Australia's National Bowel Cancer Screening Program: does it work for Indigenous Australians?

Although bowel cancer is rarely noted as a significant health issues for Indigenous Australians, it is the third most common cancer after lung and breast in women, and lung and prostate in men, accounting for about 10% and 9% of all cancers respectively [13]. However, the lack of national data has meant that little information is available surrounding the impact of bowel cancer on Indigenous Australians. The Australian Institute of Health and Welfare (AIHW) reports that incidence and deaths from bowel cancer are much lower than the non-Indigenous population [13], although the true magnitude of bowel cancer is not clear due to problems associated with under-ascertainment of Indigenous status in cancer data collection registries. The Australian Government recognizes three criteria as determining Aboriginal and Torres Strait Islander status: - Indigenous descent; self-identification as an Indigenous person and acceptance of the individual as Indigenous by the Indigenous community [14]. Administrative data collections generally focus on self-identification to classify an individual as Indigenous, and while ascertainment of Indigenous status is improving over time in Australia, under ascertainment is an acknowledged problem with gaps in the data that affect the accuracy of statistics recorded [14]. Consequently, Indigenous cancers are thought to be underestimated by up to 20% and mortality up to 35% [13, 15, 16].

The observed low incidence and mortality from bowel cancer may be influenced by a number of other factors, including the high rate of incomplete or inadequate death certification and the large proportion of cancers of undefined primary site in registered Indigenous cancer records [13, 17]. Additionally, as cancer risk increases with advancing age, the substantially lower life expectancy of Indigenous Australians means they may die from other chronic conditions prior to cancer diagnosis [18]. Lower participation in cancer screening and the possibility that Indigenous people with bowel cancer may avoid seeking medical care for symptoms or accessing treatment could also reduce ascertainment of disease [19‐22].

Age-standardized incidence rates from 2002-2004 suggest that the rate of bowel cancer in Indigenous Australians is about half that of the non-Indigenous population (39.7 vs 76.4 per 100 000 in males and 36.6 vs 52.4 in females; rate ratios are 0.5 for males and 0.7 for females) [13]. Incidence data collected from cancer registries (2002-2006) in three Australian states which have the most reliable data collections gave a slightly lower overall age standardized rate of 29.9 per 100 000 among Indigenous people compared to 64.7 per 100 000 in non-Indigenous [16]. However, these figures mask higher rates that exist in younger Indigenous Australians. Age-specific rates show that incidence was similar in non-Indigenous and Indigenous people up until age 50. Notably, men in the 40-49 year age group were found to have higher rates of rectal cancer than non-Indigenous men, attributed to greater consumption of alcohol. Supporting the increased burden at a younger age are findings from the Northern Territory (NT) indicating that incidence rate ratios were higher in Indigenous Australians under 64 years compared to those over 64 years (0.6 vs 0.2; 1991-2001) [3, 4].

Colorectal cancer (CRC) mortality estimates vary in different jurisdictions. In a NT study (1987-95), age standardised mortality rates in Indigenous people were about half that of non-Indigenous people [22], while in rural and remote Queensland and in New South Wales mortality rates were not significantly different between Indigenous and non-Indigenous people [15, 23]. A more recent NT analysis showed mortality rates were similar in Indigenous compared to non-Indigenous people under 65 years, but substantially lower in older ages, suggesting an interaction between age and Indigenous status [24]. This was further supported by recent results from four Australian states, where despite somewhat lower age standardised death rates in Indigenous people (17.9 per 100 000 compared to 19.8 in the non-Indigenous population) age-specific mortality rates were actually higher in Indigenous people under 60 years compared to the non-Indigenous population [25].

