Socioeconomic inequalities in prostate cancer survival: A review of the evidence and explanatory factors

Social Science & Medicine 142 (2015) 9e18

Contents lists available at ScienceDirect

Social Science & Medicine journal homepage: www.elsevier.com/locate/socscimed

Review article

Socioeconomic inequalities in prostate cancer survival: A review of the evidence and explanatory factors Jens Klein*, Olaf von dem Knesebeck Department of Medical Sociology, University Medical Center Hamburg-Eppendorf, Martinistr. 52, 20246 Hamburg, Germany

a r t i c l e i n f o

a b s t r a c t

Article history: Received 29 January 2015 Received in revised form 5 June 2015 Accepted 7 July 2015 Available online 11 July 2015

Although survival rates after prostate cancer diagnosis have improved in the past two decades, survival analyses regarding the socioeconomic status (SES) suggest inequalities indicating worse prognosis for lower SES groups. An overview of the current literature is lacking and moreover, there is an ongoing discussion about the underlying causes but evidence is comparatively sparse. Several patient, disease and health care related factors are discussed to have an important impact on disparities in survival. Therefore, a systematic review was conducted to sum up the current evidence of survival inequalities and the contribution of different potential explanatory factors among prostate cancer patients. The PubMed database was screened for relevant articles published between January 2005 and September 2014 revealing 330 potentially eligible publications. After systematic review process, 46 papers met the inclusion criteria and were included in the review. About 75% of the studies indicate a significant association between low SES and worse survival among prostate cancer patients in the fully adjusted model. Overall, hazard ratios (low versus high SES) range from 1.02 to 3.57. A decrease of inequalities over the years was not identified. 8 studies examined the impact of explanatory factors on the association between SES and survival by progressive adjustment indicating mediating effects of comorbidity, stage at diagnosis and treatment modalities. Eventually, an apparent majority of the obtained studies indicates lower survival among patients with lower SES. The few studies that intend to explain inequalities found out instructive results regarding different contributing factors but evidence is still insufficient. © 2015 Elsevier Ltd. All rights reserved.

Keywords: Socioeconomic status Survival inequalities Explanatory factors Mediating effects Prostate cancer Systematic review

1. Introduction Prostate cancer has emerged as one of the most prevalent cancers worldwide (Bray et al., 2013). Particularly in Europe, North America and Australia prostatic neoplasms are widespread. In terms of inequalities studies mostly have shown higher incidence rates among higher socioeconomic status (SES) groups (Aarts et al., 2010; Clegg et al., 2009; Faggiano et al., 1997; Gilligan, 2005; National Cancer Intelligence Network, 2009; Shafique et al., 2012a). Studies analysing population-based prostate cancer mortality data found inconsistent associations with SES (Albano et al., 2007; Elstad et al., 2011; Krieger et al., 2012; Menvielle et al., 2008). Therefore, it is remarkable that previous overviews highlighted lower survival rates and higher excess mortality especially

* Corresponding author. E-mail addresses: [email protected] (J. Klein), [email protected] (O. von dem Knesebeck). http://dx.doi.org/10.1016/j.socscimed.2015.07.006 0277-9536/© 2015 Elsevier Ltd. All rights reserved.

for lower status groups among prostate cancer patients (Coleman et al., 2004; Gilligan, 2005; Kogevinas and Porta, 1997; Kravdal, 2000; Quaglia et al., 2013; Woods et al., 2006) whilst overall, the survival concerning prostate cancer has improved (De Angelis et al., 2014). One can assume that the identification of potential explanatory factors could point to reduce inequalities, and furthermore the number of avoidable deaths (Ellis et al., 2012). Moreover, as survival is considered as a potential quality of care indicator for prostate cancer, improvements in health care could be deduced (Spencer et al., 2003). Information about underlying causes to explain socioeconomic differences in prostate cancer survival and case fatality is sparse. Possible explanations can be divided into three groups (Auvinen and Karjalainen, 1997; Frederiksen et al., 2009; Woods et al., 2006): factors linked to the tumour (stage at diagnosis, biological characteristics), the patient (comorbidity, health behaviour, psychosocial factors) and the health care (treatment, medical expertise, screening).

10

J. Klein, O. von dem Knesebeck / Social Science & Medicine 142 (2015) 9e18

There is an ongoing discussion about the role of health care management as a contributing factor to social disparities in survival among prostate cancer patients (Chu and Freedland, 2010). For instance, Lyratzopoulos et al. (2010) found out that patients from England with lower SES were less likely to receive radical surgery or radiotherapy (and watchful waiting more likely) than those from least deprived SES groups, also when age, disease stage, period of diagnosis, tumour type or hospital (but not comorbidity) were taken into consideration. Data from other studies conducted in England, Australia and the USA also has shown that socioeconomically disadvantaged men have a decreased likelihood of having radical prostatectomy compared to patients with lower SES who received more often hormone therapy, active surveillance, watchful waiting and partly radiation (Fairley et al., 2009; Hayen et al., 2008; Krupski et al., 2005). Furthermore, screening uptake is lower among prostate cancer patients with low SES in different health care settings (Ross et al., 2011; Williams et al., 2011). Also, stage at diagnosis is discussed extensively as an explanatory factor (Auvinen and Karjalainen, 1997; De Angelis et al., 2014; Woods et al., 2006). Recent Anglo-American studies have shown an association between lower SES and an advanced stage at diagnosis for prostate cancer while adjusting for several covariates (Clegg et al., 2009; Lyratzopoulos et al., 2013). Moreover, patient factors as comorbidity or health behaviour can interact with treatment modalities or disease stage and additionally have a potential impact on inequalities in survival (Berglund et al., 2011; Hall et al., 2005b). Berglund et al. (2011) report in their study an increased likelihood of surveillance as treatment among patients with severe comorbidity while radical prostatectomy was significantly less likely to be offered. Furthermore, all cause and competing cause mortality but not prostate cancer specific mortality was higher in patients with severe comorbidity. However, despite the increasing efforts in research about social disparities in prostate cancer survival, the latest comprehensive (non-systematic) overview dates back to nearly one decade (Woods et al., 2006). It comprises 14 studies reporting data for prostate cancer patients, and moreover, it is still unclear which are the most relevant factors contributing to the differences. Therefore, a systematic review was conducted to address two major topics: first, to give a current overview of the studies and their evidence about the association between socioeconomic status and prostate cancer survival since 2005, and second, to work out which explanatory factors contribute to these differences following Woods et al. (2006) and Auvinen and Karjalainen (1997). Thus, patient, disease and health care factors are considered as potential mediators. 2. Methods A systematic review in the PubMed database was performed on the basis of the PRISMA guidelines (Moher et al., 2009) using a combination of following keywords in title and abstract: socio*, inequalit*, income, education*, occupation*, survival, excess mortality, case fatality, prognosis, prostate, cancer, common cancer*, major cancer*, cancer site*. The search strategy was completed by the two MeSH Terms ‘prostatic neoplasms’ and ‘survival analysis’. Publications that have been released between January 2005 and September 2014 were included as the last comprehensive review examining social inequalities in cancer survival including prostate cancer was published online 2005 (Woods et al., 2006). Additionally, the bibliographic references of the eligible studies were screened for further relevant publications. For more detailed information about the search strategy see Appendix I. Studies were rated as eligible if they (1) were written in English or had English abstract, (2) were published or in-press in a peer reviewed journal, (3) reported data from a primary study and not

an editorial or review, (4) were prospective or retrospective cohort studies that analysed survival, case fatality or excess mortality among prostate cancer patients, (5) introduced indicators of SES as predictor or covariate that enables to identify the impact of SES on survival. Studies that only adjust for SES without explicitly reporting the effects on prostate cancer survival were excluded. Furthermore, SES had to be determined by indicators of education, income or occupational position on individual level or regional level indicated by, for instance, census tracts. Analyses using macro-economic factors on country level were not enclosed. Studies and research that focus solely on ethnic or racial disparities were also excluded, just as studies only referring to insurance status. Data was extracted regarding author and publication date, location of the study, period of diagnosed cancer cases, sample size, SES indicators, type of measurement, adjusted variables and main findings in terms of survival or risk of death of patients with low versus high SES. In a second step, the extracted studies were screened for potential explanatory factors contributing to social inequalities in survival that were identified by progressive adjustment in multivariate analyses. To calculate the percentage reducing contribution of these factors to inequalities in survival, the change of hazard ratio (HR) or relative excess risk (RER) was assessed by using the formula: ([HR/RER Basic Model e HR/RER Basic model þ explanatory factors])/ [HR/RER Basic model  1])  100 (Louwman et al., 2010; Skalicka et al., 2009). Basic model was the model only adjusted for age, race/ ethnicity, year of diagnosis and in one case for stage. A meta-analysis was not conducted as the included studies showed a considerable methodological heterogeneity regarding their designs including varying time frames as well as measurement instruments to capture potential predictors and outcomes. In addition, the performed statistical analyses varied largely, the number and quality of considered confounders were diverse, and the reported effect measures were heterogeneous. 3. Results The PubMed search generated 330 publications that were screened by title and abstract resulting in 78 potential relevant articles. Of these, 40 were included in this review after extensive full-text screening. Main reasons for exclusion were that indicators of SES were missing in the analyses or SES was just introduced as confounding variable without presenting its impact on survival, no survival analyses among a patient cohort were conducted, no prostate cancer but other cancer sites were examined or the paper did not contain original data. For more information about the study selection see Appendix II. Additionally, 6 studies were identified by scanning the reference lists resulting in 46 studies in total that were included in the review (Table 1). Most of the studies were conducted in the USA (n ¼ 15; thereof one study in USA and Canada) and Europe (n ¼ 20), i.e. UK (n ¼ 9), Netherlands (n ¼ 3), Sweden (n ¼ 3), Ireland, Germany, Denmark, Finland, Switzerland (in each case n ¼ 1). Further countries of investigation were Australia (n ¼ 6), New Zealand (n ¼ 2), Colombia (n ¼ 1), Japan (n ¼ 1) and Taiwan (n ¼ 1). 33 studies indicate a significant association between SES and survival among prostate cancer patients (fully adjusted model if multivariate analyses were conducted) (Aarts et al., 2013a, 2013b; Australian Institute of Health and Walfare (2013); Berglund et al., 2012; Bravo et al., 2014; Burns et al., 2014; Byers et al., 2008; Chang et al., 2012; Du et al., 2006; Freeman et al., 2011; Hall et al., 2005a; Hellenthal et al., 2010; Hussain et al., 2008; Jansen et al., 2014; Jeffreys et al., 2009; Li et al., 2012; Louwman et al., 2010; Luo et al., 2013; Marsa et al., 2008; Niu et al., 2010; Prasad et al., 2014; Rachet et al., 2010; Robbins et al., 2007a, 2007b;

