Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 7-Day Trial for You or Your Team.

Learn More →

Inequalities in colorectal cancer screening participation in the first round of the national screening programme in England

Inequalities in colorectal cancer screening participation in the first round of the national... British Journal of Cancer (2009) 101, S60 – S63 & 2009 Cancer Research UK All rights reserved 0007 – 0920/09 $32.00 www.bjcancer.com Full Paper Inequalities in colorectal cancer screening participation in the first round of the national screening programme in England ,1 1 2 3 4 4 1 C von Wagner , A Good , D Wright , B Rachet , A Obichere , S Bloom and J Wardle Department of Epidemiology and Public Health, Cancer Research UK Health Behaviour Research Centre, University College London, Gower Street, London WC1E 6BT, UK; Department of Clinical Biochemistry, Northwick Park Hospital, North West London Hospitals NHS Trust, Watford Rd., London, Harrow, UK; Department of Epidemiology and Population Health, Cancer Research UK Cancer Survival Group, Non-Communicable Disease Epidemiology Unit, London School of Hygiene and Tropical Medicine, Keppel Street, London, UK; Department of Gastroenterology, University College London Hospital, Maple House, 25 Grafton Way, London, UK BACKGROUND: Introduction of organised, population-based, colorectal cancer screening in the United Kingdom using the faecal occult blood test (FOBT) has the potential to reduce overall colorectal cancer mortality. However, socio-economic variation in screening participation could exacerbate existing inequalities in mortality. METHODS: This study examined FOBT uptake rates in London, England in relation to area-level socio-economic deprivation over the first 30 months of the programme during which 401 197 individuals were sent an FOBT kit. Uptake was defined as return of a completed test kit within 3 months. Area-level deprivation in each postcode sector was indexed with the Townsend Material Deprivation Index. Analyses controlled for area-level household mobility, ethnic diversity and poor health, each of which was associated with lower return rates. RESULTS: The results showed a strong socio-economic gradient in FOBT uptake, which declined from 49% in the least deprived quintile of postcodes to 38% in the middle quintile and 32% in the most deprived quintile. Variation in socio-economic deprivation between sectors accounted for 62% of the variance in return rates, with little attenuation as a result of controlling for ethnic diversity, household mobility or health status. CONCLUSION: These results highlight the need to understand the causes of socio-economic gradients in screening participation and address barriers that could otherwise increase disparities in colorectal cancer survival. British Journal of Cancer (2009) 101, S60 – S63. doi:10.1038/sj.bjc.6605392 www.bjcancer.com & 2009 Cancer Research UK Keywords: colorectal cancer screening; socio-economic status; ethnicity; self-reported health; health inequality Reducing health inequalities through equality of access to health with medical services and health professionals, could lie behind care is a cornerstone of the UK Government’s health policy, and SES differences in participation. has been re-emphasised in the Department of Health (2007). Whether the same SES differential will be seen for the latest However, associations between socio-economic status (SES) and addition to the UK Cancer Screening Programmes, biennial bowel health have persisted through major changes in social conditions cancer screening using the faecal occult blood test (FOBT), is an and provision of medical care, and it has been argued that important question. In the Bowel Cancer Screening Programme innovations in health technology may widen inequalities if people (BCSP), all 60–69-year olds are sent an invitation to participate with more knowledge, money or power are better able to harness around a week before the test kit is mailed to the home address. the beneficial effects (Link et al, 1998; Baker and Middleton, 2003; The kit is accompanied by step-by-step instructions on how to Coleman et al, 2004). complete the test at home and a hygienic freepost envelope is The UK Breast and Cervical Cancer Screening Programmes were provided to return the samples to the laboratory for processing. adopted more enthusiastically in more affluent areas (Gatrell et al, Individuals are informed of the result and advised if further tests 1998; Webb et al, 2004; Dailey et al, 2007), and at least in the case are needed (repeat FOBT or colonoscopy) within 2 weeks. of cervical cancer, this has had a measurable impact on disparities Faecal occult blood test has the potential to reduce colorectal in mortality (Link et al, 1998). As part of the National Health cancer mortalityby16% (Hewitson et al, 2008) but SES differences Service (NHS), cancer screening in the United Kingdom incurs no in participation could increase existing differentials in survival direct financial costs, so affordability cannot explain variation in (Mitry et al, 2008). The BCSP has some features that ought to min- uptake. However, both breast and cervical screening require imise social inequalities: it incurs no direct or indirect financial cost attendance for a clinic appointment, and therefore factors such as to the individual, no time off work is required, and the test is pressures of time, transport problems or discomfort in interacting delivered direct to the home, is self-administered and returned in a ‘freepost’ envelope. This did not prevent SES differences in the UK trial of FOBT screening (Whynes et al, 2003) or the second round of the UK colorectal screening pilot (Weller et al, 2007). However, as *Correspondence: Dr C von Wagner; E-mail: [email protected] inequalities in participation in the cervical screening programme Inequalities in colorectal cancer screening C von Wagner et al S61 have narrowed over time (Baker and Middleton, 2003), it is possible 60 that this will occur in the context of the BCSP. This study assessed FOBT participation in relation to neigh- bourhood socio-economic characteristics in the London area in the first 30 months of the national roll out of the BCSP. Data were based on test kit return rates for 401 197 individuals aged 60–69, resident in 808 postcode sectors, who were sent FOBT kits between October 2006 and January 2009. 