Breast and bowel cancer screening uptake patterns over 15 years for UK south Asian ethnic minority populations, corrected for differences in socio-demographic characteristics

Breast cancer is the most common cancer in women in the United Kingdom (UK); bowel cancer is the second most common, and the third most common cancer in men after prostate and lung cancer [1]. Five year survival figures for bowel cancer are on average below 50% with poor survival largely attributable to late detection of the disease [2]. In contrast, five year survival for women diagnosed with breast cancer is 76% [3] and for cancers detected by screening 93% [4]. A population screening programme for breast cancer, first introduced nationally in 1988, is now well established in the UK.

Mammography uptake is approximately 71% at first invitation, with values ranging from 55% in London to 76% in other regions [5]. In 2000, the Department of Health set up pilot site areas to assess the feasibility of introducing a population screening programme for bowel cancer using faecal occult blood test (FOBT) kits completed at home and returned to a laboratory for processing [6]. An overall uptake level of 60% was reported in the pilot [7].

A number of international studies have highlighted inequalities in access to cancer screening for people from black and minority ethnic communities. In the UK, low uptake levels for breast screening have been observed, especially among South Asians [8–10]. In the United States (US), low breast screening rates have been reported for African-American women [11]. Although differences have narrowed over time [12–14], rates for Hispanic women appear to have remained lower than those for other groups [15]. For bowel cancer screening, the literature also provides evidence of lower uptake by ethnic minority communities. In the US, differences have been reported both for established ethnic minority groups such as African Americans [16–19] and for more recent immigrant Asian populations including Koreans, Japanese, Chinese and South-East Asians [20–22]. Low levels of bowel cancer screening have also been reported for immigrant populations in Europe [23]. In the UK, the impact of population diversity on uptake of cancer screening was not assessed in randomised controlled trials of FOBT screening [24] or flexible sigmoidoscopy screening [25]. Studies which compare breast and bowel cancer screening behaviour in the same population are limited; US research has reported that uptake is far lower for bowel screening than for breast screening among African American and Hispanic women [15]. Unfortunately, most studies fail to correct observed uptake differences for deprivation, so the possibility of socio-economic status acting as a confounding factor cannot be ruled out [26]. One study which has addressed this issue found that deprivation could not fully explain observed differences in uptake for bowel cancer screening, especially for older US ethnic minority populations [15].

The present study has analysed uptake patterns in a common UK population for two cancer screening programmes over time: breast screening (beginning with round 1 in 1989) and bowel cancer screening which started in 2000. Uptake patterns for South Asian minority groups have been compared to those for the majority population, adjusted for differences in demographics and socio-economic status. The research was funded by the National Health Service (NHS) Cancer Screening Programmes and the project was awarded Coventry Research Ethics Committee approval (ref: 05/Q2802/2) on 27th January 2005.

In the UK, various policy documents [39, 40] and the legal requirements set out in the Race Relations (Amendment) Act 2000 [41] have all called for a change from the concept of 'average citizen' to one that recognises diversity. A recent review of 'ethnic issues in screening' produced by the NHS National Screening Committee [42] has recommended that all nationally managed screening programmes ensure effective demographic data collection (including data on ethnicity) and assess the likely impact of any new policy on the promotion of race equality.

The findings presented in this paper are important because they represent the first body of systematic research evidence based on individual patients in different South Asian populations. This provides robust evidence of disparities in screening uptake for existing and new cancer screening programmes even after adjusting for differences in demographics and socio-economic status. With nearly one in ten UK citizens from a minority ethnic background, improved understanding of how the offer of cancer screening is received by different populations is an increasingly important issue [43, 44]. Our findings support evidence from earlier analyses based on practice-level data, indicating that cancer screening uptake might be significantly lower in certain ethnic minority populations such as South Asians [8].

Our analysis over screening rounds shows that breast screening uptake has improved for the South Asian population at a faster rate than for the majority population, with the net result that differences between the two populations have reduced significantly from 14.6% in 1989–92 to 10.9% in 2003–05, although a significant disparity remains. Our findings are similar to those from the US where differences in mammography uptake are reported to have narrowed over time [12–14]; this decrease has occurred against the background of a general increase in breast screening uptake rates for eligible women [45], similar to the pattern we identify, and of lower breast screening uptakes by women of African-American, Hispanic and Native-American origin [11, 46–51]. The difference we report is however smaller than the 78% vs. 53% disparity reported by another 2001 UK study [52], although this study was undertaken in an area with 61% Muslims in the South Asian population [53]. At the subgroup level, our findings demonstrate that the Hindu-Gujarati population achieved parity with the majority population over the period up to 2005, while Muslim women are the only group for whom breast screening uptake has not improved over time. A slight increase in abnormal mammograms was observed for Muslims over time (2.7% to 4.2% in rounds 1 and 5 respectively). It is interesting to note that uptake rates for Hispanic women also appear to have remained lower than for other groups such as African-American or Caucasian women in the US [15]. Although our findings indicate that breast screening uptake has improved for all age groups over time in both populations, South Asian women, especially older women, continue to exhibit a significantly lower uptake. In the US, mammography uptake is similarly reported to fall with age; the regularity with which women undertake screening also varies, with older women, African-Americans and those from more deprived populations showing more infrequent use [11, 12]. In our study a slight increase in abnormal mammograms was observed for Muslims over time which might be linked to less frequent screening use.

