Abstract

Colorectal cancer (CRC) accounts for 9% of all new cancer cases worldwide and affects over 1 million people each year. Screening can reduce the mortality associated with the disease, yet participation rates are suboptimal. Compliers with CRC screening are less deprived; they have higher education than noncompliers and tend to be male, white and married. Likely reasons for nonparticipation encompass several 'modifiable' factors that could be targeted in interventions aimed at increasing participation rates. Successful intervention strategies include organizational changes, such as increasing access to fecal occult blood test (FOBT) kits, providing reminders to healthcare providers or users about screening opportunities, and educational strategies to improve awareness and attitudes towards CRC screening. Multifactor interventions that target more than one level of the screening process are likely to have larger effects. The biggest challenge for future research will be to reduce inequalities related to socio–economic position and ethnicity in the uptake of screening.

Introduction

This review aims to provide a comprehensive overview of colorectal cancer (CRC) screening programs and uptake rates worldwide, the factors that have been associated with participation and strategies designed to improve uptake rates. Information on uptake rates was collected using a pragmatic approach that included searching electronic databases, emailing named contacts from screening programs across the world and searching websites such as the International Agency for Research on Cancer. Studies that addressed factors associated with participation in CRC screening and interventions to improve uptake were identified using a systematic search of electronic databases (Embase, Medline, Web of Knowledge, Cumulative Index to Nursing and Allied Health Literature [CINAHL], Psyc Info, Cochrane Database of Systematic Reviews [CDSR], and the Health Technology Assessment [HTA] database). Abstracts of all the articles identified were reviewed and relevant studies were obtained.

Colorectal Cancer & Screening

Worldwide, CRC accounts for over 9% of all new cancer cases and is the third most common cause of cancer death. Incidence rates are similar in men and women, but are higher in more developed countries, with over half of all cases occurring in Europe or North America.[1] Screening for CRC offers the opportunity to detect cancer earlier, as well as remove precancerous adenomas and prevent development of the disease. There is strong evidence that screening reduces mortality;[2] a recent meta-analysis of studies evaluating screening using fecal occult blood tests (FOBT; FOBT using the guaiac smear is currently being replaced in many countries by fecal immunochemical tests [FITs][3]) estimated the mortality reduction to be 15%.[4] Four large-scale randomized clinical trials are evaluating flexible sigmoidoscopy (FS) as a screening tool in the UK,[5] Norway,[6] Italy[7] and the USA.[8] Interim results from the Norwegian Colorectal Cancer Prevention (NORCCAP) Trial indicated a trend towards reduced mortality, but reported no effect on incidence after a 7-year follow-up.[9] However, the Telemark trial showed an 80% reduction in CRC incidence rates at the 13-year follow-up,[10] and case–control studies demonstrate reductions in CRC mortality of between 45 and 79%.[11–13] The UK FS Trial is examining the efficacy of a single FS screen offered to asymptomatic individuals aged 55–64 years, and data on the effect on mortality will be available soon. Further results from the NORCCAP trial and the UK FS Trial will help establish the magnitude and duration of benefit from FS screening. There is also evidence for the effectiveness of screening colonoscopy (CS). CS with polypectomy has been shown to reduce the incidence of CRC by 76–90% in cohort studies.[10,14,15] Although CS involves a greater cost – both to the health service and the patient (preparation and sometimes sedation are required) – it is regarded by some as the 'gold standard' screening method and is the preferred procedure following a positive result using other techniques.[3]

Availability & Organization of CRC Screening

CRC screening is available in the USA, Australia and several countries in Europe and Asia. A more limited number of countries offer screening via an organized program in which tests are provided systematically to a specific target group on the basis of an explicit policy about who is eligible for screening, screening modality and frequency. In 2004, the International Colorectal Cancer Screening Network identified 35 organized screening initiatives in 17 countries, ten of which were classified as routine screening programs, although only four of these were national programs at the time.[16] More recent data on activities indicate that 12 countries now offer routine national screening programs (Australia, Cyprus, Czech Republic, England, Finland, France, Israel, Japan, Korea, Poland, Portugal, Scotland, Slovenia [Belgium, Canada, Italy and Sweden run regional CRC screening programs, Bulgaria, Romania, Slovak Republic, and Wales are planning to introduce national programs]).[17] The most common screening modality is FOBT, but several countries offer endoscopy (FS or CS) as a follow-up test for positive FOBTs. In England, the National Health Service (NHS) invites individuals aged 60–69 years (to be extended to 75 years in 2010) to complete a FOBT every 2 years as part of their Bowel Cancer Screening programme (NHS BCSP).[201] Several countries offer opportunistic screening where a test is offered to asymptomatic individuals when they present to a healthcare practitioner for reasons unrelated to CRC (Austria, Germany, Latvia, Luxembourg, Malaysia, New Zealand, Poland, Spain, Switzerland and the USA). For example, the USA promotes CRC screening through the dissemination of guidelines and media campaigns, although some organized programs are run through health plans and local health departments.

