Papers

Intersectional inequalities in advanced stage diagnosis of colorectal cancer in England: a cross-sectional study of National Cancer Registry data from 2013 to 2019

Background Inequalities in colorectal cancer (CRC) staging and outcomes exist across numerous sociodemographic axes. Early-stage CRC diagnosis is important for treatment success and survival. In this study, we investigate inequalities in CRC staging using registry data for 186 713 first-time CRC cancer diagnoses from 2013 to 2019 in England.

Methods We employ the novel Multilevel Analysis of Individual Heterogeneity and Discriminatory Accuracy (MAIHDA) approach to National Cancer Registry data. We investigate inequalities in CRC staging (early vs advanced stage) via a logistic MAIHDA. We examine a range of intersectional inequalities in CRC staging, across different age, ethnicity, gender and area-level deprivation groups.

Results Just over half of the staged cancers in the sample were diagnosed at advanced stage (62%). Results demonstrate notable inequalities in the risk of advanced CRC staging, with a gap of 17 percentage points between the strata with the lowest and highest predicted probability of advanced stage CRC diagnosis. These inequalities exist between age groups, ethnicity and deprivation level, with no evidence of gender-related inequalities when other variables are controlled. However, unexpectedly, we find these inequalities to be almost entirely additive in nature.

Conclusions These results suggest substantial inequalities in advanced stage CRC diagnosis exist, but that these are driven largely by universal processes of inequality, rather than disadvantages associated with single intersectional strata beyond an additive layering of disadvantage. Policy tools to encourage prompt screening engagement and symptom awareness campaigns in pre-screening age groups may benefit from considering the groups most disadvantaged by that additive layering.

Click here to read the paper.

The association between macro-level structural discrimination and alcohol outcomes: A systematic review

Bright, S., Buckley, C., Holman, D., Henney, M., Tabb, L. P. and Purhouse, R. Social Science and Medicine, 2025.

Alcohol consumption is a major risk factor for death and disability, disproportionately harming disadvantaged groups. While a positive association between interpersonal discrimination and alcohol use is established, structural discrimination’s impact remains unclear. We conducted a systematic review of the association between macro-level structural discrimination and alcohol consumption or related health outcomes. We searched four databases and grey literature, identifying 25 eligible studies, which considered racism (n = 11), sexism (n = 7), heterosexism (n = 4), and intersectional discrimination (n = 3). Most considered alcohol consumption (n = 17); fewer addressed harm (n = 4) or both (n = 4). The majority were US-based (n = 21), with four making cross-country comparisons. Associations varied by discrimination type, exposure measurement, alcohol outcome, and sociodemographic factors, though differential effects by sociodemographic factors remain understudied.

Most structural racism studies considered segregation as the exposure, but findings were inconsistent, even when grouped by outcome. Emerging evidence suggests increased race-based poverty ratios and incarceration gaps are associated with higher consumption and harm, especially for Black and Hispanic populations. Studies of structural sexism often used composite measures, like state-level gender inequality indices. Evidence suggests that as gender equality increases, women are more likely to drink, while greater structural sexism may be linked to higher rates of risky drinking and alcohol-related mortality. Findings on heavy episodic drinking and drinking frequency were mixed, while associations with volume and quantity were mostly non-significant. The limited available evidence suggests structural heterosexism may be positively associated with high intensity drinking and alcohol use disorders among sexually minoritized groups. The simultaneous impact of multiple forms of structural discrimination remains underexplored. Advancing this field requires consensus on how to operationalize structural discrimination within alcohol studies and greater adoption of intersectional and longitudinal approaches.

Click here to read the paper.

The Statistical Advantages of Multilevel Analysis of Individual Heterogeneity and Discriminatory Accuracy for Estimating Intersectional Inequalities

Leckie, G., Bell, A., Merlo, J. and Evans, C. SoxArXiv, 2025.