20 MATERIALS AND METHODS Implementation of the screening programme in London Quintiles of area-level deprivation In 2006, the Department of Clinical Biochemistry at Northwick Figure 1 FOBT return by quintile of area-based socio-economic Park and St Marks Hospitals became the London ‘Hub’ for the deprivation (N¼ 401 197). BCSP, responsible for sending out and analysing FOBT kits, and supported by six screening centres carrying out follow-up Table 1 Univariate predictors of area-level uptake of FOBT investigations. The analyses reported here are based on 401 197 first invitations sent out by the Hub across 808 postcode sectors 2 Bb 95% CI R between October 2006 (when the programme went ‘live’) and January 2009. Kits were recorded as returned if they were received Deprivation 0.78** 1.74 1.849, 1.636 0.62 by the screening centre within 3 months. Ethnic diversity 0.62** 24.49 26.774, 22.205 0.38 Area mobility 0.36** 23.55 28.691, 18.417 0.13 Owing to the way the programme is organised, not all eligible Poor health 0.45** 1.903 2.234, 1.572 0.21 individuals would have been invited during the first 30 months of the programme. We therefore divided the number of invitations a Abbreviation: FOBT¼ faecal occult blood test. Transformed variables. **Significant sent to a postcode sector by the number of people in the age range at Po0.001. Bold values are effects size (R ) – the same level of significance as shown living in that sector (data from April 2009) to estimate screening for the unstandardised regression coefficient (B) applies for these and thus does not coverage. On average, 497 invitations were sent per postcode need to be added to the bold values. sector, giving an average of 68% coverage (s.d.¼ 17.0) during the study period. Samples was used to examine the independent effect of deprivation Area-level deprivation after controlling for other area-level factors. As the percentage of the eligible population who had been sent a Postcode sectors are areas defined by the first inward digit of the kit during the first 30 months of the programme varied from 2% to postcode and contain an average of 3000 addresses. Socio- 97% across the 808 postcodes, and there was a marginally economic deprivation for each postcode sector was indexed with significant trend for variation in coverage across deprivation the Townsend Material Deprivation Index (Townsend et al, 1988). quintiles (P¼ 0.07), data were weighted by coverage in further This is based on four area-level indicators from the most recent analyses. Census: levels of unemployment among those who are economic- ally active, owner-occupancy (% of homes owned by the occupier), car ownership (% with a car) and home overcrowding (% with no RESULTS communal living space or less than one bedroom per single adult, The percentage of individuals returning kits increased from 32% in couple or pair of children) (Mackenzie et al, 1998). Higher the most deprived quintile, through 34%, 38% and 42%, to 49% in Townsend scores indicate greater deprivation. the least deprived quintile (see Figure 1). An important observa- tion here is that the results indicate a gradient across quintiles Other area-level measures (linear trend: Po0.001) rather than a sudden ‘drop-off’ in uptake for the most deprived quintile. Univariate regression analysis (see We used 2001 Census data (Office for National Statistics, 2001, Table 1) demonstrated that area deprivation accounted for 62% of 2004) to include the following additional factors in the analyses: the variance in response rates across sectors. the proportion of people in each postcode sector belonging to non- Test kit return was also lower for postcode sectors with higher white ethnic groups (i.e. all ethnic groups other than ‘white ethnic diversity, area mobility or poor health (all Pso0.001), char- British’, ‘white Irish’ and ‘white Other’), levels of reported poor acteristics that were all associated with area deprivation. Regres- health (% of people describing their health as ‘not good’) and sion analyses were used to examine the extent to which deprivation household mobility (% of people who had reported moving out of was associated with screening uptake independently of ethnic an area within 12 months before the 2001 Census). Household diversity, poor health or mobility. The results (see Table 2) showed mobility may cause patient lists to be inflated by records of that associations between each of these neighbourhood character- patients no longer residing in the area. istics and uptake were substantially reduced after including deprivation in the model. Meanwhile the b coefficient for depri- Analyses vation was reduced only slightly by the addition of any of these predictors and was actually increased by controlling for poor Analyses were carried out using SPSS, version 14.0 (SPSS, Inc., health. Chicago, IL, USA). Postcode sectors were grouped into quintiles of Townsend scores to examine the association between area-level deprivation and uptake using the w test for linear trend. Ethnic DISCUSSION diversity and area mobility variables were transformed to correct for skewed distributions: [log (ethnic diversityþ 6); log (area These analyses demonstrate a strong SES differential in uptake of 10 10 mobilityþ 1)]. Linear regression analysis using SPSS Complex the FOBT-based BCSP in London. Residents in the most affluent & 2009 Cancer Research UK British Journal of Cancer (2009) 101(S2), S60 – S63 Percentage of FOBT kits returned Inequalities in colorectal cancer screening C von Wagner et al S62 Table 2 Predicting area level uptake of FOBT: multivariate models from the Indian subcontinent (Whynes et al, 2003). In this ethnic weighted by coverage group, there was evidence that people had more difficulty in completing the test and were more likely to perceive it as BB 95% CI unhygienic, embarrassing and distasteful, pointing to some possible avenues for intervention. Model 1 The finding that areas with a larger proportion of the population Deprivation 0.64** 1.42 1.541, 1.304 a reporting poor health achieved higher uptake accords with Ethnic diversity 0.25** 9.98 11.918, 8.032 individual-level studies identifying perception of good health and R 0.66 absence of symptoms as barriers to early detection (Ogedegbe Model 2 et al, 2005) including uptake of FOBT (Ioannou et al, 2003). Deprivation 0.77** 1.71 1.826, 1.587 However area-based measures are less easily interpreted because of Area mobility 0.03 2.14 6.782, 1.606 the possibility that high concentrations of individuals with poor R 0.62 health may influence the provision or quality of health-care services (for example, in the form of health promotion efforts). Model 3 Nevertheless, the fact that controlling for poor health increased the Deprivation 0.86** 1.91 2.068, 1.757 observed SES gradient highlights the need to