Zoe Aitken

Screen-film mammographic density and breast cancer risk: a comparison of the volumetric Standard Mammogram Form and the Interactive Threshold Measurement Methods

Background: Mammographic density is a strong risk factor for breast cancer, usually measured by an area-based threshold method that dichotomizes the breast area on a mammogram into dense and nondense regions. Volumetric methods of breast density measurement, such as the fully automated standard mammogram form (SMF) method that estimates the volume of dense and total breast tissue, may provide a more accurate density measurement and improve risk prediction. Methods: In 2000-2003, a case-control study was conducted of 367 newly confirmed breast cancer cases and 661 age-matched breast cancer-free controls who underwent screen-film mammography at several centers in Toronto, Canada. Conditional logistic regression was used to estimate odds ratios of breast cancer associated with categories of mammographic density, measured with both the threshold and the SMF (version 2.2β) methods, adjusting for breast cancer risk factors. Results: Median percent density was higher in cases than in controls for the threshold method (31% versus 27%) but not for the SMF method. Higher correlations were observed between SMF and threshold measurements for breast volume/area (Spearman correlation coefficient = 0.95) than for percent density (0.68) or for absolute density (0.36). After adjustment for breast cancer risk factors, odds ratios of breast cancer in the highest compared with the lowest quintile of percent density were 2.19 (95% confidence interval, 1.28-3.72; Pt <0.01) for the threshold method and 1.27 (95% confidence interval, 0.79-2.04; Pt = 0.32) for the SMF method. Conclusion: Threshold percent density is a stronger predictor of breast cancer risk than the SMF version 2.2β method in digitized images. Cancer Epidemiol Biomarkers Prev; 19(2); 418–28

The effect of ethnicity on the prevalence of diabetes and associated chronic kidney disease

BACKGROUND The effect of ethnicity on the prevalence of diabetes mellitus (DM) and associated chronic kidney disease (CKD) is unknown. AIM To establish the impact of ethnicity on the prevalence and severity of diabetes mellitus and associated CKD. DESIGN Cross-sectional study of 34 359 adult diabetic patients in three primary care trusts in the UK. METHODS Read coded data from general practice computers was used to analyse the relationship between ethnicity, DM and CKD. RESULTS The prevalence of DM was 3.5% for Whites, 11% for South Asians and 8% for Black groups. The prevalence of CKD (stages 3-5) among diabetics was 18%. CKD stage 3 was more prevalent in Whites compared to South Asians--OR 0.79 (95% CI: 0.71-0.87) and Blacks--OR 0.49 (95% CI: 0.43-0.57). Among all CKD patients severity (CKD stages 4, 5) was associated with Black (OR 1.39, 95% CI: 1.06-1.81) and South Asian (OR 1.54, 95% CI: 1.26-1.88) ethnicity compared to Whites. Less than 50% of diabetics with CKD met the target blood pressure (BP) of 130/80 mmHg. The prevalence of a blood pressure > 150/90 mmHg in diabetics with CKD was South Asian 15.6%, White 13.9%, Black 21.8% (P < 0.001). Proteinuria was present in 8.6% of all diabetic patients. However, this increased to 18.6% in patients with CKD, and was more frequent in Black (22.6%) and South Asian (21%) patients compared to White patients (14.1%) (P < 0.001). CONCLUSION Significant disparities exist between the major ethnic groups in both disease prevalence and management. Future studies examining the management of CKD need to take variation by ethnicity into account.