J. Klein, O. von dem Knesebeck / Social Science & Medicine 142 (2015) 9e18

11

Table 1 Summary of studies examining the impact of socioeconomic status (SES) on survival among prostate cancer patients. Author, year

Sample characteristicsa

SES indicators

Type of measurement

Adjustment

Findingsb, c, d (Survival or risk of death of patients with low versus high SES)

Aarts et al., 2013a

Netherlands, 1991 e2008, n ¼ 271

Individual level (education)

Age, year of diagnosis, stage, comorbidity, health behaviour

3.10 (1.60e5.80)e,

Aarts et al., 2013b

Netherlands, 1998 e2008, n ¼ 11,086

Area level (composite score of income and value of the house)

Hazard ratio (HR) and 5-year survival (%) Hazard ratio and 10-year survival (%)

Year of diagnosis, comorbidity, treatment

Australian Institute of Health and Welfare (2013)

Australia, 2006e2010, n ¼ not available

Area level (Index of Relative Socioeconomic Disadvantageg)

5-year relative survival (%)

e

Berglund et al., 2012

Sweden, 1997e2006, n ¼ 4304 Colombia, 1995e2004, n ¼ 3999 Ireland, 1998e2009, n ¼ 26,816

Individual level (occupation) Area level (income)

Hazard ratio

Age, calendar period, clinical subgroups, comorbidity Age, calendar period

1.08 (0.65e1.77) to 1.72 (0.90 e3.31) (Age 59) 1.21 (1.01e1.64) to 1.46 (1.22 e1.74) (Age 60e74) 1.13 (0.96e1.34) to 1.35 (1.10 e1.64) (Age 75) (stratified analyses for localized and regional stage) Prostate cancer had significantly higher survival in the highest SES quintile than for any other quintile (no exact data is documented). 1.43 (1.01e2.04)

Byers et al., 2008

USA, 1997, n ¼ 4332

Area level (education and income)

Hazard ratio

Chang et al., 2012

Taiwan, 2002, n ¼ 1247

Individual (income) and area level (income)

Hazard ratio and 5year survival (%)

Comorbidity, stage, treatment, hospital characteristics

Du et al., 2006

USA, 1992e1999, n ¼ 61,228

Area level (education, poverty level, income and composite score)

Hazard ratio and 3-, 5- and 10-year survival (%)

Age, race/ethnicity, year of diagnosis, region, grade, stage, treatment

Dutta Roy et al., 2005

England, 1986e2000, n ¼ 33,280 (all periods combined) USA, 1986e1990, n ¼ 833

Area level (Townsend Indexg)

1- and 5-year relative survival (%)

e

Area level (composite score of poverty, unemployment, female-headed households, education) Area level (Index of Relative Socioeconomic Disadvantage) Area level (New Zealand Deprivation Indexg) Area level (composite score of income, education, poverty, house value) Individual level (education) Area level (The German Index of Multiple Deprivationg)

Hazard ratio

Age, race/ethnicity, stage, tumour differentiation, health care setting comorbidity, treatment

2.37 (1.76e3.18)

Hazard ratio

Age, calendar period, race/ ethnicity, marital status, comorbidity, treatment

1.34 (1.10e1.64)

Hazard ratio

1.02 (not significant; CI not reported)

Hazard ratio

Age, race/ethnicity, travel to primary care and cancer centre, stage Age, race/ethnicity

Hazard ratio

Age, time period

Relative excess risk (RER) (3 month, 1 and 5 years) and 5year survival (%)

Age, stage

Bravo et al., 2014 Burns et al., 2014

Freeman et al., 2011

Hall et al., 2005a

Australia, 1982e2001, n ¼ 14,123

Haynes et al., 2008

New Zealand, 1994 e2004, n ¼ 25,078

Hellenthal et al., 2010

USA, 1996e2005, n ¼ 123,953

Hussain et al., 2008

Sweden, 1990e2004, n ¼ 35,256 Germany, 1997e2006, n ¼ 50,354

Jansen et al., 2014

Area level (Carstairs Indexg)

Excess mortality rate ratio Odds ratio

Age, stage, grade, marital status, region, method of presentation, health provider, smoking status Age, race/ethnicity, comorbidity, stage

f

3.57 (2.37e5.40) Overall: 1.20 (1.09e1.33) 3-year: 1.26 (1.18e1.52) 7-year: 1.34 (1.18e1.52) In poverty or undereducated: 1.22 (1.03e1.44) Both in poverty and undereducated: 1.14 (0.95 e1.37) Low individual and neighbourhood SES: 1.80 (0.65e4.95) (Age <65 years) 2.54 (1.19e5.44) (Age 65 years) Education: 1.39 (1.20e1.61) Poverty: 1.31 (1.13e1.52) Income: 1.37 (1.20e1.64) Composite score: 1.40 (1.20 e1.64) 1-year: 89 vs. 85 5-year: 62 vs. 75

2.20 (1.38e3.50) (radical prostatectomy) 2.21 (1.66e2.95) (beam radiation therapy) 1.59 (1.43e1.75) 3-month:1.55 (1.46e1.84) 1-year conditional on 3 month:1.62 (1.30e2.02) 5-year conditional on 1 year: 1.13 (0.99e1.28) (Relative survival of patients living in the most deprived district was compared to survival of patients living in all other districts) (continued on next page)

12

J. Klein, O. von dem Knesebeck / Social Science & Medicine 142 (2015) 9e18

Table 1 (continued ) Author, year

Sample characteristicsa

SES indicators

Type of measurement

Adjustment

Findingsb, c, d (Survival or risk of death of patients with low versus high SES)

Jeffreys et al., 2009

New Zealand, 1994 e2003, n ¼ 9632

Area level (New Zealand Deprivation Index)

Age

76 vs. 88 Deprivation gap: ¡15.00 (¡27.00 to ¡0.03)

Li et al., 2012

Sweden, 1990e2008, n ¼ 73,159

Age, marital status, immigration status, urban/rural status, mobility, comorbidity, area and individual SES adjusted for each other

Louwman et al., 2010

Netherlands, 1997 e2006, n ¼ 9987

1.35 (1.26e1.45) (individual/ income) 0.98 (0.92e1.04) (individual/ education) 1.19 (1.10e1.29) (area/ composite score) 1.36 (1.10e1.70)

Luo et al., 2013

Australia, 1999e2007, n ¼ 39,852

Marsa et al., 2008

Denmark, 1994e2003, n ¼ 7995

Individual (income and education) and area level (composite score of education, income, unemployment, social welfare assistance) Area level (composite score of income and value of the house) Area level (Index of Relative Socioeconomic Disadvantage) Individual level (education, income, social class, size of dwelling, housing tenure)

Relative survival rate (%) (5-year), deprivation gaph (%) Odds ratio

McPhail et al., 2013

England, 2006e2008, n ¼ 11,204

Area level (income)

Miki et al., 2014

Japan, 1990e2009, n ¼ 732 USA, 1997e2006, n ¼ 788 USA, 1993e1999, n ¼ 41,999 Australia, 1995e2000 (time of treatment), n ¼ 1984 (radical prostatectomy only) USA, 2004e2007, n ¼ 41,275

Area level (Japanese Deprivation Indexf) Area level (income)

Movsas et al., 2014 Niu et al., 2010 Papa et al., 2014

Prasad et al., 2014

Area level (poverty) Areal level (Index of Relative Socioeconomic Disadvantage) Area level (education and income)

Hazard ratio and 1year survival (%)

Age, comorbidity

Hazard ratio

Age, year of diagnosis, stage, country of birth, region

1.29 (1.12e1.48)

1- and 5-year relative survival (%)

e

Excess mortality rate ratio (0e12 months) Hazard ratio

Age, presentation (emergency/ non-emergency), comorbidity, stage Age, area, population density, occupation, health behaviour e

