Inequalities in breast cancer stage at diagnosis in the Trent region, and implications for the NHS Breast Screening Programme

Transcription

1 Journal of Public Health Vol. 31, No. 3, pp doi: /pubmed/fdp042 Advance Access Publication 7 May 2009 Inequalities in breast cancer stage at diagnosis in the Trent region, and implications for the NHS Breast Screening Programme Sarah A. Cuthbertson 1, Elizabeth C. Goyder 2, Jason Poole 1 1 Trent Cancer Registry, 5 Old Fulwood Road, Sheffield S10 3TG, UK 2 ScHARR, University of Sheffield, Regent Court, 30 Regent Street, Sheffield S1 4DA, UK Address correspondence to Sarah A. Cuthbertson, sarah.cuthbertson@nhs.net ABSTRACT Background This study investigates risk factors for diagnosis with late-stage breast cancer in order to identify inequalities and inform the understanding of barriers affecting access to mammography screening. Methods Data from the Trent Cancer Registry were used to identify all women with invasive breast cancer, diagnosed in Risk of diagnosis with late-stage breast cancer was calculated to quantify strength of association between exposure and outcome. Results Women outside the age group for routine screening were approximately 30% [,50 years, relative risk (RR) ¼ 1.34 (95% confidence interval, CI: ) and.70 years, RR ¼ 1.27 (95% CI: )] more likely to be diagnosed with late-stage breast cancer; the most deprived women were 37% [RR ¼ 1.37 (95% CI: )] more likely to be diagnosed with Stage IV breast cancer; ethnic minority women were 15% [RR ¼ 1.15 (95% CI: )] more likely to be diagnosed with late-stage breast cancer and women resident in 5 of 11 Trent Primary Care Trusts (PCT) had a greater than 30% increased risk of diagnosis with late-stage breast cancer than those in Nottingham City PCT. Conclusion These findings highlight the need for appropriate targeted interventions to address compositional and contextual inequalities that are evident in breast cancer stage at diagnosis. Keywords breast cancer, inequalities, screening Introduction Breast cancer is the most common cancer in women in the UK. More than women are diagnosed with breast cancer each year and it remains the second biggest cause of cancer deaths. 1 Detecting and treating breast cancer at an early stage in the disease pathway is an effective way to reduce morbidity and prevent mortality associated with breast cancer. 2 To improve early detection and treatment, it is essential to identify risk factors associated with diagnosis at advanced stages that result in higher breast cancer mortality rates. 2,3 Many studies have therefore emphasized the need to address inequalities in stage of disease at time of diagnosis. 4 Mammography is one method of detecting breast cancer at an early stage, with research showing that early diagnosis is significantly correlated with population rates of mammography screening. 5,6 Since the mid-1990s, the NHS Breast Screening Programme (NHSBSP) has invited all eligible women aged to attend a screening appointment every 3 years, with this recently extended to include women aged Women over 70 years continue to be screened on request. 7 However, many women in England continue to be diagnosed with late-stage breast cancer, and research suggests socioeconomic gradients of risk within the NHSBSP. 3,8,9 The degree to which a population group is at risk is strongly influenced by demographic, socioeconomic and geographical barriers, yet the relative importance of these barriers in affecting the risk of late-stage breast cancer is complex and poorly understood, 3 and previous research has produced inconsistent results. Research to date shows that the role of age, deprivation, ethnicity and geography in defining breast cancer inequalities Sarah A. Cuthbertson, Senior Cancer Information Analyst Elizabeth C. Goyder, Reader in Public Health Jason Poole, Head of Cancer Analysis 398 # The Author 2009, Published by Oxford University Press on behalf of Faculty of Public Health. All rights reserved.