emphasise to the Poor health 0.12* 0.49 0.147, 0.825 2 public that screening is for everyone, and not just people with R 0.63 symptoms (Brotherstone et al, 2006). This may have implications for the information materials that accompany the test kit. Model 4: full multivariate model Deprivation 0.71** 1.57 1.799, 1.367 It has been argued that innovations in health technology can Ethnic diversity 0.28** 10.97 12.978, 8.966 widen inequalities if people with more knowledge, money or power Area mobility 0.02 1.48 6.570, 3.256 are better able to harness the beneficial effects (Link et al, 1998; Poor health 0.14* 0.61 0.191, 0.980 Baker and Middleton, 2003; Coleman et al, 2004). While it is not R 0.68 immediately obvious that ‘knowledge, money or power’ explain SES differences in performance of a comparatively simple home test that has no associated cost and requires no social interaction, Abbreviation: FOBT¼ faecal occult blood test. Transformed variables. **Significant at Po0.001, *significant at Po0.05. Bold values are effects size (R ) – the same level the fact that the gradient matches that observed for other forms of of significance as shown for the unstandardised regression coefficient (B) applies for screening and other health behaviours, highlights the importance these and thus does not need to be added to the bold values. of understanding broader influences on engagement with pre- ventive health behaviours, particularly in relation to cancer prevention. areas were significantly more likely to return the test kits than In the meantime, there is an urgent need to identify midstream those in slightly less affluent (though nonetheless still well-to-do) or downstream strategies that can ameliorate inequalities in areas, and this pattern continued linearly through the quintiles of participation in the BCSP. Explanations put forward for social deprivation. Controlling for area mobility as an indicator of differentials in health behaviours often focus specifically on the unreliable address records made no difference to the gradient. lowest SES group (e.g. severe difficulties in health literacy), and Uptake in the most affluent quintile was 50% higher than in the although addressing these may not affect the whole gradient, they most deprived quintile, showing as much of a gradient as was seen could contribute to better health outcomes in the most disadvan- in the early days of the breast or cervical screening programmes taged sector (von Wagner et al, 2009). One potential explanatory (Majeed et al, 1995; Baker and Middleton, 2003). While long- factor that has been shown to have a graded association with SES is itudinal analysis of cervical screening uptake showed a narrowing the perceived value of early detection of cancer (Wardle et al, of the gradient over time (Baker and Middleton, 2003), and this 2004). Fatalism about cancer has been shown to be associated with SES (Powe and Finnie, 2003; Wardle et al, 2004) and could may also turn out to be true for the bowel screening programme, there is a need for immediate action to reduce inequity in delivery. contribute to reluctance to engage with screening if early detection The magnitude of the deprivation effect is particularly worrying is perceived as creating a longer period of suffering before given that the greater economic and socio-cultural mix in London inevitable death. Strategies to promote a more positive view of might have been expected to dilute the association between cancer outcomes could have a role here. Another factor is future indicators of area-level deprivation and screening uptake. Other orientation. There is some evidence that SES is related to the area-level indicators that are known to be associated with relative value attached to future (vs present) gains and losses deprivation and are of particular relevance in London, such as (Wardle and Steptoe, 2003). If the perceived benefit of better ethnic diversity of area mobility, did not explain the socio- survival a long way ahead is more highly discounted by lower SES economic variation. Some of the shortcomings associated with our groups, it could be insufficient to offset the unpleasantness of measures (discussed in more detail below) should be considered collecting faecal samples in the short term. There is no doubt that when interpreting this finding. The measure of ethnic diversity was some people find the idea of completing this test at least relatively simplistic and data regarding area mobility would have unpleasant if not ‘disgusting’ (Chapple et al, 2008), particularly benefited from a more up-to-date source, but notwithstanding because touching faecal material is taboo in all societies. The these limitations, it remains an important finding of the present numbers of toilets or private places in the home to store the analysis that there was a strong and linear gradient of deprivation samples could have a role in compliance; which could be related to on uptake in bowel cancer screening in London that could not be SES, but interestingly, there is also some evidence for a social explained by ethnic diversity or area mobility. Consequently, gradient in the ‘disgust’ response (Rozin et al, 2000). Modifying the strategies such as translating the materials to other languages or sample collection procedure by, for example, introducing alter- addressing culturally specific attitudes to screening, important as native testing modalities requiring fewer samples, could contribute these are in their own right, may only achieve a small impact on to reducing the SES gradient in test participation. the deprivation effect. The above finding should not detract from the observation that Limitations ethnic diversity contributed independently to lower uptake, similar to the effect observed in the English pilot study of the BCSP, which There were limitations to this study. Area-based measures of found lower uptake in areas with higher proportions of residents socio-economic deprivation may be out-of-date because they are British Journal of Cancer (2009) 101(S2), S60 – S63 & 2009 Cancer Research UK Inequalities in colorectal cancer screening C von Wagner et al S63 usually based on census data collected at 10-yearly intervals, but CONCLUSIONS this should serve only to underestimate the deprivation effect. This analysis has demonstrated a striking and linear associa- There can also be large variations in levels of personal deprivation tion between area-level deprivation and the proportion of within an area; which are not identified with area-based measures individuals returning an FOBT kit as part of the UK BCSP; which (Drever, 