Our findings indicate a general pattern of lower breast screening uptake in more deprived groups for both South Asian and non-South Asian populations, although the effect is far less pronounced for South Asians. Most importantly, our multivariate analysis shows that lower breast screening uptake rates in the South Asian population cannot be explained by factors such as deprivation. At a GP practice level, other researchers have identified social deprivation and ethnic-mix in the local population as correlated with breast screening rates in the UK [54]. However, the possible confounding effect of deprivation has generally not been separated from that of ethnicity. In the US, various socio-economic characteristics e.g. income, education, insurance status appear to characterise populations with lower rates of breast cancer screening, with some research evidence emerging to indicate that ethnic disparities in uptake remain even after allowing for socio-economic differences [11, 55–57].

For bowel cancer screening, our results show that the global uptake level achieved (60%) masks significantly lower uptake rates for South Asians; differences are even greater than those observed for breast screening. Both men and women from the South Asian community are far less likely to return an initial home test kit or to subsequently successfully complete the testing process, than are non-Asians. Evidence from other countries demonstrates a similarly low uptake of bowel cancer screening (FOBT, colonoscopy and flexible sigmoidoscopy) by ethnic minority populations [15–19, 23, 58–60]. Studies of American Asian immigrant populations e.g. Koreans and Vietnamese report similar findings [20–22, 61]. Once again, very few of these studies have controlled uptake for differences in socio-economic status, but those that have report inconsistent findings [15, 58, 59]. The present study provides clear evidence that 'ethnicity', although it may correlate with socio-economic status, exerts a separate effect on response to bowel cancer screening. There is also evidence of some delay in the South Asian population before undertaking colonoscopy.

The findings reported in this paper add to the developing evidence base on disparities in access to a range of services by ethnic minorities [62, 63]. It might be anticipated that bowel cancer screening would present greater barriers for ethnic minorities than breast screening. Home FOBT testing requires that individuals not only understand the benefits of screening, but also that they are able to follow specific written instructions in order to collect and preserve samples over a number of days. A separate analysis of round 1 bowel screening data indicates that South Asians who attempt an initial home testing are more likely to be sent 4 or more kits before successful completion (e.g. 28% of Muslims compared with only 3% of non-Asians) [54]. South Asian women are found to exhibit particularly low FOBT completion rates compared to their non-Asian peers, especially among older women. This may be linked to the provision of written materials, including kit instructions, since Muslim and older South Asian women are known to have particularly low literacy levels [64, 65].

The ethnicity of the GP also appeared to be associated with lower uptake of bowel cancer screening by South Asians in our study. The same effect was not observed for breast screening. It is difficult to see why this should be the case since GPs are not directly involved in the bowel screening process. However, studies in other health care systems have identified physician recommendation as a significant predictor of FOBT uptake by ethnic minorities [8, 61, 66, 67].

The implications of our findings for roll-out of the UK bowel cancer screening programme will be especially significant for inner city areas where South Asian populations can reach 40% [53]. In order to ensure equity as well as efficiency in the new bowel cancer screening programme, this population will require special attention. Interventions targeted at particular groups may be needed in order to achieve more equitable uptake rates e.g. Muslims or older South Asian women. At present there is limited research evidence to indicate what types of intervention would prove most effective [68–71]. Equally importantly, the continued disparity in breast screening uptakes after two decades needs to be addressed. A recent review of the literature on interventions to improve breast screening uptake in diverse populations suggests that combined approaches using access-enhancing and individual-directed strategies are most effective [72].

Finally, it might be argued that the incidence of cancer among South Asian populations in the UK is low, and that the observed differences in screening uptake are therefore relatively unimportant. However, the low cancer incidences reported historically are increasingly seen as an artefact of the stage of migration with reports of increased incidence beginning to emerge [73–75]. One limitation of the present study is that we are unable to draw any conclusions about the African Caribbean population because of incomplete ethnic monitoring data and the fact that name recognition software cannot identify these individuals. However, there was a relatively small population in the study area (< 1%). The need to urgently improve routine ethnic data collection to provide improved statistics on cancer incidence and survival for all ethnic minority groups in the UK has recently been highlighted [76].

We conclude that the low breast and bowel cancer screening uptake rates observed in the South Asian population cannot be attributed to socio-economic or age/gender population differences. Although disparities in breast screening have reduced over time, they are still significant. We would suggest that both programmes need to identify and assess culturally appropriate interventions to reduce these observed differences, including provision of tailored health promotion materials for certain South Asian subgroups. It would appear that Muslim invitees registered with a South Asian GP are a prime target for improved bowel screening uptake. More detailed examination of behaviour across the two screening programmes may also help to identify women who have responded positively to breast screening and could therefore be encouraged to complete bowel screening successfully.