2 INEQUALITIES IN BREAST CANCER STAGE AT DIAGNOSIS 399 is not clear cut, but that some sort of potentially avoidable inequality gradients may be in operation. Given the international scope of the research, and the small pool of UKspecific literature, alongside the importance of health inequalities on the UK health agenda, and government promises to improve the early detection of cancer, 10 further research to identify risk factors associated with advanced stages of breast cancer is therefore warranted. This study aimed to use routinely available cancer registry data to inform understanding of barriers affecting the risk of latestage breast cancer in an English region, to inform public health planning and the development of interventions to address inequalities. Methods Using a population-based, cross-sectional study design, all unduplicated cases of women reported to the Trent Cancer Registry (TrCR) with invasive breast cancer (International Classification of Diseases-10 C50), between 1998 and 2006, resident in the Trent region at time of diagnosis, were eligible for inclusion in the study. In situ breast cancer was not considered due to the likelihood that cases are underreported and due to some controversy as to its potential to progress to invasive carcinoma. Consequently, it is often treated separately when examining breast cancer stages. 2 TrCR defines stages of breast cancer with the TNM system. The Malmö mammographic screening trial reported that % of breast cancer deaths occurred among women diagnosed with Stages II, III or IV breast cancer. 11 Therefore, for the purpose of this study, early-stage breast cancer is defined as Stage I and late-stage breast cancer is defined as Stages II, III and IV (Stage IIþ). The primary outcome for investigating inequalities in breast cancer stages at diagnosis is an estimation of the risk ratio. Age, deprivation, ethnicity and Primary Care Trust (PCT) of residence are the risk factors used to assess the extent of inequalities in stage at diagnosis. These variables were readily available from the TrCR dataset, and they have been extensively used in previous research to measure inequalities in breast cancer stage at diagnosis, and therefore offer a robust measure. Data analysis Since women aged are routinely invited for screening, inequalities in stage at diagnosis were evaluated by comparing,50, and.70 age groups. Deprivation was defined by the Index of Multiple Deprivation 2007: a measure of area-based deprivation at the Lower Super Output Area level. Ethnicity was populated in the Registry dataset with data from Hospital Episode Statistics and is coded according to the Office for National Statistics Classification for Ethnic Categories. 12 Area of patient residence (based on PCT boundaries) was derived from the postcode for each record. Nottingham City PCT is used as the baseline for the PCT analysis because it most closely reflects the regional average, with no significant differences in the percentage of breast cancer recorded at each stage, compared with the regional average. To account for small numbers, analyses by ethnic group were conducted using the full 9-year cohort of data ( ), from the year the population-based NHSBSP began to the most current year of complete registry data. Analyses by deprivation, age and PCT were conducted using a 2-year cohort of women diagnosed in This is because the numbers are large enough for these explanatory variables, and because for a period in 2004, stage was underrecorded by the Registry. The relative differences between the variables of interest are expressed as unadjusted percentages with 95% confidence intervals (CI) using the Wilson Score method, as described by APHO. 13 To adjust for confounding variables, risk ratios were estimated in Stata using generalized linear models, with a modified Poisson method proposed by Zou. 14 Post-estimation was used to perform a goodnessof-fit test of the model. Results About women were diagnosed with invasive breast cancer in the Trent region from 1998 to 2006 and 7814 women were diagnosed with invasive breast cancer in the Trent region from 2005 to 06. Around 7176 (22%) women were excluded from the 1998 to 2006 cohort and 1141 (15%) women from the 2005 to 06 cohort because they did not have ethnicity recorded. This gave a total of women in the cohort and 6673 women in the cohort. Results are presented for overall late stage (IIþ), Stage IV the most advanced stage with the poorest prognosis and for unknown stage to highlight the significance of the under-reporting of stage at diagnosis. Variation by age Table 1 shows that the unadjusted proportion of breast cancers diagnosed as late stage is significantly lower in the age group for routine invitation to breast screening (50 70 year olds) [43.6% (95% CI: ] compared with the