Indigenous Australians generally have higher mortality following a cancer diagnosis compared to non-Indigenous Australians, especially for those cancers which are amenable to early diagnosis and treatment [22, 26]. Later staging at diagnosis means patients present with more advanced disease and experience poorer treatment outcomes; Indigenous people also appear less likely to undergo and comply with treatment [19, 21]. This has also been shown to be true for CRC [21]. Thus, once diagnosed, Indigenous people are more likely to die from CRC compared to their non-Indigenous counterparts as reflected in the poorer five year survival rate. Condon et al. [21] found that nearly 90% of Indigenous people with CRC in the NT presented with advanced disease at diagnosis as opposed to 68% of the non-Indigenous group, and that five year survival was only 31% compared to 55% in the non-Indigenous group [4]. This pattern is consistent with CRC mortality and survival observed in Maori, Native Americans, Native Hawaiians, First Nations and Inuit [6, 27‐31].

Reasons behind late presentation and delayed diagnosis have been explored in studies examining breast and cervical cancer in Indigenous people and include a low awareness of symptoms, delay in seeking medical advice, reduced access to and low quality primary care, diagnostic and specialist services, and a disinclination to seek advice for symptoms because of certain beliefs regarding cancer and the chance of successful treatment [4, 32]. Late presentation and poor survival is also influenced by communication and cultural barriers which may impact the effectiveness and choice of treatment, as well as the decision to engage in preventive and diagnostic activities including cancer screening [19, 20, 33]. The presence of other chronic diseases or environmental conditions also limits provision of treatment due to the risk of complications [4, 21, 26, 34].

National cancer screening programs can partly address the survival discrepancy in Indigenous Australians by facilitating early diagnosis. If implemented in a culturally sensitive manner, cancer screening can aid in reducing mortality from the disease [26]. This has been demonstrated in the NT and WA where cervical cancer mortality declined substantially following the provision of a culturally appropriate holistic cancer screening services [35‐37]. However, Australia's national bowel cancer screening program in its current form has not taken such an approach, reflected in the very low uptake by the Indigenous population.

Bowel cancer screening is an effective means of reducing the mortality and burden of CRC in the community through the early detection of abnormal changes in the bowel [38, 39]. It is an ideal malignancy to target for population screening and fits the World Health Organisation's criteria for suitability for population based screening [38]. Importantly, it has an identifiable early precursor lesion, usually in the form of an adenomatous polyp which in most cases is asymptomatic; it has a long latency period (for a polyp to become cancerous), thus providing an opportunity for early diagnosis; and there is a high cure rate if found with available effective treatment. Five year survival rates can be almost 90% if a tumour is detected in the earliest stages (Stage A) while still asymptomatic and localized. Once cancer has proceeded through the bowel wall, to the lymph nodes and beyond, five year survival is reduced to only 7% [38, 40]. In South Australia (SA), a monitoring system showed that only 15% of bowel cancers were detected at Stage A indicating that substantial potential improvements in survival could be achieved with earlier diagnosis [41].

Australia is one of a few countries implementing a formal, government funded, population-based CRC screening program - others include the UK, Canada, France, Italy and Finland [42‐44]. Following a bowel cancer screening pilot program in 2002-2004 which examined the feasibility, acceptability and cost-effectiveness of the program in the Australian community, the first phase of the National program was rolled out in 2006 and was offered to people turning 55 and 65 years of age. The second phase began in July 2008 and expanded eligibility to all men and women turning 50, 55 or 65. The program targets only specific ages due to limited funding and to ensure that health services can cope with the increased service demands [9]. The target population age range was chosen based on the National Health and Medical Research Council guidelines which indicated the risk of developing bowel cancer increases significantly over the age of 50 years, with over 90% of those diagnosed with CRC in Australia aged over 50 [25].

Funding for the NBCSP has been secured only until 2011, allowing for once-off testing of the population - despite the fact that the mortality benefit associated with screening depends on regular participation [45]. The program is freely available to people aged 50, 55 and 65 years only and no plans have been announced for second yearly screening. If fully implemented to a broader age range and with ongoing screening rounds, the program could potentially save up to 2,000 lives per year in Australia [46]. Furthermore, sequential screening can also provide an estimation of incidence (new disease) since last screen, as opposed to the initial screening round which cannot differentiate between new and old cases.