1-year/5-year relative survival (in %): 88 vs. 92/47 vs. 59 (education) 88 vs. 93/47 vs. 56 (income) 92 vs. 92/47 vs. 59 (social class) 91 vs. 96/51 vs. 52 (housing tenure) 80 vs. 92/38 vs. 55 (size of dwelling) (Significance calculated from the data; p < 0.001) 1.0 (0.9e1.2) (6e12 Months)

1.10 (0.76e1.59)

Age, stage, race/ethnicity

1.39 (1.24e1.56)

Age, region, insurance status, PSA level, Gleason score, tumour stage, biochemical recurrence Age, year of diagnosis, marital status, comorbidity, race/ ethnicity, region, population density, tumour grade, treatment

1.11 (0.55e2.22)

Hazard ratio Hazard ratio and 5year survival (%) Hazard ratio

Hazard ratio

Pokhrel et al., 2010

Finland, 1971e2005, n ¼ 59,308

Individual level (education)

Relative risk and 5year survival (%)

Age, stage

Rachet et al., 2010

England, 1996e2006, n ¼ 265,753 (all periods combined) Switzerland, 1995 e2005, n ¼ 2738

Area level (income)

Deprivation gap (%) (1-year relative survival) Hazard ratio and 5year survival (%)

e

Rapiti et al., 2009

Robbins et al., 2007a

USA, 1995e2004, n ¼ 116,916 (White and Asian men)

Robbins et al., 2007b

USA, 1995e2004, n ¼ 122,374 (White and Black men)

Rowan et al., 2008

England and Wales, 1986e1999, n z 201,000

Individual level (occupation)

Area level (composite score of income, education, poverty, house value) Area level (composite score of income, education, poverty, house value) Area level (income)

1.22 (0.75e2.00)

Stratified for NCCN (National Comprehensive Cancer Network) risk group (significant results): 1.16 (1.02e1.33) (education/ high risk) 1.52 (1.25e1.85) (income/ intermediate risk) Better survival in higher SES groups (80.3% vs. 87.4%), also indicated by relative risk (exact data and significance not documented). ¡2.90 (p < 0.001)

Age, nationality, period of diagnosis, stage, grade, method of detection, sector of care, treatment Age, race/ethnicity, year of diagnosis, stage, grade, treatment

1.20 (0.80e1.60)

Hazard ratio

Age, race/ethnicity, year of diagnosis, stage, grade, treatment

1.51 (1.39e1.64)

Deprivation gap (%) (1-, 5- and 10-year relative survival)

e

1-year: ¡4.10 (¡4.90 to ¡3.30) 5-year: ¡7.20 (¡9.00 to ¡5.50) 10-year: ¡4.90 (¡7.30 to ¡2.50)

Hazard ratio

1.51 (1.39e1.66)

J. Klein, O. von dem Knesebeck / Social Science & Medicine 142 (2015) 9e18

13

Table 1 (continued ) Author, year

Sample characteristicsa

SES indicators

Type of measurement

Adjustment

Findingsb, c, d (Survival or risk of death of patients with low versus high SES)

Schwartz et al., 2009

USA, 1988e1992, n ¼ 8679

Area level (composite score of education, poverty, occupation)

Hazard ratio

Age, race/ethnicity, grade, treatment

Shack et al., 2007

Scotland, 1986e2000, n ¼ 9370

Deprivation gap (%) (5-year relative survival)

e

Shafique et al., 2012

Scotland, 2000e2007, n ¼ 897

Area level (Carstairs Index (1986e1995) and Scottish Index of Multiple Deprivationg (1996e1999)) Area level (Scottish Index of Multiple Deprivation)

1.39 (1.14e1.69) (localized stage) 1.37 (0.91e2.04) (regional stage) 6.90 (10.30 to 3.40)

Relative excess risk and 5-year relative survival (%)

Age, grade, two systemic inflammation based prognostic scores

Shafique et al., 2013

Scotland, 2000e2006, n ¼ 744

Area level (Scottish Index of Multiple Deprivation)

Relative excess risk and 5- and 10-year relative survival (%)

Shafique and Morrison 2013

Scotland, 1991e2007, n ¼ 15,292

Area level (Scottish Index of Multiple Deprivation)

Slogett et al., 2007

England and Wales, 1981e1997, n ¼ 1714

Area level (social class, housing tenure, car access and Carstairs Index)

Relative excess risk and deprivation gap (%) (1-, 3- and 5-year relative survival) Relative excess mortality

Age, grade, PSA level, previous inpatient bed days, systemic inflammation based prognostic score Age

Tewari et al., 2006 (see also Tewari et al., 2005) Tewari et al., 2009

USA, 1980e1997, n ¼ 3159

Area level (income)

USA, 1990e2000, n ¼ 2046

Area level (income)

Relative risk and 15-year survival rate (%) Hazard ratio

Tyson et al., 2013

USA, 1988e2003, n ¼ 115,922 USA, 2000e2009, n ¼ 132,887e132,993

Area level (income)

Hazard ratio

Area level (composite indices of occupation, house and car ownership, poverty, unemployment, education, house value, overcrowding) Area level (education, occupation)

5-year survival rate (%)

Yu et al., 2014a

Yu et al., 2008

Australia, 1996e2000, n ¼ 30,441

Yu et al., 2014b

Australia, 1982e2007, n ¼ 68,686

Zhang-Salomons et al., 2006

USA (1988e1992) and Canada (1989e1993), n ¼ 6843 (Toronto)/ 14,656 (Detroit)

Area level (Index of Relative Socioeconomic Disadvantage) Area level (poverty, income)

Relative excess risk and 5-year relative survival (%) Relative excess risk and 10-year relative survival (%) Relative risk and 5year relative survival (%)

Age, year of follow-up, period of diagnosis, marital status, geographical zone, all other SES indicators

Age, year of diagnosis, race/ ethnicity, comorbidity, grade

1.39 (0.61e3.18) (sampled before diagnosis) 0.75 (0.46e1.24) (sampled after diagnosis) 1.24 (0.81e1.89)

0.84 (0.39e1.82) to 2.61 (1.09 e6.26) (Relative excess risk stratified for different Gleason grades) 0.99 (0.88e1.11) (social class) 1.14 (0.83e1.57) (car access) 1.02 (0.76e1.38) (housing tenure) 1.07 (0.97e1.18) (Carstairs Index) 1.11 (1.04e1.19)

Age, race/ethnicity, stage, insurance status, grade, comorbidity, treatment Age, race/ethnicity, stage, grade, marital status e

2.13 (1.05e4.34)

Age, year of follow-up, stage

1.09 (p ¼ 0.68)

Age, calendar period, geographical remoteness, stage, incidence rate

1.40 (1.29e1.53)

Age

Canada: 1.22 (income), 1.12 (poverty) USA: 2.04 (income), 1.85 (poverty) (p < 0.05)

1.06 (1.02e1.11) No clear trend in survival by SES category was observed (stratified analyses for race/ ethnicity and stage; no exact data is documented).

a

Country, period of diagnosis, sample size. If crude survival analysis and multivariate analysis are conducted only the results from the latter are shown (fully adjusted model only). c If overall survival and prostate cancer-specific survival are outcome variables only the results of the latter are shown. d If multiple calendar periods are analysed only the results for the most current calendar period are shown. e 95% confidence interval (CI) or p-value in parentheses. f Statistically significant values are bold. If the reference category in the analyses is low SES and therefore hazard ratios lower than 1 are potentially documented, the inverse was calculated. g Index of Relative Socio-economic Disadvantage: education, occupation, overcrowding, marital status, children, health, language skills, car ownership, housing costs, internet access; Carstairs Index: social class, car ownership, employment, overcrowding; Townsend Index: employment, overcrowding, car ownership, house ownership; New Zealand Deprivation Index: car access, housing tenure, benefit receipt, employment, income, internet access, single parent families, qualifications, overcrowding; The German Index of Multiple Deprivation: income, employment, education, district revenue, social capital, environment, security; Japanese Deprivation Index: housing tenure, partner and children in household, occupation, employment; Scottish Index of Multiple Deprivation: income and benefits, employment, health, education, access to services and transport, crime rates, housing quality, overcrowding. h A negative value means that survival in the most deprived group is lower than survival in the most affluent group. b

Rowan et al., 2008; Schwartz et al., 2009; Shack et al., 2007; Shafique and Morrison, 2013; Tewari et al., 2005, 2006, 2009; Tyson et al., 2013; Yu et al., 2014b; Zhang-Salomons et al., 2006),

one study also reports this trend without presenting data about significance (Pokhrel et al., 2010), and another study does not indicate significant results regarding the most current calendar