1995) making it impossible to draw definite conclusions is not explained by differences in ethnic diversity, area mobility about associations between uptake and individual SES. The or health status. It does not provide explanations, but does use of smaller geographical units (e.g. Lower Super Output Areas) emphasise the urgency of constructing more comprehensive multi- as well as commercial geo-demographic classification systems, level explanatory models. In the meantime, the fact that area- such as ACORN or MOSAIC that supplement census level data level characteristics predict participation provides an opportunity with up-to-date demographic and lifestyle data, would give a finer to address midstream factors that maintain disparities in uptake resolution. We also lacked data on sex or age, which might have through tailored educational and instructional materials to offset influenced or interacted with factors in our analysis such as some of the negative effects of deprivation. perceived health status, but these require individual-level data that were not available for these analyses. Variability in coverage may have compromised the extent to which we could generalize the ACKNOWLEDGEMENTS current findings although we did control for coverage in our analysis. It was also reassuring to find that excluding postcode This study was supported by a programme grant from Cancer sectors with less than 50% coverage (N¼ 201) did not alter the Research UK to JW. We acknowledge Gianluca Baio at the effect of area-based socio-economic deprivation. Finally, this Department of Epidemiology and Public Health, UCL who analysis was restricted to test kits sent out in London, which, as provided guidance on a draft of the manuscript. noted above, has a very heterogeneous population. Future studies addressing inequalities in bowel cancer screening will benefit from data from the ‘hubs’ delivering the programme throughout the Conflict of interest country, which will provide information about additional issues such as differences between urban and rural areas. The authors declare no conflict of interest. REFERENCES Baker D, Middleton E (2003) Cervical screening and health inequality in Mitry E, Rachet B, Quinn MJ, Cooper N, Coleman MP (2008) Survival from England in the 1990s. J Epidemiol Community Health 57: 417–423 cancer of the colon in England and Wales up to 2001. Br J Cancer 99: Brotherstone H, Miles A, Robb KA, Atkin W, Wardle J (2006) The impact of S26 –S29 illustrations on public understanding of the aim of cancer screening. Office for National Statistics (2001) Census 2001: Key Statistics for postcode Patient Educ Couns 63: 328–335 sectors in England and Wales. Crown Copyright: London Coleman MP, Rachet B, Woods LM, Mitry E, Riga M, Cooper N, Quinn MJ, Office for National Statistics (2004) Census 2001: Key Statistics for postcode Brenner H, Esteve J (2004) Trends and socioeconomic inequalities in cancer sectors in England and Wales: supplementary files and corrections/CSV survival in England and Wales up to 2001. Br J Cancer 90: 1367 –1373 version 1 August 2004. Crown Copyright: London Chapple A, Ziebland S, Hewitson P, McPherson A (2008) What affects the Ogedegbe G, Cassells AN, Robinson CM, DuHamel K, Tobin JN, Sox CH, uptake of screening for bowel cancer using the faecal occult blood test Dietrich AJ (2005) Perceptions of barriers and facilitators of cancer early (FOBt): a qualitative study. Soc Sci Med 66: 2425–2435 detection among low-income minority women in community health Dailey AB, Kasl SV, Holford TR, Calvocoressi L, Jones BA (2007) centers. J Natl Med Assoc 97: 162–170 Neighborhood-level socioeconomic predictors of nonadherence to Powe BD, Finnie R (2003) Cancer fatalism: the state of the science. Cancer mammography screening guidelines. Cancer Epidemiol Biomarkers Prev Nurs 26: 454– 465 16: 2293–2303 Rozin P, Haidt J, McCauley CR (2000) Disgust. In Handbook of Emotions,Lewis Drever F (1995) Mortality in regions and local authority districts in the MJ, Haviland-Jones JM (eds), pp 637–653. Guildford Press: New York 1990s: exploring the relationship with deprivation. Popul Trends 86: 19–26 Townsend P, Phillimore P, Beattle A (1988) Health and Deprivation: Inequality Department of Health (2007) NHS Cancer Reform Strategy. Department of and the North.London: Croom Helm Health: London von Wagner C, Semmler C, Good A, Wardle J (2009) Health literacy and GatrellA, Garnett S, RigbyJ,Maddocks A,KirwanM(1998) Uptake of self-efficacy for participating in colorectal cancer screening: the role of screening for breast cancer in South Lancashire. Public Health 112: 297–301 information processing. Patient Educ Couns 75(3): 352–357 Hewitson P, Glasziou P, Watson E, Towler B, Irwig L (2008) Cochrane Wardle J, McCaffery K, Nadel M, Atkin W (2004) Socioeconomic systematic review of colorectal cancer screening using the fecal occult differences in cancer screening participation: comparing cognitive and blood test (hemoccult): an update. Am J Gastroenterol 103: 1541–1549 psychosocial explanations. Soc Sci Med 59: 249–261 Ioannou GN, Chapko MK, Dominitz JA (2003) Predictors of colorectal Wardle J, Steptoe A (2003) Socioeconomic differences in attitudes and beliefs cancer screening participation in the United States. Am J Gastroenterol about healthy lifestyles. J Epidemiol Community Health 57: 440–443 98: 2082–2091 Webb R, Richardson J, Esmail A, Pickles A (2004) Uptake for cervical Link BG, Northridge ME, Phelan JC, Ganz ML (1998) Social epidemiology screening by ethnicity and place-of-birth: a population-based cross- and the fundamental cause concept: on the structuring of effective cancer sectional study. J Public Health 26: 293–296 screens by socioeconomic status. Milbank Q 76: 375–402 Weller D, Coleman D, Robertson R, Butler P, Melia J, Campbell C, Parker R, Mackenzie IF, Nelder R, Maconachie M, Radford G (1998) ‘My ward is Patnick J, Moss S (2007) The UK colorectal cancer screening pilot: results more deprived than yours’. J Public Health 20: 186–190 of the second round of screening in England. Br J Cancer 97: 1601 –1605 Majeed FA, Cook DG, Given-Wilson R, Vecchi P, Poloniecki J (1995) Whynes DK, Frew EJ, Manghan CM, Scholefield JH, Hardcastle JD (2003) Do general practitioners influence the uptake of breast cancer screening? Colorectal cancer, screening and survival: the influence of socio- J Med Screen 2: 119–124 economic deprivation. Public Health 117: 389–395 & 2009 Cancer Research UK British Journal of Cancer (2009) 101(S2), S60 – S63 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png British Journal of Cancer Springer Journals