3 400 JOURNAL OF PUBLIC HEALTH Table 1 Distribution and risk of stage at diagnosis by age group ( ) Age group (years) by stage n % Cases a (95% CI) RR b (95% CI) Late stage (Stage IIþ) 2456/ ( ) Late versus early stage,50 617/ ( ) 1.34 ( ) / ( ) / ( ) 1.27 (1.19 to 1.36) Stage IV 154/ ( ) Stage IV versus I,50 31/ ( ) 1.97 ( ) / ( ) / ( ) 3.45 ( ) Unknown stage 1725/ ( ) Unknown versus known stage,50 238/ ( ) 0.99 ( ) / ( ) / ( ) 0.69 ( ) a Except for unknown stage, % cases are for those with known stage by age group. b RR is adjusted for deprivation (5 groups), ethnicity (6 groups) and PCT (12 groups). regional average [49.6% (95% CI: )] and to those outside the screening age group [59.0% (95% CI: ) and 55.8% (95% CI: ) in the,50 and.70 age groups, respectively]. The unadjusted proportion of breast cancer diagnosed as Stage IV is significantly higher in the.70 age group [6.1% (95% CI: )] compared with the regional average, the,50 age group and to the screening age group [3.1% (95% CI: ); 3.0% (95% CI: ) and 2.0 (95% CI: )]. After adjusting for ethnic group, deprivation and PCT of residence, the clear increased risk of diagnosis with late-stage breast cancer for women outside the screening age group relative to women in the screening age group remains: relative risk (RR) ¼ 1.34 (95% CI: ) and RR ¼ 1.27 (95% CI: ) in the,50 and.70 age groups, respectively. There is also a clear increased risk of diagnosis with Stage IV breast cancer for women outside the screening age group: RR ¼ 1.97 (95% CI: ) and RR ¼ 3.45 (95% CI: ) in the,50 and.70 age groups, respectively. Variation by deprivation Table 2 shows that for this cohort of patients, the only significant difference between deprivation quintile groups in the unadjusted proportions of cancers recorded at each stage is for women in deprivation quintile Group 4 who have a significantly higher proportion of Stage IV cancers compared with women in quintile Group 1 [4.1% (95% CI: ) and 1.9% (95% CI: ), respectively]. However, after adjusting for age, ethnic group and PCT of residence, women in deprivation quintile Groups 3, 4 and 5 have a significantly increased risk of diagnosis with Stage IV breast cancer [RR ¼ 1.87 (95% CI: ); RR ¼ 2.00 (95% CI: ) and RR ¼ 1.37 (95% CI: ), respectively]. Women in deprivation Groups 3 and 4 also show a significantly increased risk of diagnosis with late-stage breast cancer [RR ¼ 1.09 (95% CI: ) and RR ¼ 1.11 (95% CI: ), respectively]. Variation by ethnicity Table 3 shows that compared with the proportions of latestage breast cancer recorded for the White British group [51.1% (95% CI: )], the White Other and All Other groups have a significantly larger proportion of latestage breast cancers recorded [59.9% (95% CI: ) and 60.8% (95% CI: ), respectively]. After adjusting for age, deprivation and PCT of residence, the ethnic minority groups (except for the Mixed group) have a significantly increased risk of diagnosis with late-stage breast cancer relative to the White British group. For the Black/Black British and Chinese/Other ethnic groups, risk of diagnosis with late-stage breast cancer is more than 25% higher [RR ¼ 1.28 (95% CI: ) and RR ¼ 1.26 (95% CI: )]. For the minority ethnic groups combined, the All Other group has a significantly increased risk [RR ¼ 1.15 (95% CI: )]. Variation by geography After adjusting for age, deprivation and ethnicity, nine Trent PCTs have a significantly increased risk of diagnosis with late-stage breast cancer, relative to women resident in

4 INEQUALITIES IN BREAST CANCER STAGE AT DIAGNOSIS 401 Table 2 Distribution and risk of stage at diagnosis by deprivation quintile group ( ) Deprivation group by stage n % Cases a (95% CI) RR b (95% CI) Late stage (Stage IIþ) 2456/ ( ) Late versus early stage 1 (least deprived) 475/ ( ) / ( ) 1.04 ( ) 3 505/ ( ) 1.09 ( ) 4 487/ ( ) 1.11 ( ) 5 (most deprived) 442/ ( ) 1.03 ( ) Stage IV 154/ ( ) Stage IV versus I 1 (least deprived) 19/ ( ) / ( ) 1.47 ( ) 3 38/ ( ) 1.87 ( ) 4 39/ ( ) 2.00 ( ) 5 (most deprived) 24/ ( ) 1.37 ( ) Unknown stage 1725/ ( ) Unknown versus known stage 1 (least deprived) 319/ ( ) / ( ) 1.00 ( ) 3 356/ ( ) 0.99 ( ) 4 338/ ( ) 0.98 ( ) 5 (most deprived) 346/ ( ) 0.95 ( ) a Except for unknown stage, % cases are for those with known stage by deprivation group. b RR is adjusted for age (3 groups), ethnicity (6 groups) and PCT (12 groups). Nottingham City PCT (Table 4). This varies from a 22% increase in risk for women resident in Doncaster PCT [RR ¼ 1.22 (95% CI: )] to 42% for women resident in Bassetlaw PCT and Leicester City PCT [RR ¼ 1.42 (95% CI: ) and RR ¼ 1.42 (95% CI: ), respectively]. Women resident in Rotherham PCT have a significantly lower risk of diagnosis with late-stage breast cancer [RR ¼ 0.67 (95% CI: )]. Risk of diagnosis with Stage IV breast cancer is highest for women resident in Bassetlaw and Sheffield [RR ¼ 5.59 (95% CI: ) and RR ¼ 3.11 (95% CI: ), respectively]. Discussion Main findings of this study This study has identified evidence of inequalities in stage at