CRC Screening Uptake Rates

Higher participation rates in screening are associated with greater screening efficacy in terms of mortality reduction,[4] and can increase its cost–effectiveness – particularly in the case of FOBT screening.[18–20]

Data from routine population-based programs offering FOBT with national coverage in 2004 (Czech Republic, Finland, Israel, Japan and Korea) report uptake levels between 20% (Czech Republic[21]) and 71% (Finland[22]) for FOBT. Feasibility studies and studies reporting on uptake rates in countries undergoing national roll-out have shown rates of around 45% (Australian pilot 45.5%,[202] Italy mid-roll-out 45%,[23] France pilot 42%,[24] England and Scotland 52%[23]). Participation in FS screening in trial or pilot settings is lower: Italy 15%,[23] and UK 39%[25] (see Table 1 for more details).

The National Health Interview Survey (NHIS) is the leading source of health information on the civilian, noninstitutionalized population in the USA, which asks survey respondents whether people have had a screening test, if they are up-to-date with testing, and differentiates between tests carried out for diagnostic and screening purposes. Data from the 2003 survey indicate that uptake of screening tests for CRC are around 26% for either FOBT or endoscopy, but only 18% for endoscopy alone.[26] In more recent reports from the 2005 survey, only 35% report ever having a CRC screening test.[27]

Factors Associated with Uptake

The comparatively low rates of public participation in bowel cancer screening programs have prompted considerable interest in the determinants of screening behavior.

Who Attends Screening?

A number of socio–demographic and economic factors (summarized in Box 1) are associated with participation in screening and help to identify who does and does not attend screening.

Socio–Economic Status

The biggest individual factor predicting participation is socio–economic deprivation, with uptake levels in deprived groups as much as 30% lower than uptake in more affluent groups.[26,28–32] Inability to afford healthcare is a contributory factor in countries such as the USA, but marked socio–economic disparities are still observed in countries where screening is free at the point of delivery (e.g., the UK). For example, uptake of FOBT in the English FOBT pilot fell for each increase in deprivation quintile (Index of Multiple Deprivation, 2004) from 67% in the lowest quintile to 37% in the highest (the trend was significant at p < 0.001),[32] and similar deprivation effects were observed in the UK FS Trial.[33,34] Some evidence suggests that cognitive factors such as attitudes towards CRC screening may explain socio–economic status (SES) differences in uptake.[35] This is an important avenue for future research.

Participation in CRC screening has been consistently associated with higher educational attainment in several large-scale surveys of screening participation, including the 2004 Behavioral Risk Factor Surveillance System (BRFSS),[36] the 2003 Health Information National Trends Survey (HINTS)[36] and the 2003 NHIS.[26] In 2003, among men (n =4647) and women (n = 6778) there was a consistent gradient for education and reported use of home FOBT in the past 12 months. Participation among men without high-school qualification was 12%, compared with 19% for college graduates. The same pattern emerged for women (11 vs 19%), with a gradient across intermediate qualification levels in both sexes. The same pattern was observed for sigmoidoscopy use.[26] An Australian cohort study of 30,853 men and women aged 50 years and over also found that there was a significant trend of increasing likelihood of CRC testing (FOBT, sigmoidoscopy or CS) with increasing education, after adjusting for gender, age and family history of CRC.[37]

Literacy, and more recently health literacy, have been used as variables to explain the impact of education on health behavior in general and screening participation in particular.[38–40] The evidence base describing the role of health literacy is small and inconsistent, because there is no agreement about the definition of health literacy and objective assessment is difficult to implement as part of large-scale surveys. However, as health literacy comprises a relatively well-defined and circumscribed set of skills, it remains an important and useful variable to address as part of health education and promotion strategies to increase CRC screening participation.