Inequalities in colorectal cancer screening participation in the first round of the national screening programme in England

Download PDF
4 pages

Loading next page...
 
/lp/springer-journals/inequalities-in-colorectal-cancer-screening-participation-in-the-first-0KwkZj0muc

References (51)

Publisher
Springer Journals
Copyright
Copyright © 2009 by The Author(s)
Subject
Biomedicine; Biomedicine, general; Cancer Research; Epidemiology; Molecular Medicine; Oncology; Drug Resistance
ISSN
0007-0920
eISSN
1532-1827
DOI
10.1038/sj.bjc.6605392
Publisher site
See Article on Publisher Site

Abstract

British Journal of Cancer (2009) 101, S60 – S63 & 2009 Cancer Research UK All rights reserved 0007 – 0920/09 $32.00 www.bjcancer.com Full Paper Inequalities in colorectal cancer screening participation in the first round of the national screening programme in England ,1 1 2 3 4 4 1 C von Wagner , A Good , D Wright , B Rachet , A Obichere , S Bloom and J Wardle Department of Epidemiology and Public Health, Cancer Research UK Health Behaviour Research Centre, University College London, Gower Street, London WC1E 6BT, UK; Department of Clinical Biochemistry, Northwick Park Hospital, North West London Hospitals NHS Trust, Watford Rd., London, Harrow, UK; Department of Epidemiology and Population Health, Cancer Research UK Cancer Survival Group, Non-Communicable Disease Epidemiology Unit, London School of Hygiene and Tropical Medicine, Keppel Street, London, UK; Department of Gastroenterology, University College London Hospital, Maple House, 25 Grafton Way, London, UK BACKGROUND: Introduction of organised, population-based, colorectal cancer screening in the United Kingdom using the faecal occult blood test (FOBT) has the potential to reduce overall colorectal cancer mortality. However, socio-economic variation in screening participation could exacerbate existing inequalities in mortality. METHODS: This study examined FOBT uptake rates in London, England in relation to area-level socio-economic deprivation over the first 30 months of the programme during which 401 197 individuals were sent an FOBT kit. Uptake was defined as return of a completed test kit within 3 months. Area-level deprivation in each postcode sector was indexed with the Townsend Material Deprivation Index. Analyses controlled for area-level household mobility, ethnic diversity and poor health, each of which was associated with lower return rates. RESULTS: The results showed a strong socio-economic gradient in FOBT uptake, which declined from 49% in the least deprived quintile of postcodes to 38% in the middle quintile and 32% in the most deprived quintile. Variation in socio-economic deprivation between sectors accounted for 62% of the variance in return rates, with little attenuation as a result of controlling for ethnic diversity, household mobility or health status. CONCLUSION: These results highlight the need to understand the causes of socio-economic gradients in screening participation and address barriers that could otherwise increase disparities in colorectal cancer survival. British Journal of Cancer (2009) 101, S60 – S63. doi:10.1038/sj.bjc.6605392 www.bjcancer.com & 2009 Cancer Research UK Keywords: colorectal cancer screening; socio-economic status; ethnicity; self-reported health; health inequality Reducing health inequalities through equality of access to health with medical services and health professionals, could lie behind care is a cornerstone of the UK Government’s health policy, and SES differences in participation. has been re-emphasised in the Department of Health (2007). Whether the same SES differential will be seen for the latest However, associations between socio-economic status (SES) and addition to the UK Cancer Screening Programmes, biennial bowel health have persisted through major changes in social conditions cancer screening using the faecal occult blood test (FOBT), is an and provision of medical care, and it has been argued that important question. In the Bowel Cancer Screening Programme innovations in health technology may widen inequalities if people (BCSP), all 60–69-year olds are sent an invitation to participate with more knowledge, money or power are better able to harness around a week before the test kit is mailed to the home address. the beneficial effects (Link et al, 1998; Baker and Middleton, 2003; The kit is accompanied by step-by-step instructions on how to Coleman et al, 2004). complete the test at home and a hygienic freepost envelope is The UK Breast and Cervical Cancer Screening Programmes were provided to return the samples to the laboratory for processing. adopted more enthusiastically in more affluent areas (Gatrell et al, Individuals are informed of the result and advised if further tests 1998; Webb et al, 2004; Dailey et al, 2007), and at least in the case are needed (repeat FOBT or colonoscopy) within 2 weeks. of cervical cancer, this has had a measurable impact on disparities Faecal occult blood test has the potential to reduce colorectal in mortality (Link et al, 1998). As part of the National Health cancer mortalityby16% (Hewitson et al, 2008) but SES differences Service (NHS), cancer screening in the United Kingdom incurs no in participation could increase existing differentials in survival direct financial costs, so affordability cannot explain variation in (Mitry et al, 2008). The BCSP has some features that ought to min- uptake. However, both breast and cervical screening require imise social inequalities: it incurs no direct or indirect financial cost attendance for a clinic appointment, and therefore factors such as to the individual, no time off work is required, and the test is pressures of time, transport problems or discomfort in interacting delivered direct to the home, is self-administered and returned in a ‘freepost’ envelope. This did not prevent SES differences in the UK trial of FOBT screening (Whynes et al, 2003) or the second round of the UK colorectal screening pilot (Weller et al, 2007). However, as *Correspondence: Dr C von Wagner; E-mail: [email protected] inequalities in participation in the cervical screening programme Inequalities in colorectal cancer screening C von Wagner et al S61 have narrowed over time (Baker and Middleton, 2003), it is possible 60 that this will occur in the context of the BCSP. This study assessed FOBT participation in relation to neigh- bourhood socio-economic characteristics in the London area in the first 30 months of the national roll out of the BCSP. Data were based on test kit return rates for 401 197 individuals aged 60–69, resident in 808 postcode sectors, who were sent FOBT kits between October 2006 and January 2009. 