The Bowel Cancer Screening Implementation Committee chose the faecal occult blood test (FOBT), a simple non-invasive test that detects small amounts of blood in the bowel motion, as the test of choice for population screening in Australia. The presence of blood in faeces is a sign that abnormalities such as a precancerous lesion or cancer may be present in the bowel. Landmark trials have shown that regular population screening using FOBT can reduce the risk of death from CRC by up to 33% in those who (regularly) participate [39, 45, 47]. FOBT is the only screening test for CRC where evidence from RCTs demonstrated a reduction in mortality [48], although the trials used the guaiac FOBT (gFOBT) as opposed to immunochemical FOBTs (iFOBT), the test used in Australia. One recently completed RCT using the iFOBT demonstrated increased detection and compliance, although long-term mortality benefits have yet to be demonstrated [49].

The screening pathway is shown in Figure 1. Screening test kits are mailed to individuals following identification of eligible program participants from enrolment in the national healthcare insurance scheme (Medicare). Kits are completed at home by taking samples of two consecutive bowel motions and mailed back to a pathology laboratory for testing without cost to the patient. Results are sent back to participants and their nominated GP (if one has been nominated). For positive tests, a letter advises participants to visit their GP to discuss a follow-up procedure, namely colonoscopy. A bowel cancer screening register has been established within Medicare to assist in follow-up of those with positive tests not undergoing colonoscopies, rescreening of participants, and to aid data collection, monitoring and evaluation [9]. The data obtained are collected via forms filled in by the individual, their GP, colonoscopist and others involved. As completion of the forms is not compulsory, a large amount of information is incomplete or missing [25].

The sensitivity and specificity of FOBT varies by test type and manufacturer. Sensitivity is in the order of 60-85% for iFOBTs when used regularly within a program, although with once-off testing this drops to 30-50% [48]. The iFOBT (MagStream 1000/Hem SP) currently used in the Australian program has a sensitivity of 66% for cancer and 20% for polyps over 1 cm [50‐52]. Hence, many false positives occur, resulting in high numbers of follow-up colonoscopies. Of those who have a positive FOBT, about 5-10% will have CRC, and 20-35% will have an adenoma identified [41]. Test outcomes and the ability of the test to identify CRC are not usually fully understood or made apparent to the public, yet are important when weighing up the risks and benefits associated with participation in the program [53].

In 2009, nearly half a million testing kits were recalled and the program was temporarily suspended for six months after an unusual decline in the number of positive test results, assumed to be due to faulty kits. Testing resumed in November 2009 with a new kit that was initially distributed to re-screen those who received the faulty kits [54]. However, the new kit is recommended for use only up to 30 degrees Celsius creating potential problems for many regions, particularly in the north of Australia where a large proportion of Indigenous people reside. The new screening kit will not be sent to these regions initially, creating even further inequities in access. Much greater thought and effort needs to be put into the problem of differential access to the program.

Adequate levels of participation are essential for the success of a screening program in reducing mortality from CRC and for population benefit equating to that attained in randomized controlled trials [8]. It is also an important contributor to the cost-effectiveness of the program in terms of number of lives saved and costs saved by averting a cancer diagnosis [8, 55]. Monitoring reports completed in 2007, 2008 and 2009 details performance of the program from 2006-2008 [10, 25, 56].

Overall participation in the Australian NBCSP was 37.6% by the end of 2008 however, participation was lower than that observed during the pilot (45.4%). The average positivity rate was 6.6%, significantly higher among men than women which is reflected in higher incidence rates [25]. Table 1 details variation in participation, positivity rate and proportion of correctly completed tests according to demographic factors.

Analyses up until mid-2008 showed Indigenous participation was 17% - less than half the non-Indigenous rate [10]. In the 2009 monitoring report which documented results from 2008 only, Indigenous people were 2.2 times less likely than non-Indigenous to participate in the screening program [25]. However, Indigenous identification of participants in the program was through self-report and was completed by only 63% of participants. Another important finding from the 2008 report was that Indigenous people were significantly less likely to complete the FOBT correctly, with Indigenous females aged over 65 years having the lowest rate of correctly completed tests [10]. Indigenous people also had higher FOBT positivity rates although this was not significant due to the small sample size [25]. The proportion of correctly completed tests was also significantly lower for people who nominated a language other than English as their preferred language [10]. Significant differences were also observed according to geographic location, with inner regional areas having the highest rates of participation while those in remote and very remote areas (where 26% of Indigenous Australians reside) were significantly less likely to participate. Participation also fell with lower SES [10], and those in the most disadvantaged lowest quintile of the population were significantly less likely to follow up a positive result with their GP. Indigenous Australians are disproportionately represented in this quintile.