presentation of breast cancer by age, deprivation, ethnicity and geography and these persist after adjustment for potential confounders. Cancer registry data are a timely and reliable source with a good level of completeness, all of which is necessary to inform appropriate targeted policies. Given the limited current research occurring in the UK to inform health-care policy on inequalities in breast cancer stage at diagnosis, this study has demonstrated the value of registry data to bridge a gap in knowledge, quantifying the extent of risk factors at a regional level. What is already known on this topic Eaker et al. 15 showed that as a consequence of women over 70 years old being offered mammography screening much less often than younger women, they are more likely to present with late-stage disease, and this is associated with poorer survival. They also found that breast cancers in older women were more likely to be un-staged due to missing information on lymph node involvement as a consequence of lower treatment activity. Similar findings have been reported by Barchelli and Balzi 16 who found that 43% of older patients (80) had no axillary staging, compared with only 3 4% of younger patients (,70). The underlying socioeconomic factors for gradients in late-stage breast cancer incidence are complex and multidimensional as lower screening levels have been observed for deprived women in England; 17,18 and some studies have reported a significant association between measures of socioeconomic status and stage at diagnosis, 2,8,19 24 while others have reported the absence of such an association. 4,25 Research suggests that deprivation alone is not sufficient to explain variation in breast cancer stage among women, but that a large proportion of this observed effect can be

5 402 JOURNAL OF PUBLIC HEALTH Table 3 Distribution and risk of stage at diagnosis by ethnic group ( ) Ethnic group by stage n % Cases a (95% CI) RR b (95% CI) Late stage (Stage IIþ) 9437/ ( ) Late versus early stage White British 8882/ ( ) 1.00 White Other 255/ ( ) 1.12 ( ) Mixed 15/ ( ) 1.26 ( ) Asian/Asian British 184/ ( ) 1.11 ( ) Black/Black British 62/ ( ) 1.28 ( ) Chinese/Other Ethnic 39/ ( ) 1.26 ( ) All Other 555/ ( ) 1.15 ( ) Stage IV 659/ ( ) Stage IV versus I White British 633/ ( ) 1.00 White Other 11/ ( ) 0.89 ( ) Mixed Asian/Asian British 9/ ( ) 1.19 ( ) Black/Black British Chinese/Other Ethnic All Other 26/ ( ) 1.07 ( ) Unknown stage 6576/ ( ) Unknown versus known stage White British 6278/ ( ) 1.00 White Other 154/ ( ) 1.03 ( ) Mixed 5/ ( ) 1.02 ( ) Asian/Asian British 89/ ( ) 0.97 ( ) Black/Black British 29/ ( ) 0.98 ( ) Chinese/Other Ethnic 21/ ( ) 0.96 ( ) All Other 298/ ( ) 1.00 ( ) represents small numbers suppressed. a Except for unknown stage, % cases are for those with known stage by ethnic group. b RR is adjusted for age (3 groups), deprivation (5 groups) and PCT (12 groups). accounted for by ethnic variation. 19,22 There are several studies in US populations, 2,4,6,11 but few comparable studies were identified for the UK. Chiu 26 reports that the lack of systematic monitoring of variation between ethnic groups means that, for example, the current record of uptake rates for mammography screening can easily mask inequality of access among certain groups of women. There is an ongoing debate in the literature as to the extent to which differences between areas reflect compositional or contextual effects, and the evidence to date is that both are important. 27 Research suggests that despite uniform breast cancer care programmes within regions and within countries, geographical differences in breast cancer survival and stage at diagnosis persist. 3,5,15,17 In summary, research to date shows that the role of age, deprivation, ethnicity and geography in defining breast cancer inequalities is not clear cut, but that some sort of potentially avoidable inequality gradients may be in operation. What this study adds The results by age show that risk of diagnosis with late-stage breast cancer is greatest for those women outside the screening age group. Given the natural history of breast cancer, characterized by more aggressive tumours among premenopausal women, 20 we would have expected to see greater risk of diagnosis with late-stage breast cancer in the,50 age group than in the.70 age group. However, both age groups show similar risk, which implies that these results may be due to differences in access to screening. Women in the three most deprived quintile groups all have a significantly increased risk of diagnosis with the most advanced stage of breast cancer (Stage IV), compared with the least deprived women, which suggests that deprived women are less likely to participate in the screening programme and that they are more likely to ignore symptoms for longer. 2,17,22,28 Consequently, the delay in diagnosis and treatment affords the most deprived women the poorest prognosis.