20 MATERIALS AND METHODS Implementation of the screening programme in London Quintiles of area-level deprivation In 2006, the Department of Clinical Biochemistry at Northwick Figure 1 FOBT return by quintile of area-based socio-economic Park and St Marks Hospitals became the London ‘Hub’ for the deprivation (N¼ 401 197). BCSP, responsible for sending out and analysing FOBT kits, and supported by six screening centres carrying out follow-up Table 1 Univariate predictors of area-level uptake of FOBT investigations. The analyses reported here are based on 401 197 first invitations sent out by the Hub across 808 postcode sectors 2 Bb 95% CI R between October 2006 (when the programme went ‘live’) and January 2009. Kits were recorded as returned if they were received Deprivation 0.78** 1.74 1.849, 1.636 0.62 by the screening centre within 3 months. Ethnic diversity 0.62** 24.49 26.774, 22.205 0.38 Area mobility 0.36** 23.55 28.691, 18.417 0.13 Owing to the way the programme is organised, not all eligible Poor health 0.45** 1.903 2.234, 1.572 0.21 individuals would have been invited during the first 30 months of the programme. We therefore divided the number of invitations a Abbreviation: FOBT¼ faecal occult blood test. Transformed variables. **Significant sent to a postcode sector by the number of people in the age range at Po0.001. Bold values are effects size (R ) – the same level of significance as shown living in that sector (data from April 2009) to estimate screening for the unstandardised regression coefficient (B) applies for these and thus does not coverage. On average, 497 invitations were sent per postcode need to be added to the bold values. sector, giving an average of 68% coverage (s.d.¼ 17.0) during the study period. Samples was used to examine the independent effect of deprivation Area-level deprivation after controlling for other area-level factors. As the percentage of the eligible population who had been sent a Postcode sectors are areas defined by the first inward digit of the kit during the first 30 months of the programme varied from 2% to postcode and contain an average of 3000 addresses. Socio- 97% across the 808 postcodes, and there was a marginally economic deprivation for each postcode sector was indexed with significant trend for variation in coverage across deprivation the Townsend Material Deprivation Index (Townsend et al, 1988). quintiles (P¼ 0.07), data were weighted by coverage in further This is based on four area-level indicators from the most recent analyses. Census: levels of unemployment among those who are economic- ally active, owner-occupancy (% of homes owned by the occupier), car ownership (% with a car) and home overcrowding (% with no RESULTS communal living space or less than one bedroom per single adult, The percentage of individuals returning kits increased from 32% in couple or pair of children) (Mackenzie et al, 1998). Higher the most deprived quintile, through 34%, 38% and 42%, to 49% in Townsend scores indicate greater deprivation. the least deprived quintile (see Figure 1). An important observa- tion here is that the results indicate a gradient across quintiles Other area-level measures (linear trend: Po0.001) rather than a sudden ‘drop-off’ in uptake for the most deprived quintile. Univariate regression analysis (see We used 2001 Census data (Office for National Statistics, 2001, Table 1) demonstrated that area deprivation accounted for 62% of 2004) to include the following additional factors in the analyses: the variance in response rates across sectors. the proportion of people in each postcode sector belonging to non- Test kit return was also lower for postcode sectors with higher white ethnic groups (i.e. all ethnic groups other than ‘white ethnic diversity, area mobility or poor health (all Pso0.001), char- British’, ‘white Irish’ and ‘white Other’), levels of reported poor acteristics that were all associated with area deprivation. Regres- health (% of people describing their health as ‘not good’) and sion analyses were used to examine the extent to which deprivation household mobility (% of people who had reported moving out of was associated with screening uptake independently of ethnic an area within 12 months before the 2001 Census). Household diversity, poor health or mobility. The results (see Table 2) showed mobility may cause patient lists to be inflated by records of that associations between each of these neighbourhood character- patients no longer residing in the area. istics and uptake were substantially reduced after including deprivation in the model. Meanwhile the b coefficient for depri- Analyses vation was reduced only slightly by the addition of any of these predictors and was actually increased by controlling for poor Analyses were carried out using SPSS, version 14.0 (SPSS, Inc., health. Chicago, IL, USA). Postcode sectors were grouped into quintiles of Townsend scores to examine the association between area-level deprivation and uptake using the w test for linear trend. Ethnic DISCUSSION diversity and area mobility variables were transformed to correct for skewed distributions: [log (ethnic diversityþ 6); log (area These analyses demonstrate a strong SES differential in uptake of 10 10 mobilityþ 1)]. Linear regression analysis using SPSS Complex the FOBT-based BCSP in London. Residents in the most affluent & 2009 Cancer Research UK British Journal of Cancer (2009) 101(S2), S60 – S63 Percentage of FOBT kits returned Inequalities in colorectal cancer screening C von Wagner et al S62 Table 2 Predicting area level uptake of FOBT: multivariate models from the Indian subcontinent (Whynes et al, 2003). In this ethnic weighted by coverage group, there was evidence that people had more difficulty in completing the test and were more likely to perceive it as BB 95% CI unhygienic, embarrassing and distasteful, pointing to some possible avenues for intervention. Model 1 The finding that areas with a larger proportion of the population Deprivation 0.64** 1.42 1.541, 1.304 a reporting poor health achieved higher uptake accords with Ethnic diversity 0.25** 9.98 11.918, 8.032 individual-level studies identifying perception of good health and R 0.66 absence of symptoms as barriers to early detection (Ogedegbe Model 2 et al, 2005) including uptake of FOBT (Ioannou et al, 2003). Deprivation 0.77** 1.71 1.826, 1.587 However area-based measures are less easily interpreted because of Area mobility 0.03 2.14 6.782, 1.606 the possibility that high concentrations of individuals with poor R 0.62 health may influence the provision or quality of health-care services (for example, in the form of health promotion efforts). Model 3 Nevertheless, the fact that