Substantial numbers of unreturned health provider forms make it impossible to determine how many Indigenous (or other) people with positive FOBTs went on to complete a colonoscopy. For the entire program, only 42% of participants with a positive FOBT had a colonoscopy recorded. The remaining participants either did not have a colonoscopy or the colonoscopy report was not returned [25]. Lack of quality surveillance information compromises the assessment of actual population benefits and the impact on the incidence and mortality from CRC, particularly on population subgroups. Despite this, a recent analysis by Ananda et al [57] assessing the initial impact of the NBCSP from mid 2006 to mid-2008, reported its success in identifying CRC at an earlier stage. This study utilised linked data from hospitals across Australia to gather information on CRC cases reported by surgeons and found that 40% of cancers identified through the NBCSP (N = 40) were at Stage 1- the most treatable stage, compared to 14% among individuals who presented with symptoms to their doctor [57].

An individual's decision to engage in CRC screening is complex and influenced by socio-cultural, historical, system and organizational factors, and individual behavioural factors - all of which impact on screening uptake at a population level [8, 58]. Unfortunately, the design of screening programs often does not reflect the diversity of social and cultural values of a nation's population, creating significant barriers to participation in CRC screening.

Bowel cancer screening using the FOBT differs quite considerably from other screening programs in its complexity, and is unique in many ways which makes it a difficult program to promote. A self-administered test requires that one complete it at home and return it, hence relying on individuals to take initiative and follow instructions. Moreover, people may find it an unpleasant test to complete, and the first faecal sample must be stored appropriately until the other is taken. CRC screening targets both males and females, which not only increases the size of the target population, but men's lack of previous involvement in screening programs poses additional challenges for recruitment. There is also some confusion surrounding the age groups for screening with inadequate understanding by and explanation to participants as to why only certain people are chosen. As with most screening tests the FOBT is not diagnostic and only identifies those at high risk, so further investigation is necessary for a definitive diagnosis to be made, making the process longer and more stressful for the client. Thus, participation must be addressed along the entire screening pathway - at the screening stage, if further investigation with colonoscopy is required, then again when treatment is required. Where ongoing screening is incorporated into the program, difficulties arise getting people to return for subsequent two-yearly screens.

Lower screening uptake by certain racial/ethnic and minority groups has consistently been documented in countries implementing CRC screening programs [59‐62]. Lower participation also occurs among the less educated, lower income groups, and those from non-English speaking backgrounds [10, 58, 63, 64]. Studies controlling for SES and other factors such as distance from services and education show that race and ethnicity are independent predictors of CRC screening uptake [28, 65].

Research reveals that beliefs and attitudes of ethnic and minority groups in regard to cancer and cancer screening can differ quite significantly from the majority mainstream population which may have important implications for participation in cancer prevention efforts and treatment [17]. A number of studies in the USA have identified differing views and beliefs that could act as motivators or barriers to CRC screening [66‐68]. A study of American Indian and Alaska Natives (AI/AN) found that having a family history of cancer, higher education and income, presence of other chronic conditions, speaking English at home, urban residence, and older age were predictors of CRC screening participation [69]. Researchers examining other ethnic groups have identified that being physically active, having a greater perceived susceptibility to CRC, a doctor's recommendation and participation in other screening programs as important predictors of screening adherence [70], while lack of knowledge, lower perceived risk and perceived self-efficacy, inaccurate beliefs, language and communication barriers, fatalism, the belief that screening is not necessary without symptoms, lower levels of utilisation of health services, the failure of a doctor to recommend screening and cost are significant barriers to screening [66, 67, 71, 72].