Few studies have assessed the impact of employment, and on the whole most have found no significant association.[41–43] However, data from the UK has shown a positive effect with higher attendance at FS screening among working individuals than those who were not working (78 vs 73%).[44] There is also some evidence that retirement (compared with being employed or unemployed) is positively associated with screening participation.[45,46]

Ethnicity

Rates of CRC screening are lower in ethnic minority groups than the White majority in the USA,[36] Australia,[37] and The Netherlands,[47] and this remains true when controlling for income[48,49] or education,[50,51] suggesting that the effect is independent of any association with socio–economic deprivation. In the UK, the south Asian population has been shown to have significantly lower participation despite adjustments for deprivation (individual and area-level deprivation, respectively).[52,53] Evidence from the USA indicates higher burden of CRC among African–Americans,[54] but more generally, ethnic minority groups have been considered to be at lower risk of CRC than the majority population.[54–57] However, if ethnic minority groups are less likely to take part in CRC screening, their CRC mortality rate may begin to increase relative to the majority population, despite their lower risk. Efforts are needed to ensure equality of access across all ethnic groups.

Age

National population-based surveys in the USA have documented an association between older age (>60 years) and participation in CRC screening (BRFSS 2004,[58] NHIS 2000,[50] and NHIS 2003[26]); this was also true in Finland[22] and Norway,[6] but the reverse has been shown in Italy.[59] More detailed analysis reveals that participation rates increase up to around 75 or 80 years and then decrease.[23,37,60,61] Associations with age could be explained by changes in health insurance cover according to age (Medicare only covers members 65 years and older). However, a study by Seeff and colleagues found that age and health insurance were independently associated with CRC screening participation, suggesting that additional variables may explain the relationship.[50] In some organized systems there is an upper age limit to screening that means people would no longer be sent an FOBT kit automatically (or sent an invitation for endoscopy), but would only be able to have screening on request. This shift from external to self-determined screening is likely to have a detrimental impact on uptake and may also send out the implicit message that their risk of getting CRC reduces after a certain age.[62]

Gender

When you consider any CRC screening, the majority of research finds that men are more likely to take part than women,[37,44,50,58] although the opposite was shown in Norway, with women more likely to take part in CRC screening, and this was independent of screening modality.[6] Nevertheless, recent data from the BRFSS indicates that the gender gap in CRC screening may be decreasing (at least in the USA); men were more likely to have taken part in CRC screening in 2002 and 2004, but there was no sex difference in screening participation in 2006.[63]

Marital Status

Married individuals are more likely to take part in CRC screening than unmarried individuals; the 2003 HINTS survey showed that among those aged 65–74 years (n = 398), 73% of married individuals were 'up-to-date' with CRC screening (FOBT in previous year or endoscopy in previous 10 years), compared with 28% who were not married (divorced, separated, widowed or never married).[36] Similar findings were reported in the 2000 and 2003 NHIS (e.g., 46 vs 36%[50]). Positive effects on screening participation of being married have also been noted elsewhere, in Australia,[37] Sweden[64] and in the UK,[65] where it was also shown that inviting couples to attend FS screening together further increased participation.

Facilitators & Barriers to Participation

Nonattendance at CRC screening is associated with health system factors, as well as barriers on the part of physicians and patients.

System-based Factors

Within organized programs, screening is often offered directly to the patient and is free at the point of delivery, and therefore cost and healthcare provider recommendation typically play a less prominent role in determining uptake. In countries where screening is opportunistic, however, access to screening is more determined by factors such as individual insurance coverage and characteristics associated with the healthcare provider. Screening participation rates are reported to be higher in areas with organized screening programs than those where screening is opportunistic,[66] and when screening is provided by a general practitioner compared with a hospital.[67]

Uptake rates also vary according to the type of screening offered. Four identified randomized controlled trials (RCTs) compared uptake for FOBT alone to sigmoidosocpy plus FOBT,[59,68–70] of which two found that offering FOBT plus sigmoidoscopy resulted in lower uptake than FOBT alone[69,70] and one reported that FOBT plus sigmoidoscopy was associated with higher participation.[68] One study reported no difference.[59] In Norway, a comparison of FS with FS plus FOBT revealed higher participation rates for FS alone (67 vs 63%).[6] One trial compared FOBT plus sigmoidoscopy with FOBT plus CS, and both produced similar uptake rates.[71] Similarly, no differences in uptake were found when comparing conventional FOBT kits with more advanced FOBT kits that minimize fecal manipulation (e.g., Haemoccult®, 49.1% vs Early Detector 52.1%; and Coloscreen Self-Test® 50.6%),[72] suggesting that while people may have preferences for either FOBT or endoscopy, there is little difference in uptake within each modality.