controlling for poor health increased the Deprivation 0.86** 1.91 2.068, 1.757 observed SES gradient highlights the need to emphasise to the Poor health 0.12* 0.49 0.147, 0.825 2 public that screening is for everyone, and not just people with R 0.63 symptoms (Brotherstone et al, 2006). This may have implications for the information materials that accompany the test kit. Model 4: full multivariate model Deprivation 0.71** 1.57 1.799, 1.367 It has been argued that innovations in health technology can Ethnic diversity 0.28** 10.97 12.978, 8.966 widen inequalities if people with more knowledge, money or power Area mobility 0.02 1.48 6.570, 3.256 are better able to harness the beneficial effects (Link et al, 1998; Poor health 0.14* 0.61 0.191, 0.980 Baker and Middleton, 2003; Coleman et al, 2004). While it is not R 0.68 immediately obvious that ‘knowledge, money or power’ explain SES differences in performance of a comparatively simple home test that has no associated cost and requires no social interaction, Abbreviation: FOBT¼ faecal occult blood test. Transformed variables. **Significant at Po0.001, *significant at Po0.05. Bold values are effects size (R ) – the same level the fact that the gradient matches that observed for other forms of of significance as shown for the unstandardised regression coefficient (B) applies for screening and other health behaviours, highlights the importance these and thus does not need to be added to the bold values. of understanding broader influences on engagement with pre- ventive health behaviours, particularly in relation to cancer prevention. areas were significantly more likely to return the test kits than In the meantime, there is an urgent need to identify midstream those in slightly less affluent (though nonetheless still well-to-do) or downstream strategies that can ameliorate inequalities in areas, and this pattern continued linearly through the quintiles of participation in the BCSP. Explanations put forward for social deprivation. Controlling for area mobility as an indicator of differentials in health behaviours often focus specifically on the unreliable address records made no difference to the gradient. lowest SES group (e.g. severe difficulties in health literacy), and Uptake in the most affluent quintile was 50% higher than in the although addressing these may not affect the whole gradient, they most deprived quintile, showing as much of a gradient as was seen could contribute to better health outcomes in the most disadvan- in the early days of the breast or cervical screening programmes taged sector (von Wagner et al, 2009). One potential explanatory (Majeed et al, 1995; Baker and Middleton, 2003). While long- factor that has been shown to have a graded association with SES is itudinal analysis of cervical screening uptake showed a narrowing the perceived value of early detection of cancer (Wardle et al, of the gradient over time (Baker and Middleton, 2003), and this 2004). Fatalism about cancer has been shown to be associated with SES (Powe and Finnie, 2003; Wardle et al, 2004) and could may also turn out to be true for the bowel screening programme, there is a need for immediate action to reduce inequity in delivery. contribute to reluctance to engage with screening if early detection The magnitude of the deprivation effect is particularly worrying is perceived as creating a longer period of suffering before given that the greater economic and socio-cultural mix in London inevitable death. Strategies to promote a more positive view of might have been expected to dilute the association between cancer outcomes could have a role here. Another factor is future indicators of area-level deprivation and screening uptake. Other orientation. There is some evidence that SES is related to the area-level indicators that are known to be associated with relative value attached to future (vs present) gains and losses deprivation and are of particular relevance in London, such as (Wardle and Steptoe, 2003). If the perceived benefit of better ethnic diversity of area mobility, did not explain the socio- survival a long way ahead is more highly discounted by lower SES economic variation. Some of the shortcomings associated with our groups, it could be insufficient to offset the unpleasantness of measures (discussed in more detail below) should be considered collecting faecal samples in the short term. There is no doubt that when interpreting this finding. The measure of ethnic diversity was some people find the idea of completing this test at least relatively simplistic and data regarding area mobility would have unpleasant if not ‘disgusting’ (Chapple et al, 2008), particularly benefited from a more up-to-date source, but notwithstanding because touching faecal material is taboo in all societies. The these limitations, it remains an important finding of the present numbers of toilets or private places in the home to store the analysis that there was a strong and linear gradient of deprivation samples could have a role in compliance; which could be related to on uptake in bowel cancer screening in London that could not be SES, but interestingly, there is also some evidence for a social explained by ethnic diversity or area mobility. Consequently, gradient in the ‘disgust’ response (Rozin et al, 2000). Modifying the strategies such as translating the materials to other languages or sample collection procedure by, for example, introducing alter- addressing culturally specific attitudes to screening, important as native testing modalities requiring fewer samples, could contribute these are in their own right, may only achieve a small impact on to reducing the SES gradient in test participation. the deprivation effect. The above finding should not detract from the observation that Limitations ethnic diversity contributed independently to lower uptake, similar to the effect observed in the English pilot study of the BCSP, which There were limitations to this study. Area-based measures of found lower uptake in areas with higher proportions of residents socio-economic deprivation may be out-of-date because they are British Journal of Cancer (2009) 101(S2), S60 – S63 & 2009 Cancer Research UK Inequalities in colorectal cancer screening C von Wagner et al S63 usually based on census data collected at 10-yearly intervals, but CONCLUSIONS this should serve only to underestimate the deprivation effect. This analysis has demonstrated a striking and linear associa- There can also be large variations in levels of personal deprivation tion between area-level deprivation and the proportion of within an area; which are not identified with area-based measures individuals returning