14

J. Klein, O. von dem Knesebeck / Social Science & Medicine 142 (2015) 9e18

period but former periods (Dutta Roy et al., 2005). 11 studies did not verify a significant association in the fully adjusted model (Haynes et al., 2008; McPhail et al., 2013; Miki et al., 2014; Movsas et al., 2014; Papa et al., 2014; Rapiti et al., 2009; Shafique et al., 2013, 2012b; Sloggett et al., 2007; Yu et al., 2014a, 2008). SES was assessed both on individual and, more frequently, on area level. On the majority, hazard ratio was the type of measurement in multivariate analyses ranging from 1.02 to 3.57 (low versus high SES). The type and amount of adjusted control variables vary strongly (e.g. age, race/ethnicity, year of diagnosis, calendar period, region, marital status, health behaviour, comorbidity, stage, grade, treatment). In some cases analyses were stratified for age, stage, grade, time after diagnosis or race/ethnicity. Age was introduced most frequently followed by diseases related factors like stage and grade. Also, year of diagnosis, treatment, comorbidity, and particular in US-studies race/ethnicity were included in a number of analyses. In Table 2, the studies examining potential explanatory factors are summarised. Results of multivariate analyses which calculated the mediating effects of these factors by progressive adjustment are shown. In 8 studies the impact of comorbidity (n ¼ 4), stage (n ¼ 4), treatment (n ¼ 3), method of detection (n ¼ 1), grade (n ¼ 1) and health behaviour (n ¼ 1) was analysed (Aarts et al., 2013a, 2013b; Byers et al., 2008; Jansen et al., 2014; Louwman et al., 2010; Pokhrel et al., 2010; Rapiti et al., 2009; Schwartz et al., 2009). Reductions of the association between SES and survival are ranging from 6% to 25% for comorbidity, 3% to 73% for stage at diagnosis and 22% to 50% for treatment. Lifestyle behaviour as explanatory factor was examined in one study but did not contribute to any reductions of survival inequalities. While the majority of the studies report a reduction after introducing explanatory factors, just in a few cases the associations change to an insignificant level. Two analyses even show an increase of inequality after adjustment. Finally, treatment and tumour stage seems to have the most impact. 4. Discussion This is the first review since 2005 that gives a systematic overview of the evidence about social inequalities in survival among prostate cancer patients. An apparent majority of the studies (about 75%) supports and widens the findings of former reviews and studies investigating the association between SES and survival regarding this neoplasm (Coleman et al., 2004; Kogevinas and Porta, 1997; Schrijvers et al., 1995; Schrijvers and Mackenbach, 1994; Woods et al., 2006). Men with lower SES have an increased risk of death after having been diagnosed with prostate cancer. Moreover, the findings of this review suggest that survival inequalities were not decreasing in the past 10 years when comparing the findings with the studies' results of the past reviews (Kogevinas and Porta, 1997; Schrijvers and Mackenbach, 1994; Woods et al., 2006). In addition, more recent studies that analysed average changes in the deprivation gap regarding different calendar periods do not report any decrease of inequality (Rachet et al., 2010; Shafique and Morrison, 2013), although generally, the survival rates for prostate cancer survival have improved in the past decade (De Angelis et al., 2014). A striking difference between the continents and countries where the studies were conducted could not be identified. The studies that do not reveal significant associations between lower SES and lower survival rates either report a trend towards a disadvantage among deprived groups or no impact of SES. None of the studies indicates a pattern towards a trend of lower survival among higher status groups. Generally, it is difficult to compare the reviewed studies as they differ in methodological aspects. These aspects are also a major concern when estimating the studies' conclusions. Firstly, the

studies differ in the assessment of SES as different single indicators (education, income, occupation, housing etc.) or a composite score is used. Particularly, if deprivation is surveyed using a censusderived or area-based score and not individual data, the determination of the patients' SES is less precise. Therefore social inequalities could be biased and results are even prone to ecological fallacy (Quaglia et al., 2013; Woods et al., 2005). Furthermore, the size and origin of the samples vary between the studies and make comparisons difficult as health care systems can differ in access and supply, like for instance screening programs and treatment opportunities. This also applies to the available outcome measurement (overall survival, cancer-specific survival), the chosen variables to adjust (e.g. stage, grade, year of diagnosis, comorbidity) and the type of measurement conducted (e.g. relative survival rate, hazard ratio) (Dickman and Adami, 2006; Sarfati et al., 2010). Furthermore, this paper summarises the studies that examine the impact of potential contributing factors relating to different patient, disease and health care characteristics to find out explanations for the observed survival inequalities. Especially, comorbidity (Aarts et al., 2013a, 2013b; Byers et al., 2008; Louwman et al., 2010), stage at diagnosis (Byers et al., 2008; Jansen et al., 2014; Pokhrel et al., 2010; Rapiti et al., 2009) and treatment modalities (Aarts et al., 2013b; Rapiti et al., 2009; Schwartz et al., 2009) are shown to have an mediating effect on the association between SES and survival after diagnosis. However, it is important to differentiate the underlying causes. Notably, in health care research it has to be distinguished between access and utilization. Although they often overlap, access is primarily a supply-side factor (availability of resources, waiting time, user charges etc.) while utilization refers mainly to the demandside and reflects the patients' preferences including the level of need (Goddard and Smith, 2001). Thus, a later stage at diagnosis could refer to patient, doctor or system delay (Hansen et al., 2008) which were shown to be potentially associated with lower SES of prostate cancer patients (Macleod et al., 2009; Neal and Allgar, 2005). On the one hand, delay may be due to a worse awareness and appraisal of cancer symptoms, on the other hand due to more barriers regarding access to screening programs (Lyratzopoulos et al., 2013), and health care in general which in turn lead to small chances of incidental findings. It was shown that cancer awareness is lower in more deprived groups (Robb et al., 2009), but a recent study found out that lower SES may predict a longer interval between prostate biopsy and treatment (radical prostatectomy) (Pitman et al., 2012). Indeed, current findings hardly suggest an association between curative treatment delay and survival (Redaniel et al., 2013; van den Bergh et al., 2013). In this context, artificial effects as the lead time and length bias have to be considered while interpreting the results (Auvinen and Karjalainen, 1997; Berry, 2014; de Vries et al., 2010; Dickman and Adami, 2006; Kogevinas and Porta, 1997). Lead time bias increasingly occurred when new screening methods like PSA (prostate specific antigen) testing were introduced. It means that an earlier detection results in an increased survival but not in a delayed death. So, no additional lifetime is gained and the screening just pretend to increase survival time without altering the natural course of the disease. As the uptake of screening programs among prostate cancer patients differs by SES (Clarke, 2008; Ross et al., 2011; Williams et al., 2011), inequalities could be overestimated. Length bias refers to slow-growing tumours that are more likely to be detected by screening programs than fast-growing ones. It suggests that these tumours have an improved survival time as they already had an inherently favourable prognosis quite apart from any advantage of early detection (Berry, 2014; de Vries et al., 2010). In this context, the stage migration bias, also known as the Will Rogers phenomenon, has to be regarded (Albertsen et al., 2005; Dickman

J. Klein, O. von dem Knesebeck / Social Science & Medicine 142 (2015) 9e18

15

Table 2 Summary of studies examining the mediating effects of potential explanatory factors on inequalities in survival. Author, year

Explanatory factors examined

Further adjustment and stratification

Percentage change of social inequalities in survivala (change of ratio in parentheses)

Description of results

Aarts et al., 2013a

Comorbidity (Charlson Comorbidity Index), health behaviour (smoking, physical activity, alcohol consumption) Comorbidity (Charlson Comorbidity Index), treatment (prostatectomy, beam radiation, brachytherapy, hormonal therapy, watchful waiting)

Age, year of diagnosis, stage

11% (comorbidity) (2.90b / 2.70) þ5% (health behaviour) (2.90 / 3.00) Localized stage: Age 59 years: 21% (comorbidity) (2.32 / 2.04) 38% (treatment) (2.32 / 1.82) Age 60e74 years: 25% (comorbidity) (1.81 / 1.61) 22% (treatment) (1.81 / 1.63) Advanced stage: Age 60e74 years: 6% (comorbidity) (1.36 / 1.34) 36% (treatment) (1.36 / 1.23) Age 75: 7% (comorbidity) (1.27 / 1.23) þ37% (treatment) (1.27 / 1.37) 6% (comorbidity) (1.35 / 1.33) 33% (stage) (1.33 / 1.22)

Overall, health behaviours do not and comorbidities marginally reduce social inequalities in survival. The impact of treatment and comorbidity varies with age group and tumour stage. Both partly reduce inequalities. Overall, the adjustment for treatment leads to stronger reductions than for comorbidity.

Aarts et al., 2013b

Year of diagnosis, stratified for age and stage

Byers et al., 2008

Stepwise adjustment: stage (first step), comorbidity (Charlson Comorbidity Index; second step)

Age, race/ethnicity

Jansen et al., 2014

Stage

Age

Louwman et al., 2010

Comorbidity (Charlson Comorbidity Index) Stage

Age

Pokhrel et al., 2010

3-month: 51% (Relative excess risk: 1.91 / 1.45) 1-year conditional on 3month: 3% (1.64 / 1.62) 5-year conditional on 1year: 73% (1.48 / 1.13) 22% (1.47 / 1.36)

Age

Rapiti et al., 2009

Stepwise adjustment: stage and grade (first step), method for detection (PSA screening, symptoms, other; second step), sector of care (public, private) and treatment (prostatectomy, radiotherapy, hormonal therapy, watchful waiting, other; third step)

Age, nationality, period of diagnosis

Schwartz et al., 2009

Treatment (prostatectomy, beam radiation, no definitive therapy)

Age, race/ethnicity, grade, stratified for stage

No calculation possible as exact data is missing. 20% (stage and grade) (1.50 / 1.40) 0% (method of detection) (1.40 / 1.40) 50% (treatment and sector of care) (1.40 / 1.20)

Localized stage: 29% (1.55 / 1.39) Regional stage: 40% (1.62 / 1.37)

Inequalities are reduced after introducing stage and comorbidity. Stage at diagnosis accounts more than comorbidity. Depending on the time since diagnosis inequalities are reduced severely after adjusting for stage at diagnosis.