Neighbourhood deprivation and the price and availability of fruit and vegetables in Scotland

BACKGROUND Previous research has suggested that fruits and vegetables are more expensive and less readily available in more deprived communities. However, this evidence is mainly based on small samples drawn from specific communities often located in urban settings and thus is not generalisable to national contexts. The present study explores the influence of neighbourhood deprivation and local retail structure on the price and availability of fruit and vegetables in a sample of areas representing the diversity of urban-rural environments across Scotland, UK. METHODS A sample of 310 stores located in 10 diverse areas of Scotland was surveyed and data on the price and availability of a basket of 15 fruit and vegetable items were collected. The data were analysed to identify the influence of store type and neighbourhood deprivation on the price and availability of fruits and vegetables. RESULTS Neighbourhood deprivation and store type did not significantly predict the price of a basket of fruit and vegetables within the sample, although baskets did decrease in price as store size increased. The highest prices were found in the smallest stores located in the most deprived areas. Availability of fruit and vegetables is lower in small shops located within deprived neighbourhoods compared to similar shops in affluent areas. Overall, availability increases with increasing store size. CONCLUSIONS Availability of fruit and vegetables significantly varies by neighbourhood deprivation in small stores. Policies aimed at promoting sales of fruit and vegetable in these outlets may benefit residents in deprived areas.

Inequalities in social capital and health between people with and without disabilities

The poor mental and physical health of people with disabilities has been well documented and there is evidence to suggest that inequalities in health between people with and without disabilities may be at least partly explained by the socioeconomic disadvantage (e.g. low education, unemployment) experienced by people with disabilities. Although there are fewer studies documenting inequalities in social capital, the evidence suggests that people with disabilities are also disadvantaged in this regard. We drew on Bourdieus conceptualisation of social capital as the resources that flow to individuals from their membership of social networks. Using data from the General Social Survey 2010 of 15,028 adults living in private dwellings across non-remote areas of Australia, we measured social capital across three domains: informal networks (contact with family and friends); formal networks (group membership and contacts in influential organisations) and social support (financial, practical and emotional). We compared levels of social capital and self-rated health for people with and without disabilities and for people with different types of impairments (sensory and speech, physical, psychological and intellectual). Further, we assessed whether differences in levels of social capital contributed to inequalities in health between people with and without disabilities. We found that people with disabilities were worse off than people without disabilities in regard to informal and formal networks, social support and self-rated health status, and that inequalities were greatest for people with intellectual and psychological impairments. Differences in social capital did not explain the association between disability and health. These findings underscore the importance of developing social policies which promote the inclusion of people with disabilities, according to the varying needs of people with different impairments types. Given the changing policy environment, ongoing monitoring of the living circumstances of people with disabilities, including disaggregation of data by impairment type, is critical.

The maternal health outcomes of paid maternity leave: A systematic review

Paid maternity leave has become a standard benefit in many countries throughout the world. Although maternal health has been central to the rationale for paid maternity leave, no review has specifically examined the effect of paid maternity leave on maternal health. The aim of this paper is to provide a systematic review of studies that examine the association between paid maternity leave and maternal health. We conducted a comprehensive search of electronic databases (Medline, Embase, CINAHL, PsycINFO, Web of Science, Sociological Abstracts) and Google Scholar. We searched websites of relevant organisations, reference lists of key papers and journals, and citation indices for additional studies including those not in refereed journals. There were no language restrictions. Studies were included if they compared paid maternity leave versus no paid maternity leave, or different lengths of paid leave. Data were extracted and an assessment of bias was performed independently by authors. Seven studies were identified, with participants from Australia, Sweden, Norway, USA, Canada, and Lebanon. All studies used quantitative methodologies, including cohort, cross-sectional, and repeated cross-sectional designs. Outcomes included mental health and wellbeing, general health, physical wellbeing, and intimate partner violence. The four studies that examined leave at an individual level showed evidence of maternal health benefits, whereas the three studies conducting policy-level comparisons reported either no association or evidence of a negative association. The synthesis of the results suggested that paid maternity leave provided maternal health benefits, although this varied depending on the length of leave. This has important implications for public health and social policy. However, all studies were subject to confounding bias and many to reverse causation. Given the small number of studies and the methodological limitations of the evidence, longitudinal studies are needed to further clarify the effects of paid maternity leave on the health of mothers in paid employment.

