lack of comprehension
4 "I think it is an initiative that is well taken, but it is something long-term because I have not received anything for two years, then it is clear in two years many things could happen ... I find that the main problem (...) if this test is performed every six months it would be better than every year, an example, an example ... I find that every two years is too long, it is my opinion... " (Regular adherent, Area 2, Male)

5
"we have all been relaxed about it because none of us knew that the maximum period between analysis and analysis of the colon are two years, we did not know. And then, well, they have done it and that's it ... and it's not like that" (Irregular adherent, Area 1, Male) 6 "we have a greater chance of dying when we are older than when we are young, so surveillance should also be more stringent or more active when we become older than when we are young, so it is a request that I believe would be correct, that well, instead of cutting at 69, then after 69