Participation in screening has also been shown to be higher if FOBT kits are posted with an invitation compared with an invitation only,[72–75] if they are posted rather than provided by the physician or screening facility,[59] and when they can be returned by post rather than being hand-delivered (57 vs 37%).[76] Uptake rates are higher when advanced notification of screening invitations are provided (48 vs 40%)[77] and when additional invitations are sent, including prefixed appointment times.[59] Zorzi and colleagues reported hugely variable compliance across regional programs in Italy (7 to 78%),[23] which could be a result of cultural differences or variability in the way screening is organized. The Italian Group of Colorectal Cancer Screening (GISCoR) are carrying out a study that involves all the Italian programs, with the aim of identifying organizational factors linked to higher compliance levels.

Health Insurance

Lack of health insurance may be a specific barrier to health services in some health systems, and this is evidenced by significantly lower rates of CRC screening participation in those with no insurance compared with those with some cover.[26,36,37,58] The 2005 NHIS survey showed that 19% of respondents with no insurance reported having CRC screening (FOBT in the past year or endoscopy in the previous 5 or 10 years) compared with over 39% of those who had some sort of insurance (e.g., private, Medicare).[78] Health insurance is also associated with access to healthcare; 54% of uninsured individuals aged 19–64 years had no usual source of care, compared with 10% of those with private or Medicaid insurance.[79] Furthermore, lack of health insurance remains an important barrier to CRC screening participation across all levels of education and ethnic groups and, as a result, people without insurance are much more likely to be diagnosed with advanced-stage cancer.[78,80]

Area-level Factors

There has also been interest in screening participation in rural and urban areas. In the USA, people living in urban areas are more likely to take part in screening than those living in more rural areas.[81–84] Interestingly, studies from Europe and Japan show the opposite trend (e.g.,[64,85]). For example, a Swedish study of a population sample of 1986 residents found that living in the countryside or a smaller community was associated with better participation.[64] In an Australian study, Weber et al. [37] found that compared with participants from major cities, those living in outer regional areas were significantly more likely to report using FOBT, although the pattern was reversed for the use of other colorectal tests. Owing to the diversity of CRC screening modalities (which feature both hospital- and home-based tests) it is unlikely that there will be a single or simple answer as to what and how these types of environmental factors impact participation.

Provider-level Factors

When screening is opportunistic, receiving a recommendation to have screening from a doctor is consistently identified as a key predictor of participation.[43,45,86,87] Having a primary care physician or a usual source of care (compared with not having one) are associated with higher participation rates in CRC screening in the USA.[26,36,88] Having more than one physician is also positively associated with CRC screening participation; in an analysis of BRFSS 2004 data, Cardarelli and colleagues found that those who reported one or more personal healthcare providers were more than four-times as likely to be current with CRC screening guidelines even after controlling for age, gender, income, education and health insurance.[58] A higher frequency of visits to the physician is also linked to higher uptake; in the 2000 NHIS, 51% of respondents who had visited the doctor more than six times in the previous year reported being current with screening, participation was 45% in those who had visited between two and five times, 36% in people who visited once and only 15% in those who had not visited at all.[50]

Further support for the importance of the physicians role comes from studies that show higher screening participation in individuals who report greater satisfaction with their physician,[89] or who report their doctor as their main source of information.[87] Better doctor–patient communication is also associated with increased attendance at CRC screening.[79,90,91] For example, patients who believed their physician showed them respect, listened carefully, spent enough time with them and explained things sufficiently were more likely to have participated in CRC screening.[79]

Health Factors

One of the most prominent self-reported barriers of CRC screening is good health and a lack of bowel symptoms,[92–100] yet general perceptions of health are not consistently associated with participation in CRC screening when measured in quantitative surveys. For example, the HINTS 2003 data in the USA reported no association,[36] as did surveys carried out in Sweden[64] and Australia.[37] In contrast, data from the UK FS trial showed a positive association between good health and attendance at FS screening,[34,44] as did the BRFSS in 1999, although only among men.[101] In a review of determinants of FOBT completion, Neilson and Whynes found that individuals who had more personal experience of illness were more likely to be compliant.[30] Furthermore, there is some evidence of an association between having a chronic health condition and participation in screening.[101–103] Inconsistent findings may be a result of differences in the way screening is accessed, because people with poorer health may have increased exposure to screening opportunities.