Many other developed countries with Indigenous populations are in the beginning stages of CRC screening and while uptake by Indigenous populations has not been extensively studied, early data also indicate lower participation. Table 2 shows differences in CRC screening participation by Indigenous compared to the non-Indigenous population in four industrialised countries, outlining variations in screening recommendations, the type of screening test and how the screening is targeting individuals.

Cancer screening programs have had variable success among Indigenous people in Australia. The low participation rates observed for the NBCSP reflect similarly poor engagement of Indigenous women in cervical and breast screening programs, with participation in the order of 30-35% compared to 55-60% in non-Indigenous women [3, 37, 73]. Indigenous communities face several barriers to accessing cancer-related health services, including early detection programs [74, 75]. Evidence from studies assessing breast and cervical screening uptake by Indigenous women in Australia suggest that participation is influenced by multiple structural, socio-cultural and behavioural factors similar to those described above (see Table 3), including poor knowledge and awareness of cancer and the benefits of screening prevention services, culturally insensitive services, shame and fear regarding cancer, lack of Indigenous or female service providers, distrust of the mainstream providers, low literacy, poor coordination of services as well as distance and access barriers [3, 17, 76‐78]. The absence of sufficient Indigenous-specific, culturally relevant, educational resources has also been identified as an important factor impeding cancer prevention efforts [74, 79‐81]. To date, no published studies in Australia explore barriers specifically towards bowel cancer screening in Indigenous people other than an evaluation conducted following the pilot program in 2004 which interviewed only 15 people - 8 participants and 7 non-participants [82].

However, a study of Italian migrants in Australia found similar issues and beliefs to Indigenous Australians - fatalism, English proficiency, fear of cancer and finding out they have cancer, inadequate knowledge and misconceptions of causes of bowel cancer [83]. Weber [84] also explored uptake of cancer screening in Australian migrants and found a significantly lower participation rate in bowel cancer screening among Asian men, although as the authors noted, the NBCSP had not begun in Australia at the time of the survey [84]. This highlights the importance of identifying the needs and preferences of minority groups so as to appropriately target them for bowel cancer screening. To successfully target messages, educational resources and interventions for increasing uptake requires understanding of the current knowledge, attitudes and beliefs of the community of interest.

Potentially, screening could exacerbate inequalities in health as a result of less participation by those of lower SES or ethnic minority groups, including Indigenous populations [8]. A recently completed study in S.A has shown disparities in participation in the NBCSP according to SES, gender and age [85]. Although many countries are examining ways of increasing participation by disadvantaged groups, only N.Z has taken a serious interest in avoiding increasing health inequalities before implementing any sort of CRC screening program that may negatively impact on the health of the Maori population [86]. A CRC screening advisory group in N.Z recommended that any screening program if implemented must: adopt a holistic approach to health and wellbeing; establish partnerships with relevant Maori health organisations; incorporate Maori language in information and health promotion materials; and ensure participation of Maori in the planning and dissemination of the program [87]. Such an inclusive approach with consultation and involvement of the community is likely to have a positive impact on participation and should be explored further in Australia.

A qualitative evaluation undertaken in 2005 of opinions, attitudes and behaviours influencing participation or non-participation in the pilot bowel cancer screening program is the only available review which provides some insight into Indigenous people's view of the bowel cancer screening program, its associated screening kit, and reasons behind their decision to participate or not [82]. However, only a very small number of Indigenous people were involved (8 participants and 7 non-participants) and the target population of Indigenous people was not considered representative, so the findings need to be interpreted with caution.