6 INEQUALITIES IN BREAST CANCER STAGE AT DIAGNOSIS 403 Table 4 Distribution and risk of stage at diagnosis by PCT of patient residence ( ) PCT by stage n % Cases a (95% CI) RR b (95% CI) Late stage (Stage IIþ) 2456/ ( ) Late versus early stage Nottingham City 92/ ( ) 1.00 Bassetlaw 72/ ( ) 1.42 ( ) Rotherham 84/ ( ) 0.67 ( ) Barnsley 126/ ( ) 1.23 ( ) Sheffield 215/ ( ) 1.31 ( ) Doncaster 152/ ( ) 1.22 ( ) Derbyshire County 383/ ( ) 1.33 ( ) Derby City 100/ ( ) 1.26 ( ) Nottinghamshire County 233/ ( ) 0.85 ( ) Lincolnshire County 375/ ( ) 1.29 ( ) Leicestershire County 455/ ( ) 1.30 ( ) Leicester City 169/ ( ) 1.42 ( ) Stage IV 154/ ( ) Stage IV versus I Nottingham City 5/ ( ) 1.00 Bassetlaw 12/ ( ) 5.59 ( ) Rotherham Barnsley 8/ ( ) 1.98 ( ) Sheffield 20/ ( ) 3.11 ( ) Doncaster Derbyshire County 22/ ( ) 2.07 ( ) Derby City 8/ ( ) 2.61 ( ) Nottinghamshire County 34/ ( ) 2.43 ( ) Lincolnshire County 24/ ( ) 2.19 ( ) Leicestershire County 9/ ( ) 0.80 ( ) Leicester City 6/ ( ) 1.88 ( ) Unknown stage 1725/ ( ) Unknown versus known stage Nottingham City 64/ ( ) 1.00 Bassetlaw 27/ ( ) 1.07 ( ) Rotherham 62/ ( ) 1.06 ( ) Barnsley 89/ ( ) 0.93 ( ) Sheffield 256/ ( ) 0.78 ( ) Doncaster 47/ ( ) 1.08 ( ) Derbyshire County 224/ ( ) 0.95 ( ) Derby City 58/ ( ) 0.96 ( ) Nottinghamshire County 308/ ( ) 0.85 ( ) Lincolnshire County 342/ ( ) 0.85 ( ) Leicestershire County 191/ ( ) 1.02 ( ) Leicester City 57/ ( ) 1.06 ( ) represents small numbers suppressed. a Except for unknown stage, % cases are for those with known stage by PCT. b RR is adjusted for age (three groups), deprivation (five groups) and ethnicity (six groups). The results by ethnic group show that risk of diagnosis with late-stage breast cancer is greater for those women from ethnic minority groups than for White British women. In particular, the Black/Black British and Chinese/Other ethnic groups had a greater than 25% increased risk of diagnosis with late-stage breast cancer. In agreement with previous studies, 2,19,22 these results show that deprivation alone is not sufficient to explain variation in cancer stage at diagnosis the increase in risk is not because the ethnic minority women are from deprived

7 404 JOURNAL OF PUBLIC HEALTH areas, but is due to some other factor(s). Ethnic differentials in staging may therefore be due to lower screening rates. 11,26 Further investigation is warranted into cultural factors, such as beliefs and attitudes and knowledge about cancer, all of which are known to vary by ethnic group, and may be associated with differential access to screening mammography. 29 The results by PCT of patient residence illustrate barriers to access in different settings. Variation between PCTs is a concern for the NHSBSP: Patnick, 30 Director of NHS Cancer Screening Programmes, attributes variation in screening coverage between PCTs to generally deprived, hard to reach, mobile and inner city populations, and reducing this variation is a key target of the Cancer Reform Strategy. 10 The results for inequalities by PCT of patient residence show some of the greatest risks found in this study. In particular, women resident in Bassetlaw PCT are more than five times more likely, and women resident in Sheffield PCT are more than three times more likely to be diagnosed with a Stage IV breast cancer than women resident in Nottingham City PCT. In agreement with the study by Barry and Breen, 5 these results show the importance of place of residence in the risk of a late-stage cancer diagnosis, and suggest that those areas where there is greater risk of diagnosis with late-stage breast cancer are areas where screening uptake is lower. Therefore, alongside compositional effects (age and ethnicity), contextual effects also have a significant impact on risk of diagnosis with late-stage breast cancer and therefore on coverage for routine screening that impacts over and above individual level variables. 