Unsurprisingly, previous participation in other types of cancer screening (e.g., prostate or breast cancer screening) is associated with taking part in CRC screening.[37,44,45,50,88,104] For example, in a random digit-dialed survey of 1033 people aged 50–79 years in the USA, 69% of those who reported frequently or almost always participated in other cancer screening tests were up-to-date with CRC screening, compared with 46% who reported never, rarely or only sometimes taking part.[88]

There is also evidence to support a positive association between other healthy behaviors and screening participation; visiting the dentist more regularly,[44] being physically active,[37,44,45,50] eating fruit and vegetables[37,50] and abstaining from unhealthy behaviors such as smoking.[37,45,50]

Psychosocial Factors

Awareness & Knowledge of CRC & Screening Better knowledge about CRC and screening is related to higher participation[36,42,105–107] and is particularly important in areas where screening is opportunistic. In the HINTS 2003 survey, the majority of individuals (83%) who were up-to-date with CRC screening knew there were one or more tests for CRC, compared with less than half of adults who were not current with screening (44%). There are also differences in awareness of the recommended intervals for CRC screening; 60% of those current with screening guidelines knew the correct interval for FOBT screening, compared with 45% who were not up-to-date. Awareness of the guidelines for endoscopy screening was poor, with only 14% of those up-to-date reporting the correct interval, and 5% among those who were not current with screening.[36] This is consistent with qualitative evidence in which lack of knowledge or awareness of CRC and screening is frequently cited as a barrier to participation.[89,93,98,100,108–113] However, this could be a consequence of previous participation in screening and may not be a causal factor.

Higher awareness has been linked to various socio–demographic variables, including white ethnic background,[100,114] speaking English as a first language,[115] female gender, older age and higher education.[115,116] Awareness is also higher in adults that report a close relationship with their provider and in those who have health insurance.[117] Therefore, it is possible that knowledge could mediate some of the demographic effects associated with uptake.

Perceptions of Risk Higher perceptions of risk have been associated with screening participation using absolute risk of CRC in the future or in a specific period of time (e.g.,[118]) and comparative risk (risk of CRC compared with others of the same age and gender) (e.g.,[44]), but several studies report nonsignificant associations.[119] Null findings could be related to measurement issues because a recent meta-analysis of health beliefs and vaccination behavior found that the risk-perception–behavior relationships were largely influenced by the design and quality of the study, and measurement skew.[120] For instance, previous screening behavior could contaminate perceptions of risk such that those individuals who had previously had screening felt less at risk from CRC. There is evidence that conditioning the risk question to attendance at screening increases the accuracy of the measure; for example, asking 'If you attend screening, what do you think your chances of developing CRC are?'.[121] It is also possible that the impact of perceived risk is mediated by beliefs about the perceived benefits and barriers of screening.[122,123]

Few studies have addressed the association between perceived severity of CRC and participation, but of those that have there is no consistent relationship. In the USA, a cross-sectional survey of older women (n = 280) noted a positive association,[124] yet two prospective studies of FS screening (carried out in The Netherlands and the USA) found no relationship.[91,125] Inconsistent patterns might be a result of variability in the conceptualization and measurement of perceived severity.[119]

Attitudes Attitudes towards screening are consistently associated with CRC screening behavior. For example, beliefs about screening efficacy (screening can prevent or detect CRC early) are positively associated with taking part in screening,[36,117,118,126,127] as are perceptions that screening will provide peace of mind[44,124] and is safe.[88,124] Perceptions of the negative consequences (e.g., fear of cancer), as well as practical barriers such as getting time off work, are associated with lower attendance. There is a consistent negative effect for perceived embarrassment,[91,117,118,124,128,129] discomfort,[91,117,124,128,129] fear of screening,[130–132] lack of symptoms[87,123] and worries about cost,[36,44,118,124,128] time[44,118,124,128] and screening location.[118,128] Concerns about bowel preparation procedures before endoscopy screening are also cited as barriers to screening participation.[94,108,113,124,128,133,134] Overall, endorsing more 'pros' than 'cons' of CRC screening is related to higher screening participation.[123,135]

Fatalistic attitudes, for example, 'I would rather not know' and 'I would rather let nature take its course', have also been identified as important reasons for nonparticipation in CRC screening.[119] Higher fatalism was a significant predictor of nonattendance at FS screening in the UK,[34] but overall, few studies have assessed the link between these types of avoidant attitudes and CRC screening behavior.

Fear/Worry of CRC Studies examining the effect of fears or worry about CRC have reported contradictory findings demonstrating both a positive[36,91,135] and negative[125,126,128] association between higher fear or worry and participation in CRC screening. Moderate levels of worry have shown a positive effect; participants in the UK FS Trial who indicated they were 'a bit worried' were significantly more likely to attend FS screening than those who were 'not worried at all' or 'quite' or 'very worried'.[44] Four studies that used a prospective design[44,91,125,126] showed that fear or worry was a significant predictor of subsequent screening participation, and in three of the four studies findings indicated that a lower level of fear or worry was a facilitator.[44,125,126] This suggests that being less fearful or worried about CRC is more often predictive of screening participation rather than nonparticipation, although findings are inconsistent.