The main issues that emerged were that lack of awareness and knowledge of bowel cancer and related screening options as key barriers to participation. There was a lack of understanding of bowel cancer and the benefits of screening and early detection, with many unable to distinguish bowel cancer from other cancers. Not having any symptoms and no family history of cancer were reasons given for non-participation. In addition, there were doubts about the efficacy of the test and some had concerns about physically collecting the stool samples. Rejection of the kit was often the result of anxiety and confusion. The analysis showed those who completed the kit required much encouragement and support from health workers and GPs as they believed that they did not have the ability to correctly complete the test. Indigenous participants did not discuss the test with family members nor did they actively seek assistance from health professionals [82]. Indigenous people that participated were perceived to be more proactive, in control of their health, and more literate. Participation was influenced by personal experiences such as knowing someone with CRC, degree of community engagement and access to information in a known language. Participants in the pilot tended to have a better understanding of why they were receiving the kit and the benefits of the screening as they were more likely to have seen promotions and media campaigns prior to receiving the kit and so had an idea of what to expect and were not surprised when they received it. This was an important trigger for participation, as was contact and encouragement from community health workers, the fact the test was free, and a sense of privilege felt for being chosen for participation. However, this latter factor also instigated fear and anxiety in some people who felt that they were at risk because they were targeted individually [82].

Additional barriers identified were similar to other screening programs and included a fear of cancer and the belief that cancer essentially means death. This finding is consistent with research conducted into the perceptions of cancer among Indigenous people [88]. There is often a social stigma attached to cancer among Indigenous people [17] which may prevent people from utilizing screening.

Bowel cancer screening has not had a strong advertising campaign to increase public education and awareness, so that many people receive the test without being aware of its purpose. Poor knowledge and understanding of the program was highlighted as a major barrier to participation in the pilot [82]. A study [89] assessing television coverage of CRC concluded that the disease is under-reported in the media relative to other cancers, despite incidence and mortality rates being high. The absence of celebrities speaking publicly of their CRC experience was also noted, with publicity of affected celebrities highly effective in influencing the public's perception and increasing uptake of screening [90]. The 2005 evaluation report echoed this need for more extensive media coverage as a means to target hard-to-reach populations [82]. Unfortunately, the program has attracted some negative publicity since the incident with faulty kits which has not helped the case for promoting CRC screening.

Several features in the way the NBCSP operates, including downstream activities of follow-up and treatment, make access for Indigenous people particularly challenging. Many of these issues were highlighted in the 2005 evaluation report [91], yet to date no changes have been made to the program. Outlined below and summarised in Table 4 are specific elements of the program related to the way it is organised and disseminated that could contribute to unintentionally discouraging participation by Indigenous Australians. Particular socio-cultural barriers are also explored in relation to how they impact on program participation.

In recent years there has been a proliferation of literature investigating CRC screening in minority and underserved groups, identifying barriers to screening and examining how program design can be improved to reduce disparities in cancer outcomes. Intervention studies for improving CRC screening uptake and adherence have been undertaken in the USA. However, the number of studies targeting Indigenous people is minimal with the majority focusing on other ethnic minority and low literacy groups including African American, Latino and Chinese immigrants. However, these studies provide some information on what may or may not work in considering implementation in Australia.

The approaches taken have included education and educational sessions, community based interventions, the use of lay health workers, storytelling, mailed interventions, mail-out reminder systems, telephone interviews, and more recently the utilisation of patient navigators and electronic/educational decision aids [106‐110]. The majority are tailored to the characteristics of the target group, incorporating their own languages and motifs. A review by Vernon [111] found that those incorporating intensive and personal follow-up appeared to increase adherence more successfully [111]. A more recent systematic review of interventions identified organisational level changes as being most effective at enhancing screening behaviour. Changes included establishing separate clinics for screening and incorporating non-physicians into the screening process [106, 112]. Table 5 summarises these intervention studies. The range of interventions trialed highlights difficulties in making comparisons because of differences in outcome measures, target populations, and the type of screening test used. As many of the interventions were carried out in the USA, outcome measures were usually uptake of any type of CRC screening. Interventions to ensure that those with a positive FOBT follow through with colonoscopy are lacking. Patient navigators could play a critical role, helping individuals to overcome many of these barriers in obtaining quality cancer care [113]. Initial evaluation of a small Indigenous women's cancer support group established in regional Western Australia that helped to encourage screening, support families of those with cancer and help patients through their cancer journey could potentially be an effective model for facilitating Indigenous participation in bowel cancer screening while also providing assistance with follow-up tests and treatment [114]. Such community-based and community-led groups could make a significant contribution and further evaluation of their influence on cancer screening behaviours would provide the evidence needed for this.