31 Limitations of this study Missing data are the main potential for bias in this study. Given that the reporting of stage is not 100% complete in the Registry dataset, it is likely that this will have introduced some bias into the results as staging data does not appear to be missing entirely at random. For example, after adjusting for ethnic group, deprivation and PCT, the risk of a breast cancer not being staged in the.70 age group is significantly reduced (Table 1), which may introduce some bias into the results by age. Missing data were also a significant source of bias for the ethnicity field. Despite using a 10-year cohort to examine inequalities by ethnicity, the ethnic minority groups had persistently small numbers, with resulting poor validity and poor precision in determining the risk of late stage at diagnosis for these groups. Additionally, this study was not able to control for all established risk factors for breast cancer, such as parity, family history, co-morbidities, use of oral contraceptives, hormone replacement therapy, obesity, alcohol consumption and smoking, all of which have been discussed as potential risk factors or modifiers. 28 This study also did not have information on whether individual women had been screened. A linked dataset of registry and screening data would permit an analysis of inequalities in screening uptake, and would make a useful comparison with the results from this study. Conclusion These results show that despite a well-established national screening programme, inequalities in late presentation of breast cancer persist. The study emphasizes the need to address both compositional and contextual inequalities in the presentation of breast cancer. It remains unclear whether the reason for the increased risk for presentation with late-stage breast cancer is due to delay in diagnosis or to differing biology of cancers in the groups identified with heightened risk. 23 However, the results do give public health practitioners the ability to identify communities that may benefit from targeted interventions to reduce their burden of delayed breast cancer diagnosis and treatment. 22 Acknowledgements Thanks to Trent Cancer Registry for their support in this study and allowing the use of their data, in particular Paul Silcocks for statistical advice. References 1 Reddy M, Given-Wilson R. Screening for breast cancer. Surgery 2004;22(7): Merkin SS, Stevenson L, Powe N. Geographic socioeconomic status, race, and advanced-stage breast cancer in New York city. Am J Public Health 2002;92(1): Wang F, McLafferty S, Escamilla V et al. Late-stage breast cancer diagnosis and health care access in Illinois. Prof Geographer 2008;60(1): Lantz PM, Mujahid M, Schwartz K et al. The influence of race, ethnicity, and individual socioeconomic factors on breast cancer stage at diagnosis. Am J Public Health 2006;96(22): Barry J, Breen N. The importance of place of residence in predicting late-stage diagnosis of breast or cervical cancer. Health Place 2005;11: Sassi F, Luft HS, Guadagnoli E. Reducing racial/ethnic disparities in female breast cancer: screening and stage at diagnosis. Am J Public Health 2006;96(12):