an FOBT kit as part of the UK BCSP; which (Drever, 1995) making it impossible to draw definite conclusions is not explained by differences in ethnic diversity, area mobility about associations between uptake and individual SES. The or health status. It does not provide explanations, but does use of smaller geographical units (e.g. Lower Super Output Areas) emphasise the urgency of constructing more comprehensive multi- as well as commercial geo-demographic classification systems, level explanatory models. In the meantime, the fact that area- such as ACORN or MOSAIC that supplement census level data level characteristics predict participation provides an opportunity with up-to-date demographic and lifestyle data, would give a finer to address midstream factors that maintain disparities in uptake resolution. We also lacked data on sex or age, which might have through tailored educational and instructional materials to offset influenced or interacted with factors in our analysis such as some of the negative effects of deprivation. perceived health status, but these require individual-level data that were not available for these analyses. Variability in coverage may have compromised the extent to which we could generalize the ACKNOWLEDGEMENTS current findings although we did control for coverage in our analysis. It was also reassuring to find that excluding postcode This study was supported by a programme grant from Cancer sectors with less than 50% coverage (N¼ 201) did not alter the Research UK to JW. We acknowledge Gianluca Baio at the effect of area-based socio-economic deprivation. Finally, this Department of Epidemiology and Public Health, UCL who analysis was restricted to test kits sent out in London, which, as provided guidance on a draft of the manuscript. noted above, has a very heterogeneous population. Future studies addressing inequalities in bowel cancer screening will benefit from data from the ‘hubs’ delivering the programme throughout the Conflict of interest country, which will provide information about additional issues such as differences between urban and rural areas. The authors declare no conflict of interest. REFERENCES Baker D, Middleton E (2003) Cervical screening and health inequality in Mitry E, Rachet B, Quinn MJ, Cooper N, Coleman MP (2008) Survival from England in the 1990s. J Epidemiol Community Health 57: 417–423 cancer of the colon in England and Wales up to 2001. Br J Cancer 99: Brotherstone H, Miles A, Robb KA, Atkin W, Wardle J (2006) The impact of S26 –S29 illustrations on public understanding of the aim of cancer screening. Office for National Statistics (2001) Census 2001: Key Statistics for postcode Patient Educ Couns 63: 328–335 sectors in England and Wales. Crown Copyright: London Coleman MP, Rachet B, Woods LM, Mitry E, Riga M, Cooper N, Quinn MJ, Office for National Statistics (2004) Census 2001: Key Statistics for postcode Brenner H, Esteve J (2004) Trends and socioeconomic inequalities in cancer sectors in England and Wales: supplementary files and corrections/CSV survival in England and Wales up to 2001. Br J Cancer 90: 1367 –1373 version 1 August 2004. Crown Copyright: London Chapple A, Ziebland S, Hewitson P, McPherson A (2008) What affects the Ogedegbe G, Cassells AN, Robinson CM, DuHamel K, Tobin JN, Sox CH, uptake of screening for bowel cancer using the faecal occult blood test Dietrich AJ (2005) Perceptions of barriers and facilitators of cancer early (FOBt): a qualitative study. Soc Sci Med 66: 2425–2435 detection among low-income minority women in community health Dailey AB, Kasl SV, Holford TR, Calvocoressi L, Jones BA (2007) centers. J Natl Med Assoc 97: 162–170 Neighborhood-level socioeconomic predictors of nonadherence to Powe BD, Finnie R (2003) Cancer fatalism: the state of the science. Cancer mammography screening guidelines. Cancer Epidemiol Biomarkers Prev Nurs 26: 454– 465 16: 2293–2303 Rozin P, Haidt J, McCauley CR (2000) Disgust. In Handbook of Emotions,Lewis Drever F (1995) Mortality in regions and local authority districts in the MJ, Haviland-Jones JM (eds), pp 637–653. Guildford Press: New York 1990s: exploring the relationship with deprivation. Popul Trends 86: 19–26 Townsend P, Phillimore P, Beattle A (1988) Health and Deprivation: Inequality Department of Health (2007) NHS Cancer Reform Strategy. Department of and the North.London: Croom Helm Health: London von Wagner C, Semmler C, Good A, Wardle J (2009) Health literacy and GatrellA, Garnett S, RigbyJ,Maddocks A,KirwanM(1998) Uptake of self-efficacy for participating in colorectal cancer screening: the role of screening for breast cancer in South Lancashire. Public Health 112: 297–301 information processing. Patient Educ Couns 75(3): 352–357 Hewitson P, Glasziou P, Watson E, Towler B, Irwig L (2008) Cochrane Wardle J, McCaffery K, Nadel M, Atkin W (2004) Socioeconomic systematic review of colorectal cancer screening using the fecal occult differences in cancer screening participation: comparing cognitive and blood test (hemoccult): an update. Am J Gastroenterol 103: 1541–1549 psychosocial explanations. Soc Sci Med 59: 249–261 Ioannou GN, Chapko MK, Dominitz JA (2003) Predictors of colorectal Wardle J, Steptoe A (2003) Socioeconomic differences in attitudes and beliefs cancer screening participation in the United States. Am J Gastroenterol about healthy lifestyles. J Epidemiol Community Health 57: 440–443 98: 2082–2091 Webb R, Richardson J, Esmail A, Pickles A (2004) Uptake for cervical Link BG, Northridge ME, Phelan JC, Ganz ML (1998) Social epidemiology screening by ethnicity and place-of-birth: a population-based cross- and the fundamental cause concept: on the structuring of effective cancer sectional study. J Public Health 26: 293–296 screens by socioeconomic status. Milbank Q 76: 375–402 Weller D, Coleman D, Robertson R, Butler P, Melia J, Campbell C, Parker R, Mackenzie IF, Nelder R, Maconachie M, Radford G (1998) ‘My ward is Patnick J, Moss S (2007) The UK colorectal cancer screening pilot: results more deprived than yours’. J Public Health 20: 186–190 of the second round of screening in England. Br J Cancer 97: 1601 –1605 Majeed FA, Cook DG, Given-Wilson R, Vecchi P, Poloniecki J (1995) Whynes DK, Frew EJ, Manghan CM, Scholefield JH, Hardcastle JD (2003) Do general practitioners influence the uptake of breast cancer screening? Colorectal cancer, screening and survival: the influence of socio- J Med Screen 2: 119–124 economic deprivation. Public Health 117: 389–395 & 2009 Cancer Research UK British Journal of Cancer (2009) 101(S2), S60 – S63

Journal

British Journal of Cancer – Springer Journals

Published: Dec 3, 2009

There are no references for this article.