Comorbidity partly reduces the inequalities. Adjustment for stage reduces the inequalities. The introduction of stage and grade reduces the inequalities in survival. After adjustment for the method of detection the association is not significant although the HR is not changing. By subsequent introduction of health care characteristics (treatment, sector of care) again the inequalities are reduced severely. Regarding both stages the introduction of treatment modalities reduces survival inequalities to a comparatively strong extent.

a Percentage change of inequalities by adding explanatory factors to the models was calculated using the formula ([HR/RER Basic Model  HR/RER Basic model þ explanatory factors]/ [HR/RER Basic model  1])  100. Basic model is the model only adjusted for age, race/ethnicity, year of diagnosis and stage (one study). Percentage change is documented when hazard ratio or relative excess risk are statistically significant in basic model. b Hazard ratio if not stated otherwise. Statistically significant values are bold.

and Adami, 2006; Feinstein et al., 1985). It means that the introduction of more accurate and sensitive diagnostic technologies leads to reclassifications of tumour stages and to better prognosis of both patient groups with localized and more advanced stages without really improving survival time. As there are socioeconomic differences in the quality and intensity of diagnostic procedures,

this bias could result in lower stage-adjusted or stage-specific survival among deprived patients (Woods et al., 2006). Overall, there is still a controversial debate about the benefits and harms of prostate cancer screening and over-diagnosis as well as overtreatment (Gomella et al., 2011; Neal et al., 2009). Finally, later stage at diagnosis can partly explain socioeconomic differences in

16

J. Klein, O. von dem Knesebeck / Social Science & Medicine 142 (2015) 9e18

survival but due to statistical artefacts conclusions must be drawn carefully. Results of studies analysing the impact of stage among other cancer sites differ in their results (Booth et al., 2010; Byers et al., 2008; Frederiksen et al., 2009; Jeffreys et al., 2009; Rutherford et al., 2013; Schrijvers et al., 1995). The two further explanatory factors effecting reductions of survival inequalities are treatment and, to a lesser extent, comorbidity. It is difficult to figure out the independent effects because comorbidity and treatment are usually connected with each other as the severity of comorbidities influence the decision for treatments of prostatic neoplasms (Berglund et al., 2011; Hall et al., 2005b). Like differences in SES regarding severity and amount of comorbidities are documented (Aarts et al., 2013a, 2013b; Louwman et al., 2010), there is also clear evidence about social disparities in treatment favouring patients with higher SES to undergo radical prostatectomy (Aarts et al., 2013b; Berglund et al., 2012; Lyratzopoulos et al., 2010; Rapiti et al., 2009; Tewari et al., 2009). This surgical intervention is known to be the treatment option that provides the best prognosis to survive (Lin et al., 2009; Liu et al., 2008; Tewari et al., 2006). Nevertheless, especially among patients with localized prostate cancer there is hardly evidence and intense controversy about which treatment is the most effective (Bill-Axelson et al., 2014; Wilt et al., 2012). Moreover, discussions about treatment-related side-effects and impaired quality of life are still ongoing (Johansson et al., 2011) leading to further extensive randomized trials (Wiegel et al., 2013). Beside the occurrence of comorbidity, unequal distribution of health literacy, active decision-making and transparency and education in the doctorepatient relationship are discussed to have an impact on treatment decisions among prostate cancer patients (Aarts et al., 2013b; Chu and Freedland, 2010; Gaston and Mitchell, 2005). An extensive literature overview figured out the importance and characteristics regarding health literacy and doctorepatient communication in cancer care (Davis et al., 2002). A recent Dutch study has shown clear associations between health literacy components and various indicators of SES (van der Heide et al., 2013). Perceived barriers to care and perceptions and attitudes toward participatory decision making are unfavourably affected by lower health literacy (Yin et al., 2012). Furthermore, psychosocial factors like being married were shown to improve survival significantly after prostate cancer diagnosis and should additionally be taken into account when analysing disparities (Harvei and Kravdal, 1997; Kravdal, 2013a, 2001). And, although lifestyle behaviour plays a minor role in the reviewed studies, it could also be relevant for further analyses. There is some evidence suggesting protection by a healthier lifestyle against prostate cancer recurrence and progression (Davies et al., 2011), and moreover, a favourable lifestyle might reflect a better awareness of PSA test (Ahmed et al., 2008) and less amount and severity of comorbidities (Aarts et al., 2013a; Berglund et al., 2011). 5. Conclusion The reduction of inequalities in survival regarding one of the most prevalent tumour sites worldwide is a major public health concern. The review illustrates the current relevance of this topic despite various improvements in diagnosis and treatment of prostate cancer. Furthermore, there is an ongoing discussion if social inequalities in cancer survival have widened and no definitive evidence could be given (Kravdal, 2013b). Even though a publication bias potentially could have limited the results, this review shows that social inequalities in prostate cancer survival are not only currently evident but also do at least not seem to decrease. Moreover, the contribution of different explanatory factors was systematically explored indicating potential options for

interventions in terms of the patient (e.g. increasing awareness of symptoms and health literacy), and the health care (e.g. access to early detection and treatment options, quality of doctorepatient relationship). By specifically considering lower status groups such interventions can be able to minimize socioeconomic differences. Evidence regarding this topic is still sparse. Therefore, the explanation of survival inequalities among prostate cancer patients and the contribution of particular factors are important topics for future research. Although the introduced explanatory factors reduce the inequalities in survival and provide evidence for potential interventions, there is still unexplained variance after their inclusion. As health inequalities result from social inequalities, action across the social determinants of health e the circumstances in which people are born, grow, live, work and age e is also required (Marmot et al., 2010, 2012). That includes reduction of assets and income inequality, ensuring universal access to high-quality, affordable education as well as improvement of daily living conditions over the life course. Good quality employment and working conditions in terms of physical and psychosocial work characteristics, healthy environments and adequate public health conditions are further areas to work on health equity. In doing so, it is important to concern with the whole social gradient in health, and moreover, not only focusing on individual behaviours while ignoring the environmental drivers of these behaviours. The improvement of these circumstances and conditions is particularly addressed to fundamental drivers, i.e. economics, social policies and governance. Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.socscimed.2015.07.006. References Aarts, M.J., Kamphuis, C.B.M., Louwman, M.J., Coebergh, J.W.W., Mackenbach, J.P., van Lenthe, F.J., 2013a. Educational inequalities in cancer survival: a role for comorbidities and health behaviours? J. Epidemiol. Community Health 67, 365e373. Aarts, M.J., Koldewijn, E.L., Poortmans, P.M., Coebergh, J.W.W., Louwman, M., 2013b. The impact of socioeconomic status on prostate cancer treatment and survival in the southern Netherlands. Urology 81, 593e600. Aarts, M.J., van der Aa, M.A., Coebergh, J.W.W., Louwman, W.J., 2010. Reduction of socioeconomic inequality in cancer incidence in the south of the Netherlands during 1996e2008. Eur. J. Cancer 46, 2633e2646. Ahmed, F.S., Borrell, L.N., Spencer, B.A., 2008. Health risk behaviors and prostate specific antigen awareness among men in California. J. Urol. 180, 658e662. Albano, J.D., Ward, E., Jemal, A., Anderson, R., Cokkinides, V.E., Murray, T., Henley, J., Liff, J., Thun, M.J., 2007. Cancer mortality in the United States by education level and race. J. Natl. Cancer Inst. 99, 1384e1394. Albertsen, P.C., Hanley, J.A.H., Barrows, G.H., Penson, D.F., Kowalczyk, P.D.H., Sanders, M.M., Fine, J., 2005. Prostate cancer and the Will Rogers phenomenon. J. Natl. Cancer Inst. 97, 1248e1253. Australian Institute of Health and Walfare, 2013. Cancer survival and prevalence in Australia: period estimates from 1982 to 2010. Asia Pac. J. Clin. Oncol. 9, 29e39. Auvinen, A., Karjalainen, S., 1997. Possible explanations for social class differences in cancer patient survival. In: Kogevinas, M., Pearce, N., Susser, M., Boffetta, P. (Eds.), Social Inequalities and Cancer. IARC Sci., Lyon, pp. 377e397. Publ. No. 138. Berglund, A., Garmo, H., Robinson, D., Tishelman, C., Holmberg, L., Bratt, O., Adolfsson, J., Stattin, P., Lambe, M., 2012. Differences according to socioeconomic status in the management and mortality in men with high risk prostate cancer. Eur. J. Cancer 48, 75e84. Berglund, A., Garmo, H., Tishelman, C., Holmberg, L., Stattin, P., Lambe, M., 2011. Comorbidity, treatment and mortality: a population based cohort study of prostate cancer in PCBaSe Sweden. J. Urol. 185, 833e839. Berry, D.A., 2014. Failure of researchers, reviewers, editors, and the media to understand flaws in cancer screening studies: application to an article in cancer. Cancer 120, 2784e2791. Bill-Axelson, A., Holmberg, L., Garmo, H., Rider, J.R., Taari, K., Busch, C., Nordling, S., n, O., Palmgren, J., H€ aggman, M., Andersson, S.-O., Spångberg, A., Andre Steineck, G., Adami, H.-O., Johansson, J.-E., 2014. Radical prostatectomy or watchful waiting in early prostate cancer. N. Engl. J. Med. 370, 932e942. Booth, C.M., Li, G., Zhang-Salomons, J., Mackillop, W.J., 2010. The impact of socioeconomic status on stage of cancer at diagnosis and survival. Cancer 116,