AutoDensity: an automated method to measure mammographic breast density that predicts breast cancer risk and screening outcomes

IntroductionWhile Cumulus – a semi-automated method for measuring breast density – is utilised extensively in research, it is labour-intensive and unsuitable for screening programmes that require an efficient and valid measure on which to base screening recommendations. We develop an automated method to measure breast density (AutoDensity) and compare it to Cumulus in terms of association with breast cancer risk and breast cancer screening outcomes.MethodsAutoDensity automatically identifies the breast area in the mammogram and classifies breast density in a similar way to Cumulus, through a fast, stand-alone Windows or Linux program. Our sample comprised 985 women with screen-detected cancers, 367 women with interval cancers and 4,975 controls (women who did not have cancer), sampled from first and subsequent screening rounds of a film mammography screening programme. To test the validity of AutoDensity, we compared the effect estimates using AutoDensity with those using Cumulus from logistic regression models that tested the association between breast density and breast cancer risk, risk of small and large screen-detected cancers and interval cancers, and screening programme sensitivity (the proportion of cancers that are screen-detected). As a secondary analysis, we report on correlation between AutoDensity and Cumulus measures.ResultsAutoDensity performed similarly to Cumulus in all associations tested. For example, using AutoDensity, the odds ratios for women in the highest decile of breast density compared to women in the lowest quintile for invasive breast cancer, interval cancers, large and small screen-detected cancers were 3.2 (95% CI 2.5 to 4.1), 4.7 (95% CI 3.0 to 7.4), 6.4 (95% CI 3.7 to 11.1) and 2.2 (95% CI 1.6 to 3.0) respectively. For Cumulus the corresponding odds ratios were: 2.4 (95% CI 1.9 to 3.1), 4.1 (95% CI 2.6 to 6.3), 6.6 (95% CI 3.7 to 11.7) and 1.3 (95% CI 0.9 to 1.8). Correlation between Cumulus and AutoDensity measures was 0.63 (P < 0.001).ConclusionsBased on the similarity of the effect estimates for AutoDensity and Cumulus in models of breast density and breast cancer risk and screening outcomes, we conclude that AutoDensity is a valid automated method for measuring breast density from digitised film mammograms.

Intersections between disability, type of impairment, gender and socio-economic disadvantage in a nationally representative sample of 33,101 working-aged Australians

BACKGROUND People with disabilities are socio-economically disadvantaged and have poorer health than people without disabilities; however, little is known about the way in which disadvantage is patterned by gender and type of impairment. OBJECTIVES 1. To describe whether socio-economic circumstances vary according to type of impairment (sensory and speech, intellectual, physical, psychological and acquired brain injury). 2. To compare levels of socio-economic disadvantage for women and men with the same impairment type. METHODS We used a large population-based disability-focused survey of Australians, analyzing data from 33,101 participants aged 25-64. Indicators of socio-economic disadvantage included education, income, employment, housing vulnerability, and multiple disadvantage. Stratified by impairment type, we estimated: the population weighted prevalence of socio-economic disadvantage; the relative odds of disadvantage compared to people without disabilities; and the relative odds of disadvantage between women and men. RESULTS With few exceptions, people with disabilities fared worse for every indicator compared to people without disability; those with intellectual and psychological impairments and acquired brain injuries were most disadvantaged. While overall women with disabilities were more disadvantaged than men, the magnitude of the relative differences was lower than the same comparisons between women and men without disabilities, and there were few differences between women and men with the same impairment types. CONCLUSIONS Crude comparisons between people with and without disabilities obscure how disadvantage is patterned according to impairment type and gender. The results emphasize the need to unpack how gender and disability intersect to shape socio-economic disadvantage.

Introduction to causal diagrams for confounder selection

In respiratory health research, interest often lies in estimating the effect of an exposure on a health outcome. If randomization of the exposure of interest is not possible, estimating its effect is typically complicated by confounding bias. This can often be dealt with by controlling for the variables causing the confounding, if measured, in the statistical analysis. Common statistical methods used to achieve this include multivariable regression models adjusting for selected confounding variables or stratification on those variables. Therefore, a key question is which measured variables need to be controlled for in order to remove confounding. An approach to confounder‐selection based on the use of causal diagrams (often called directed acyclic graphs) is discussed. A causal diagram is a visual representation of the causal relationships believed to exist between the variables of interest, including the exposure, outcome and potential confounding variables. After creating a causal diagram for the research question, an intuitive and easy‐to‐use set of rules can be applied, based on a foundation of rigorous mathematics, to decide which measured variables must be controlled for in the statistical analysis in order to remove confounding, to the extent that is possible using the available data. This approach is illustrated by constructing a causal diagram for the research question: ‘Does personal smoking affect the risk of subsequent asthma?’. Using data taken from the Tasmanian Longitudinal Health Study, the statistical analysis suggested by the causal diagram approach was performed.

