Socioeconomic differences in the risk of childhood central nervous system tumors in Denmark: a nationwide register-based case-control study.

PURPOSE: Differences in the risk of childhood central nervous system (CNS) tumors by socioeconomic status (SES) may enhance etiologic insights. We conducted a nationwide register-based case-control study to evaluate socioeconomic differences in the risk of childhood CNS tumors in Denmark and examined whether associations varied by different SES measures, time points of assessment, specific tumor types, and age at diagnosis. METHODS: We identified all children born between 1981 and 2013 and diagnosed with a CNS tumor at ages 0-19 years (n = 1,273) from the Danish Cancer Registry and sampled four individually matched controls per case (n = 5,086). We used conditional logistic regression models to estimate associations with individual-level and neighborhood-level socioeconomic measures. RESULTS: We observed elevated risks of ependymoma and embryonal CNS tumors in association with higher parental education (odds ratios (ORs) of 1.6-2.1 for maternal or paternal high education and ependymoma) and higher risk of all tumor types in association with higher maternal income, e.g., OR 1.93; 95% CI 1.05-3.52 for high versus low income for astrocytoma and other gliomas. Associations were often stronger in children diagnosed at ages 5-19 years. We found little evidence for an association with neighborhood SES. CONCLUSION: This large nationwide register study with minimal risk of bias showed that having parents with higher educational level and a mother with higher income was associated with a higher risk of childhood CNS tumors. Bias or under-ascertainment of cases among families with low income or basic education is unlikely to explain our findings.

Work-unit social capital and long-term sickness absence: a prospective cohort study of 32 053 hospital employees.

OBJECTIVE: There is a lack of studies investigating social capital at the workplace level in small and relatively homogeneous work-units. The aim of the study was to investigate whether work-unit social capital predicts a lower risk of individual long-term sickness absence among Danish hospital employees followed prospectively for 1 year. METHODS: This study is based on the Well-being in HospitAL Employees cohort. The study sample consisted of 32 053 individuals nested within 2182 work-units in the Capital Region of Denmark. Work-unit social capital was measured with an eight-item scale covering elements of trust, justice and collaboration between employees and leaders. Social capital at the work-unit level was computed as the aggregated mean of individual-level social capital within each work-unit. Data on long-term sickness absence were retrieved from the employers' payroll system and were operationalised as ≥29 consecutive days of sickness absence. We used a 12-point difference in social capital as the metric in our analyses and conducted two-level hierarchical logistic regression analysis. Adjustments were made for sex, age, seniority, occupational group and part-time work at the individual level, and work-unit size, the proportion of female employees and the proportion of part-time work at the work-unit level. RESULTS: The OR for long-term sickness absence associated with a 12-point higher work-unit social capital was 0.73 (95% CI 0.68 to 0.78). Further, we found an association between higher work-unit social capital and lower long-term sickness absence across quartiles of social capital: compared with the lowest quartile, the OR for long-term sickness absence in the highest quartile was 0.51 (95% CI 0.44 to 0.60). CONCLUSION: Our study provides support for work-unit social capital being a protective factor for individual long-term sickness absence among hospital employees in the Capital Region of Denmark.

Mechanisms underlying social inequality in post-menopausal breast cancer.

This thesis is based on studies conducted in the period 2010-2014 at Department of Public Health, University of Copenhagen and at Department of Epidemiology and Population Health, Albert Einstein College of Medicine, New York. The results are presented in three scientific papers and a synopsis. The main objective of the thesis was to determine mechanisms underlying social inequality (defined by educational level) in postmenopausal breast cancer (BC) by addressing mediating effects through hormone therapy (HT) use, BMI, lifestyle and reproductive factors. The results of previous studies suggest that the higher risk of postmenopausal BC among women of high socioeconomic position (SEP) may be explained by reproductive factors and health behaviors. Women of higher SEP generally have fewer children and give birth at older ages than women of low SEP, and these factors have been found to affect the risk of BC - probably through altered hormone levels. Adverse effects on BC risk have also been documented for modifiable health behaviors that may affect hormone levels, such as alcohol consumption, high BMI, physical inactivity, and HT use. Alcohol consumption and HT use are likewise more common among women of higher SEP. The analyses were based on the Social Inequality in Cancer (SIC) cohort and a subsample of the Women's Health Initiative Observational Study (WHI-OS). The SIC cohort was derived by pooling 6 individual studies from the Copenhagen area including 33,562 women (1,733 BC cases) aged 50-70 years at baseline. The subsample of WHI-OS consisted of two case-cohort studies with measurements of endogenous estradiol (N = 1,601) and insulin (N = 791). Assessment of mediation often relies on comparing multiplicative models with and without the potential mediator. Such approaches provide potentially biased results, because they do not account for mediator-outcome confounding, exposure-dependent mediator-outcome confounding, exposure-mediator interaction and interactions between mediators. In addition, these simple methods do not allow for a decomposition of the total effect into direct and indirect pathways. The counterfactual-based methods for quantifying mediating effects in this thesis were developed specifically for this project taking into account some of the shortcomings of previous methods. The results of this thesis showed that a high versus low educational level was associated with a higher risk of postmenopausal BC and that this effect was partly mediated through HT use, fertility patterns and alcohol consumption in the SIC data. Overall BMI did not mediate the education-BC relation. The results from the WHI-OS for the effect of alcohol consumption on BC risk did not - as hypothesized - seem to be mediated by endogenous estradiol levels; however, the observed higher risk of BC with higher levels of alcohol was restricted to estrogen-receptor positive cases, which indicates a role of estrogens in this relation. In the WHI-OS subsample of non-HT users, a higher risk of BC was found with higher levels of BMI; both estradiol and insulin mediated the effect of BMI on BC. The effect of HT use on BC interacted synergistically with alcohol consumption and this combination appeared to be associated with very high serum levels of estradiol in the SIC data. For BMI combined with HT use, a modest positive association was observed for non-HT users whereas markedly higher risks were observed across all BMI groups in current HT-users with a tendency towards a U-shaped relation. In conclusion, the social inequality in postmenopausal BC seems to be largely mediated by HT use, alcohol consumption and reproductive factors. Various sources of bias - especially misclassification of mediators, but also exposure-dependent confounding - raise some concern about the observed relations. Future studies should focus on life-course perspectives to identify certain windows of susceptibility and collect data on repeated measurements of mediators to enable health behavior trajectories over time. In addition, there is a need for further development of the methodology for the quantification of mediating effects to handle current shortcomings such as exposure-dependent confounding and the potential interactions between mediators.

Cohort profile: the Social Inequality in Cancer (SIC) cohort study.

The Social Inequality in Cancer (SIC) cohort study was established to determine pathways through which socioeconomic position affects morbidity and mortality, in particular common subtypes of cancer. Data from seven well-established cohort studies from Denmark were pooled. Combining these cohorts provided a unique opportunity to generate a large study population with long follow-up and sufficient statistical power to develop and apply new methods for quantification of the two basic mechanisms underlying social inequalities in cancer-mediation and interaction. The SIC cohort included 83 006 participants aged 20-98 years at baseline. A wide range of behavioural and biological risk factors such as smoking, physical inactivity, alcohol intake, hormone replacement therapy, body mass index, blood pressure and serum cholesterol were assessed by self-administered questionnaires, physical examinations and blood samples. All participants were followed up in nationwide demographic and healthcare registries. For those interested in collaboration, further details can be obtained by contacting the Steering Committee at the Department of Public Health, University of Copenhagen, at inan@sund.ku.dk.