J. Klein, O. von dem Knesebeck / Social Science & Medicine 142 (2015) 9e18 4160e4167. Bravo, L.E., García, L.S., Collazos, P.A., 2014. Cancer survival in Cali, Colombia: a d. Cali Colomb. 45, 110e116. population-based study, 1995e2004. Colomb. Me Bray, F., Ren, J.-S., Masuyer, E., Ferlay, J., 2013. Global estimates of cancer prevalence for 27 sites in the adult population in 2008. Int. J. Cancer 132, 1133e1145. Burns, R.M., Sharp, L., Sullivan, F.J., Deady, S.E., Drummond, F.J., O Neill, C., 2014. Factors driving inequality in prostate cancer survival: a population based study. PLoS One 9, e106456. Byers, T.E., Wolf, H.J., Bauer, K.R., Bolick-Aldrich, S., Chen, V.W., Finch, J.L., Fulton, J.P., Schymura, M.J., Shen, T., Van Heest, S., Yin, X., for the Patterns of Care Study Group, 2008. The impact of socioeconomic status on survival after cancer in the United States: findings from the National Program of Cancer Registries Patterns of Care Study. Cancer 113, 582e591. Chang, C.M., Su, Y.C., Lai, N.S., Huang, K.Y., Chien, S.H., Chang, Y.H., Lian, W.C., Hsu, T.W., Lee, C.C., 2012. The combined effect of individual and neighborhood socioeconomic status on cancer survival rates. PLoS One 7, e44325. Chu, D.I., Freedland, S.J., 2010. Prostate cancer. Socioeconomic status and disparities in treatment patterns. Nat. Rev. Urol. 7, 480e481. Clarke, N.W., 2008. Survival from prostate cancer in England and Wales up to 2001. Br. J. Cancer 99, S78eS79. Clegg, L.X., Reichman, M.E., Miller, B.A., Hankey, B.F., Singh, G.K., Lin, Y.D., Goodman, M.T., Lynch, C.F., Schwartz, S.M., Chen, V.W., Bernstein, L., Gomez, S.L., Graff, J.J., Lin, C.C., Johnson, N.J., Edwards, B.K., 2009. Impact of socioeconomic status on cancer incidence and stage at diagnosis: selected findings from the surveillance, epidemiology, and end results: National Longitudinal Mortality Study. Cancer Causes Control 20, 417e435. Coleman, M.P., Rachet, B., Woods, L.M., Mitry, E., Riga, M., Cooper, N., Quinn, M.J., ve, J., 2004. Trends and socioeconomic inequalities in cancer Brenner, H., Este survival in England and Wales up to 2001. Br. J. Cancer 90, 1367e1373. Davies, N.J., Batehup, L., Thomas, R., 2011. The role of diet and physical activity in breast, colorectal, and prostate cancer survivorship: a review of the literature. Br. J. Cancer 105 (Suppl. 1), S52eS73. Davis, T.C., Williams, M.V., Marin, E., Parker, R.M., Glass, J., 2002. Health literacy and cancer communication. CA Cancer J. Clin. 52, 134e149. De Angelis, R., Sant, M., Coleman, M.P., Francisci, S., Baili, P., Pierannunzio, D., Trama, A., Visser, O., Brenner, H., Ardanaz, E., Bielska-Lasota, M., Engholm, G., Nennecke, A., Siesling, S., Berrino, F., Capocaccia, R., EUROCARE-5 Working Group, 2014. Cancer survival in Europe 1999e2007 by country and age: results of EUROCARE-5-a population-based study. Lancet Oncol. 15, 23e34. De Vries, E., Karim-Kos, H.E., Janssen-Heijnen, M.L.G., Soerjomataram, I., Kiemeney, L.A., Coebergh, J.W.W., 2010. Explanations for worsening cancer survival. Nat. Rev. Clin. Oncol. 7, 60e63. Dickman, P.W., Adami, H.-O., 2006. Interpreting trends in cancer patient survival. J. Intern. Med. 260, 103e117. Dutta Roy, S., Philip, J., Javle, P., 2005. Trends in prostate cancer incidence and survival in various socioeconomic classes: a population-based study. Int. J. Urol. 12, 644e653. Du, X.L., Fang, S., Coker, A.L., Sanderson, M., Aragaki, C., Cormier, J.N., Xing, Y., Gor, B.J., Chan, W., 2006. Racial disparity and socioeconomic status in association with survival in older men with local/regional stage prostate carcinoma: findings from a large community-based cohort. Cancer 106, 1276e1285. Ellis, L., Coleman, M.P., Rachet, B., 2012. How many deaths would be avoidable if socioeconomic inequalities in cancer survival in England were eliminated? A national population-based study, 1996e2006. Eur. J. Cancer 48, 270e278. Elstad, J.I., Torstensrud, R., Lyngstad, T.H., Kravdal, O., 2011. Trends in educational inequalities in mortality, seven types of cancers, Norway 1971e2002. Eur. J. Public Health 22, 771e776. Faggiano, F., Partanen, T., Kogevinas, M., Boffetta, P., 1997. Socioeconomic differences in cancer incidence and mortality. In: Kogevinas, M., Pearce, N., Susser, M., Boffetta, P. (Eds.), Social Inequalities and Cancer. IARC Sci., Lyon, pp. 65e176. Publ. No. 138. Fairley, L., Baker, M., Whiteway, J., Cross, W., Forman, D., 2009. Trends in nonmetastatic prostate cancer management in the Northern and Yorkshire region of England, 2000e2006. Br. J. Cancer 101, 1839e1845. Feinstein, A.R., Sosin, D.M., Wells, C.K., 1985. The Will Rogers phenomenon. Stage migration and new diagnostic techniques as a source of misleading statistics for survival in cancer. N. Engl. J. Med. 312, 1604e1608. Frederiksen, B.L., Osler, M., Harling, H., Ladelund, S., Jørgensen, T., 2009. Do patient characteristics, disease, or treatment explain social inequality in survival from colorectal cancer? Soc. Sci. Med. 69, 1107e1115. Freeman, V.L., Ricardo, A.C., Campbell, R.T., Barrett, R.E., Warnecke, R.B., 2011. Association of census tract-level socioeconomic status with disparities in prostate cancer-specific survival. Cancer Epidemiol. Biomark Prevent. 20, 2150e2159. Gaston, C.M., Mitchell, G., 2005. Information giving and decision-making in patients with advanced cancer: a systematic review, 1982 Soc. Sci. Med. 61, 2252e2264. Gilligan, T., 2005. Social disparities and prostate cancer: mapping the gaps in our knowledge. Cancer Causes Control 16, 45e53. Goddard, M., Smith, P., 2001. Equity of access to health care services: theory and evidence from the UK. Soc. Sci. Med. 53, 1149e1162. Gomella, L.G., Liu, X.S., Trabulsi, E.J., Kelly, W.K., Myers, R., Showalter, T., Dicker, A., Wender, R., 2011. Screening for prostate cancer: the current evidence and guidelines controversy. Can. J. Urol. 18, 5875e5883. Hall, S.E., Holman, C.D.J., Wisniewski, Z.S., Semmens, J., 2005a. Prostate cancer: socio-economic, geographical and private-health insurance effects on care and survival. BJU Int. 95, 51e58.