Comparison of fully and semi-automated area-based methods for measuring mammographic density and predicting breast cancer risk.

Background:Mammographic density is a strong risk factor for breast cancer but the lack of valid fully automated methods for quantifying it has precluded its use in clinical and screening settings. We compared the performance of a recently developed automated approach, based on the public domain ImageJ programme, to the well-established semi-automated Cumulus method.Methods:We undertook a case-control study within the intervention arm of the Age Trial, in which ∼54 000 British women were offered annual mammography at ages 40–49 years. A total of 299 breast cancer cases diagnosed during follow-up and 422 matched (on screening centre, date of birth and dates of screenings) controls were included. Medio-lateral oblique (MLO) images taken closest to age 41 and at least one year before the index case’s diagnosis were digitised for each participant. Cumulus readings were performed in the left MLO and ImageJ-based readings in both left and right MLOs. Conditional logistic regression was used to examine density–breast cancer associations.Results:The association between density readings taken from one single MLO and breast cancer risk was weaker for the ImageJ-based method than for Cumulus (age–body mass index-adjusted odds ratio (OR) per one s.d. increase in percent density (95% CI): 1.52 (1.24–1.86) and 1.61 (1.33–1.94), respectively). The ImageJ-based density–cancer association strengthened when the mean of left–right MLO readings was used: OR=1.61 (1.31–1.98).Conclusions:The mean of left–right MLO readings yielded by the ImageJ-based method was as strong a predictor of risk as Cumulus readings from a single MLO image. The ImageJ-based method, using the mean of two measurements, is a valid automated alternative to Cumulus for measuring density in analogue films.

Mammographic density and markers of socioeconomic status: a cross-sectional study

BackgroundSocioeconomic status (SES) is known to be positively associated with breast cancer risk but its relationship with mammographic density, a marker of susceptibility to breast cancer, is unclear. This study aims to investigate whether mammographic density varies by SES and to identify the underlying anthropometric, lifestyle and reproductive factors leading to such variation.MethodsIn a cross-sectional study of mammographic density in 487 pre-menopausal women, SES was assessed from questionnaire data using highest achieved level of formal education, quintiles of Census-derived Townsend scores and urban/rural classification of place of residence. Mammographic density was measured on digitised films using a computer-assisted method. Linear regression models were fitted to assess the association between SES variables and mammographic density, adjusting for correlated variables.ResultsIn unadjusted models, percent density was positively associated with SES, with an absolute difference in percent density of 6.3% (95% CI 1.6%, 10.5%) between highest and lowest educational categories, and of 6.6% (95% CI -0.7%, 12.9%) between highest and lowest Townsend quintiles. These associations were mainly driven by strong negative associations between these SES variables and lucent area and were attenuated upon adjustment for body mass index (BMI). There was little evidence that reproductive factors explained this association. SES was not associated with the amount of dense tissue in the breast before or after BMI adjustment. The effect of education on percent density persisted after adjustment for Townsend score. Mammographic measures did not vary according to urban/rural place of residence.ConclusionsThe observed SES gradients in percent density paralleled known SES gradients in breast cancer risk. Although consistent with the hypothesis that percent density may be a mediator of the SES differentials in breast cancer risk, the SES gradients in percent density were mainly driven by the negative association between SES and BMI. Nevertheless, as density affects the sensitivity of screen-film mammography, the higher percent density found among high SES women would imply that these women have a higher risk of developing cancer but a lower likelihood of having it detected earlier.