8 INEQUALITIES IN BREAST CANCER STAGE AT DIAGNOSIS NHSBSP. NHS Cancer Screening Programmes. Annual review, nhsbsp-annualreview2007-lower-resolution.pdf (15 August 2008, date last accessed). 8 Adams J, White M, Foreman D. Are There socioeconomic gradients in stage and grade of breast cancer at diagnosis? Cross sectional analysis of UK cancer registry data. BMJ 2004;329: Sutton S, Bickler G, Sancho-Aldridge J et al. Prospective study of predictors of attendance for breast screening in inner London. J Epidemiol Community Health 1994;48: Department of Health. Cancer Reform Strategy, gov.uk/en/publicationsandstatistics/publications/publicationspolicy AndGuidance/dh_ (12 August 2008, date last accessed). 11 Jacobellis J, Cutter G. Mammography screening and differences in stage of disease by race/ethnicity. Am J Public Health 2002;92(7): ONS. Presenting Ethnic and National Groups Data, ethnic-interim/presenting-data/index.html (14 August 2008, date last accessed). 13 APHO. Commonly Used Public Health Statistics and their Confidence Intervals. Technical Briefing 3, resource/item.aspx?rid=48457 (8 July 2008, date last accessed). 14 Zou G. A modified Poisson regression approach to prospective studies with binary data. Am J Epidemiol 2004;159(7): Eaker S, Dickman PW, Bergkvist L et al. Differences in management of older women influence breast cancer survival: results from a population-based database in Sweden, PLoS Med 2006;3(3): Barchelli A, Balzi D. Age at diagnosis, extent of disease and breast cancer survival: a population-based study in Florence, Italy. Tumori 2000;86: Maheswaran R, Pearson T, Jordan H et al. Socioeconomic deprivation, travel distance, location of service, and uptake of breast cancer screening in North Derbyshire. J Epidemiol Community Health 2006;60: Banks E, Beral V, Cameron R et al. Comparison of various characteristics of women who do and do not attend for breast cancer screening. Breast Cancer Res 2002;4(1). 19 Baquet CJ, Commisky P. Socioeconomic factors and breast carcinoma in multicultural women. Cancer 2002;88(5 Suppl.): Dalton SO, Düring M, Ross L et al. The relation between socioeconomic and demographic factors and tumour stage in women diagnosed with breast cancer in Denmark, Br J Cancer 2006;95: Kaffashain F, Godward S, Davies T et al. Socioeconomic effects on breast cancer survival: proportion attributable to stage and morphology. Br J Cancer 2003;89: Mackinnon JA, Duncan RC, Huang Y et al. Detecting an association between socioeconomic status and late stage breast cancer using spatial analysis and area-based measures. Cancer Epidemiol Biomarkers Prev 2007;16(4): Macleod U, Ross S, Gillis C et al. Socio-economic deprivation and stage of disease at presentation in women with breast cancer. Ann Oncol 2000;11: McCarthy EP, Burns RB, Freund KM et al. Mammography use, breast cancer stage at diagnosis, and survival among older women. J Am Geriatr Soc 2000;48: Thomson CS, Hole DJ, Twelves CJ et al. Prognostic factors in women with breast cancer: distribution by socioeconomic status and effect on differences in survival. J Epidemiol Community Health 2001;55: Chiu LF. Inequalities of Access to Cancer Screening: A Literature Review, Cancer Screening Series, Vol. 1, Sheffield, NHS Cancer Screening Programmes, Gatrell A, Thomas C, Bennett S et al. Understanding health inequalities: locating people in geographical and social spaces. In: Graham H (ed). Understanding Health Inequalities. Buckingham: Open University Press, Zackrisson S, Andresson I, Manjer J et al. Non-attendance in breast cancer screening is associated with unfavourable socio-economic circumstances and advanced carcinoma. Int J Cancer 2004;108: Lanin DR, Mathews HF, Mitchell JM et al. Influence of socioeconomic and cultural factors on racial differences in late-stage presentation of breast cancer; JAMA 1998;279(22): Patnick J. Changes to National Cancer Screening Programmes Linked to the Cancer Reform Strategy; UKACR, Graham H. The challenge of health inequalities. In: Graham H (ed). Understanding Health Inequalities. Buckingham: Open University Press, 2004.