17

Hall, W.H., Jani, A.B., Ryu, J.K., Narayan, S., Vijayakumar, S., 2005b. The impact of age and comorbidity on survival outcomes and treatment patterns in prostate cancer. Prostate Cancer Prostatic Dis. 8, 22e30. Hansen, R.P., Olesen, F., Sørensen, H.T., Sokolowski, I., Søndergaard, J., 2008. Socioeconomic patient characteristics predict delay in cancer diagnosis: a Danish cohort study. BMC Health Serv. Res. 8, 49. Harvei, S., Kravdal, O., 1997. The importance of marital and socioeconomic status in incidence and survival of prostate cancer. An analysis of complete Norwegian birth cohorts. Prev. Med. 26, 623e632. Hayen, A., Smith, D.P., Patel, M.I., O'Connell, D.L., 2008. Patterns of surgical care for prostate cancer in NSW, 1993e2002: rural/urban and socio-economic variation. Aust. N. Z. J. Public Health 32, 417e420. Haynes, R., Pearce, J., Barnett, R., 2008. Cancer survival in New Zealand: ethnic, social and geographical inequalities, 1982 Soc. Sci. Med. 67, 928e937. Hellenthal, N.J., Parikh-Patel, A., Bauer, K., Ralph, W., deVere, W., Koppie, T.M., 2010. Men of higher socioeconomic status have improved outcomes after radical prostatectomy for localized prostate cancer. Urology 76, 1409e1413. Hussain, S.K., Lenner, P., Sundquist, J., Hemminki, K., 2008. Influence of education level on cancer survival in Sweden. Ann. Oncol. 19, 156e162. Jansen, L., Eberle, A., Emrich, K., Gondos, A., Holleczek, B., Kajuter, H., Maier, W., Nennecke, A., Pritzkuleit, R., Brenner, H., 2014. Socioeconomic deprivation and cancer survival in Germany: an ecological analysis in 200 districts in Germany. Int. J. Cancer 134, 2951e2960. Jeffreys, M., Sarfati, D., Stevanovic, V., Tobias, M., Lewis, C., Pearce, N., Blakely, T., 2009. Socioeconomic inequalities in cancer survival in New Zealand: the role of extent of disease at diagnosis. Cancer Epidemiol. Biomark Prevent. 18, 915e921. Johansson, E., Steineck, G., Holmberg, L., Johansson, J.-E., Nyberg, T., Ruutu, M., BillAxelson, A., SPCG-4 Investigators, 2011. Long-term quality-of-life outcomes after radical prostatectomy or watchful waiting: the Scandinavian Prostate Cancer Group-4 randomised trial. Lancet Oncol. 12, 891e899. Kogevinas, M., Porta, M., 1997. Socioeconomic differences in cancer survival: a review of the evidence. In: Kogevinas, M., Pearce, N., Susser, M., Boffetta, P. (Eds.), Social Inequalities and Cancer. IARC Sci., Lyon, pp. 177e206. Publ. No. 138. Kravdal, H., 2013a. Widening educational differences in cancer survival in Norway. Eur. J. Public Health 24, 270e275. Kravdal, O., 2000. Social inequalities in cancer survival. Popul. Stud. 54, 1e18. Kravdal, Ø., 2001. The impact of marital status on cancer survival. Soc. Sci. Med. 52, 357e368. Kravdal, Ø., 2013b. The poorer cancer survival among the unmarried in Norway: is much explained by comorbidities? Soc. Sci. Med. 81, 42e52. Krieger, N., Chen, J.T., Kosheleva, A., Waterman, P.D., 2012. Shrinking, widening, reversing, and stagnating trends in US socioeconomic inequities in cancer mortality for the total, black, and white populations: 1960e2006. Cancer Causes Control 23, 297e319. Krupski, T.L., Kwan, L., Afifi, A.A., Litwin, M.S., 2005. Geographic and socioeconomic variation in the treatment of prostate cancer. J. Clin. Oncol. 23, 7881e7888. Lin, D.W., Porter, M., Montgomery, B., 2009. Treatment and survival outcomes in young men diagnosed with prostate cancer: a population-based cohort study. Cancer 115, 2863e2871. Liu, L., Coker, A.L., Du, X.L., Cormier, J.N., Ford, C.E., Fang, S., 2008. Long-term survival after radical prostatectomy compared to other treatments in older men with local/regional prostate cancer. J. Surg. Oncol. 97, 583e591. Li, X., Sundquist, K., Sundquist, J., 2012. Neighborhood deprivation and prostate cancer mortality: a multilevel analysis from Sweden. Prostate Cancer Prostatic Dis. 15, 128e134. Louwman, W.J., Aarts, M.J., Houterman, S., van Lenthe, F.J., Coebergh, J.W.W., Janssen-Heijnen, M.L.G., 2010. A 50% higher prevalence of life-shortening chronic conditions among cancer patients with low socioeconomic status. Br. J. Cancer 103, 1742e1748. Luo, Q., Yu, X.Q., Cooke-Yarborough, C., Smith, D.P., O'Connell, D.L., 2013. Characteristics of cases with unknown stage prostate cancer in a population-based cancer registry. Cancer Epidemiol. 37, 813e819. Lyratzopoulos, G., Abel, G.A., Brown, C.H., Rous, B.A., Vernon, S.A., Roland, M., Greenberg, D.C., 2013. Socio-demographic inequalities in stage of cancer diagnosis: evidence from patients with female breast, lung, colon, rectal, prostate, renal, bladder, melanoma, ovarian and endometrial cancer. Ann. Oncol. 24, 843e850. Lyratzopoulos, G., Barbiere, J.M., Greenberg, D.C., Wright, K.A., Neal, D.E., 2010. Population based time trends and socioeconomic variation in use of radiotherapy and radical surgery for prostate cancer in a UK region: continuous survey. BMJ 340, c1928. Macleod, U., Mitchell, E.D., Burgess, C., Macdonald, S., Ramirez, A.J., 2009. Risk factors for delayed presentation and referral of symptomatic cancer: evidence for common cancers. Br. J. Cancer 101 (Suppl. 2), S92eS101. Marmot, M., Allen, J., Goldblatt, P., 2010. A social movement, based on evidence, to reduce inequalities in health. Soc. Sci. Med. 71, 1254e1258. Marmot, M., Allen, J., Bell, R., Bloomer, E., Goldblatt, P., Consortium for the European Review of Social Determinants of Health and the Health Divide, 2012. WHO European review of social determinants of health and the health divide. Lancet 380, 1011e1029. Marsa, K., Johnsen, N.F., Bidstrup, P.E., Johannesen-Henry, C.T., Friis, S., 2008. Social inequality and incidence of and survival from male genital cancer in a population-based study in Denmark, 1994e2003. Eur. J. Cancer 44, 2018e2029. McPhail, S., Elliss-Brookes, L., Shelton, J., Ives, A., Greenslade, M., Vernon, S., Morris, E.J.A., Richards, M., 2013. Emergency presentation of cancer and short-

18

J. Klein, O. von dem Knesebeck / Social Science & Medicine 142 (2015) 9e18

term mortality. Br. J. Cancer 109, 2027e2034. Menvielle, G., Kunst, A.E., Stirbu, I., Strand, B.H., Borrell, C., Regidor, E., Leclerc, A., Esnaola, S., Bopp, M., Lundberg, O., Artnik, B., Costa, G., Deboosere, P., Martikainen, P., Mackenbach, J.P., 2008. Educational differences in cancer mortality among women and men: a gender pattern that differs across Europe. Br. J. Cancer 98, 1012e1019. Miki, Y., Inoue, M., Ikeda, A., Sawada, N., Nakaya, T., Shimazu, T., Iwasaki, M., Yamaji, T., Sasazuki, S., Shibuya, K., Tsugane, S., JPHC Study Group, 2014. Neighborhood deprivation and risk of cancer incidence, mortality and survival: results from a population-based cohort study in Japan. PLoS One 9, e106729. Moher, D., Liberati, A., Tetzlaff, J., Altman, D.G., 2009. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. Ann. Intern. Med. 151, 264e269. http://dx.doi.org/10.7326/0003-4819-151-4-20090818000135. Movsas, A., Ibrahim, R., Elshaikh, M.A., Lamerato, L., Lu, M., Sitarik, A., Pradhan, D., Walker, E.M., Stricker, H., Freytag, S.O., Chetty, I.J., Movsas, B., Siddiqui, F., 2014. Do sociodemographic factors influence outcome in prostate cancer patients treated with external beam radiation therapy? Am. J. Clin. Oncol. http:// dx.doi.org/10.1097/COC.0000000000000093 [Epub ahead of print]. National Cancer Intelligence Network, 2009. Cancer Incidence by Deprivation. England, 1995e2004. London, 2009. Neal, D.E., Donovan, J.L., Martin, R.M., Hamdy, F.C., 2009. Screening for prostate cancer remains controversial. Lancet 374, 1482e1483. Neal, R.D., Allgar, V.L., 2005. Sociodemographic factors and delays in the diagnosis of six cancers: analysis of data from the National Survey of NHS Patients: cancer. Br. J. Cancer 92, 1971e1975. Niu, X., Pawlish, K.S., Roche, L.M., 2010. Cancer survival disparities by race/ethnicity and socioeconomic status in New Jersey. J. Health Care Poor Underserved 21, 144e160. Papa, N., Lawrentschuk, N., Muller, D., MacInnis, R., Ta, A., Severi, G., Millar, J., Syme, R., Giles, G., Bolton, D., 2014. Rural residency and prostate cancer specific mortality: results from the Victorian Radical Prostatectomy Register. Aust. N. Z. J. Public Health 38, 449e454. Pitman, M., Korets, R., Kates, M., Hruby, G.W., McKiernan, J.M., 2012. Socioeconomic and clinical factors influence the interval between positive prostate biopsy and radical prostatectomy. Urology 80, 1027e1032. Pokhrel, A., Martikainen, P., Pukkala, E., Rautalahti, M., Sepp€ a, K., Hakulinen, T., 2010. Education, survival and avoidable deaths in cancer patients in Finland. Br. J. Cancer 103, 1109e1114. Prasad, S.M., Eggener, S.E., Lipsitz, S.R., Irwin, M.R., Ganz, P.A., Hu, J.C., 2014. Effect of depression on diagnosis, treatment, and mortality of men with clinically localized prostate cancer. J. Clin. Oncol. 32, 2471e2478. Quaglia, A., Lillini, R., Mamo, C., Ivaldi, E., Vercelli, M., SEIH (Socio-Economic Indicators, Health) Working Group, 2013. Socio-economic inequalities: a review of methodological issues and the relationships with cancer survival. Crit. Rev. Oncol. Hematol. 85, 266e277. Rachet, B., Ellis, L., Maringe, C., Chu, T., Nur, U., Quaresma, M., Shah, A., Walters, S., Woods, L., Forman, D., Coleman, M.P., 2010. Socioeconomic inequalities in cancer survival in England after the NHS cancer plan. Br. J. Cancer 103, 446e453. Rapiti, E., Fioretta, G., Schaffar, R., Neyroud-Caspar, I., Verkooijen, H.M., Schmidlin, F., Miralbell, R., Zanetti, R., Bouchardy, C., 2009. Impact of socioeconomic status on prostate cancer diagnosis, treatment, and prognosis. Cancer 115, 5556e5565. Redaniel, M.T., Martin, R.M., Gillatt, D., Wade, J., Jeffreys, M., 2013. Time from diagnosis to surgery and prostate cancer survival: a retrospective cohort study. BMC Cancer 13, 559. Robbins, A.S., Koppie, T.M., Gomez, S.L., Parikh-Patel, A., Mills, P.K., 2007a. Differences in prognostic factors and survival among white and Asian men with prostate cancer, California, 1995e2004. Cancer 110, 1255e1263. Robbins, A.S., Yin, D., Parikh-Patel, A., 2007b. Differences in prognostic factors and survival among White men and Black men with prostate cancer, California, 1995e2004. Am. J. Epidemiol. 166, 71e78. Robb, K., Stubbings, S., Ramirez, A., Macleod, U., Austoker, J., Waller, J., Hiom, S., Wardle, J., 2009. Public awareness of cancer in Britain: a population-based survey of adults. Br. J. Cancer 101, S18eS23. Ross, L.E., Taylor, Y.J., Howard, D.L., 2011. Trends in prostate-specific antigen test use, 2000e2005. Public Health Rep. 126, 228e239. Rowan, S., Rachet, B., Alexe, D.M., Cooper, N., Coleman, M.P., 2008. Survival from prostate cancer in England and Wales up to 2001. Br. J. Cancer 99, S75eS77. Rutherford, M.J., Hinchliffe, S.R., Abel, G.A., Lyratzopoulos, G., Lambert, P.C., Greenberg, D.C., 2013. How much of the deprivation gap in cancer survival can be explained by variation in stage at diagnosis: an example from breast cancer in the east of England. Int. J. Cancer 133, 2192e2200. Sarfati, D., Blakely, T., Pearce, N., 2010. Measuring cancer survival in populations: relative survival vs cancer-specific survival. Int. J. Epidemiol. 39, 598e610. Schrijvers, C.T., Mackenbach, J.P., 1994. Cancer patient survival by socioeconomic status in seven countries: a review for six common cancer sites [corrected]. J. Epidemiol. Community Health 48, 441e446. Schrijvers, C.T., Mackenbach, J.P., Lutz, J.M., Quinn, M.J., Coleman, M.P., 1995. Deprivation, stage at diagnosis and cancer survival. Int. J. Cancer 63, 324e329. Schwartz, K., Powell, I.J., Underwood, W., George, J., Yee, C., Banerjee, M., 2009.

Interplay of race, socioeconomic status, and treatment on survival of patients with prostate cancer. Urology 74, 1296e1302. Shack, L.G., Rachet, B., Brewster, D.H., Coleman, M.P., 2007. Socioeconomic inequalities in cancer survival in Scotland 1986e2000. Br. J. Cancer 97, 999e1004. Shafique, K., Morrison, D.S., 2013. Socio-economic inequalities in survival of patients with prostate cancer: role of age and Gleason grade at diagnosis. PLoS One 8, e56184. Shafique, K., Oliphant, R., Morrison, D.S., 2012a. The impact of socio-economic circumstances on overall and grade-specific prostate cancer incidence: a population-based study. Br. J. Cancer 107, 575e582. Shafique, K., Proctor, M.J., McMillan, D.C., Leung, H., Smith, K., Sloan, B., Morrison, D.S., 2013. The modified Glasgow prognostic score in prostate cancer: results from a retrospective clinical series of 744 patients. BMC Cancer 13, 292. Shafique, K., Proctor, M.J., McMillan, D.C., Qureshi, K., Leung, H., Morrison, D.S., 2012b. Systemic inflammation and survival of patients with prostate cancer: evidence from the Glasgow inflammation outcome study. Prostate Cancer Prostatic Dis. 15, 195e201. Skalicka, V., van Lenthe, F., Bambra, C., Krokstad, S., Mackenbach, J., 2009. Material, psychosocial, behavioural and biomedical factors in the explanation of relative socio-economic inequalities in mortality: evidence from the HUNT study. Int. J. Epidemiol. 38, 1272e1284. Sloggett, A., Young, H., Grundy, E., 2007. The association of cancer survival with four socioeconomic indicators: a longitudinal study of the older population of England and Wales 1981e2000. BMC Cancer 7, 20. Spencer, B.A., Steinberg, M., Malin, J., Adams, J., Litwin, M.S., 2003. Quality-of-care indicators for early-stage prostate cancer. J. Clin. Oncol. 21, 1928e1936. Tewari, A., Horninger, W., Pelzer, A.E., Demers, R., Crawford, E.D., Gamito, E.J., Divine, G., Johnson, C.C., Bartsch, G., Menon, M., 2005. Factors contributing to the racial differences in prostate cancer mortality. BJU Int. 96, 1247e1252. Tewari, A.K., Gold, H.T., Demers, R.Y., Johnson, C.C., Yadav, R., Wagner, E.H., Yood, M.U., Field, T.S., Divine, G., Menon, M., 2009. Effect of socioeconomic factors on long-term mortality in men with clinically localized prostate cancer. Urology 73, 624e630. Tewari, A., Raman, J.D., Chang, P., Rao, S., Divine, G., Menon, M., 2006. Long-term survival probability in men with clinically localized prostate cancer treated either conservatively or with definitive treatment (radiotherapy or radical prostatectomy). Urology 68, 1268e1274. Tyson, M.D., Andrews, P.E., Etzioni, D.A., Ferrigni, R.G., Humphreys, M.R., Swanson, S.K., Castle, E.K., 2013. Marital status and prostate cancer outcomes. Can. J. Urol. 20, 6702e6706. Van den Bergh, R.C.N., Albertsen, P.C., Bangma, C.H., Freedland, S.J., Graefen, M., Vickers, A., van der Poel, H.G., 2013. Timing of curative treatment for prostate cancer: a systematic review. Eur. Urol. 64, 204e215. Van der Heide, I., Rademakers, J., Schipper, M., Droomers, M., Sørensen, K., Uiters, E., 2013. Health literacy of Dutch adults: a cross sectional survey. BMC Public Health 13, 179. Williams, N., Hughes, L.J., Turner, E.L., Donovan, J.L., Hamdy, F.C., Neal, D.E., Martin, R.M., Metcalfe, C., 2011. Prostate-specific antigen testing rates remain low in UK general practice: a cross-sectional study in six English cities. BJU Int. 108, 1402e1408. Wiegel, T., Albers, P., Bussar-Maatz, R., Gottberg, A., H€ arter, M., Kieser, M., Kristiansen, G., Nettekoven, G., Martus, P., Schmidberger, H., Wellek, S., € ckle, M., 2013. PREFERE e the German prostatic cancer study: questions and Sto claims surrounding study initiation in January 2013. Urology 52, 576e579. Wilt, T.J., Brawer, M.K., Jones, K.M., Barry, M.J., Aronson, W.J., Fox, S., Gingrich, J.R., Wei, J.T., Gilhooly, P., Grob, B.M., Nsouli, I., Iyer, P., Cartagena, R., Snider, G., Roehrborn, C., Sharifi, R., Blank, W., Pandya, P., Andriole, G.L., Culkin, D., Wheeler, T., Prostate Cancer Intervention versus Observation Trial (PIVOT) Study Group, 2012. Radical prostatectomy versus observation for localized prostate cancer. N. Engl. J. Med. 367, 203e213. Woods, L.M., Rachet, B., Coleman, M.P., 2005. Choice of geographic unit influences socioeconomic inequalities in breast cancer survival. Br. J. Cancer 92, 1279e1282. Woods, L.M., Rachet, B., Coleman, M.P., 2006. Origins of socio-economic inequalities in cancer survival: a review. Ann. Oncol. 17, 5e19. Yin, H.S., Dreyer, B.P., Vivar, K.L., MacFarland, S., van Schaick, L., Mendelsohn, A.L., 2012. Perceived barriers to care and attitudes towards shared decision-making among low socioeconomic status parents: role of health literacy. Acad. Pediatr. 12, 117e124. Yu, M., Tatalovich, Z., Gibson, J.T., Cronin, K.A., 2014a. Using a composite index of socioeconomic status to investigate health disparities while protecting the confidentiality of cancer registry data. Cancer Causes Control 25, 81e92. Yu, X.Q., Luo, Q., Smith, D.P., O'Connell, D.L., Baade, P.D., 2014b. Geographic variation in prostate cancer survival in New South Wales. Med. J. Aust. 200, 586e590. Yu, X.Q., O'Connell, D.L., Gibberd, R.W., Armstrong, B.K., 2008. Assessing the impact of socio-economic status on cancer survival in New South Wales, Australia 1996e2001. Cancer Causes Control 19, 1383e1390. Zhang-Salomons, J., Qian, H., Holowaty, E., Mackillop, W.J., 2006. Associations between socioeconomic status and cancer survival: choice of SES indicator may affect results. Ann. Epidemiol. 16, 521e528.