Multilevel Analysis of Individual Heterogeneity and Discriminatory Accuracy (MAIHDA) is a multilevel regression approach grounded in intersectionality theory. It examines inequalities across intersections of social identities (e.g., gender, ethnicity, class) and is argued to provide more accurate predictions of intersectional means than conventional methods that estimate group means directly or via regressions with all interactions. This study evaluates that claim using analytic expressions and an empirical illustration to compare simple and MAIHDA-predicted means against population values. Predictive accuracy is assessed via variance, correlation, bias, and mean squared error. Results show that MAIHDA estimates generally outperform simple means, particularly when decomposing intersectional means into additive and non-additive identity effects. The magnitude of the advantage depends on inequality patterns and group sample sizes. MAIHDA is especially valuable when inequalities are subtle or data for marginalized intersections are sparse—conditions common in practice. These findings highlight MAIHDA’s practical relevance for quantitative intersectionality research.

Click here to read the paper.

Examining variation within Hispanic ethnicity: An intersectional multilevel analysis of individual heterogeneity and discriminatory accuracy (MAIHDA) of birthweight inequities in New York City, 2012–2019

Borrell, L.N., Nieves, C.I. and Evans, C.R. Social Science and Medicine Population Health, 2025.

Birthweight inequities in the United States have been persistent with variations observed across maternal age, race/ethnicity, education, and nativity status. However, the Hispanic/Latino population is often treated as a monolithic category, ignoring within-group diversity and heterogeneity of health outcomes. This study employed an intersectional MAIHDA (multilevel analysis of individual heterogeneity and discriminatory accuracy) to examine birthweight inequities among singleton births in New York City from 2012 to 2019 (n = 819,920 records of singleton births) by maternal age, race/ethnicity, education, and nativity status, with particular attention to within-group heterogeneity among Hispanic/Latino mothers. Birthweight was measured in grams and was considered continuous for analytical purposes. The analysis was conducted using both the aggregate “Hispanic” category and disaggregated into Hispanic subgroups, based on the country/region of birth as part of the racial/ethnic categories. We found that 2.7%–3.2% of the variation in birthweight means among NYC women lies between intersectional strata, suggesting both meaningful between-stratum birthweight inequities and a high-degree of between-person birthweight variation. The finding of between-stratum inequities is consistent with a difference of 384.7g. between strata with the highest and lowest predicted birthweight means. We found consistent additive inequity patterns in birthweight by maternal age, race/ethnicity, educational attainment, and nativity status. The latter explained 81.2% of the variation in birthweight inequities. While attention to subgroup differences is often limited by sample size, intersectional MAIHDA allows for the identification of between- and within-strata variations regardless of whether Hispanic ethnicity was treated as an aggregate or subgroup based on country/region of origin.

Click here to read the paper.

Making sense of MAIHDA’s history and goals: A response to “Variance partition that eludes intuition”

Borrell, L.N., Nieves, C.I. and Evans, C.R. Social Science and Medicine Population Health, 2025.

This commentary is a response to Kaufman’s (2025) editorial on Borrell et al. (2025), which discusses the application of MAIHDA (Multilevel Analysis of Individual Heterogeneity and Discriminatory Accuracy) to birthweight disparities. The authors thank Kaufman for his engagement but offer several clarifications. First, they correct a common misunderstanding regarding the origins of MAIHDA. While Merlo (2014) introduced the acronym and conceptual framing, the statistical innovation—modeling intersectional strata as level-2 units in a multilevel model—was developed by Evans (2015) and formalized in Evans et al. (2018). Second, the authors clarify their interpretation of a small variance partition coefficient (VPC ≈ 3%), which Kaufman views as limiting for individual-level prediction. They emphasize that their population-level perspective focuses on structural inequities between strata, consistent with Rose’s (1985) insight that small average differences can yield large disparities in population outcomes. Finally, they respond to Kaufman’s concern about the focus on race and ethnicity, arguing that this reflects a difference in analytic purpose rather than disagreement. They underscore the importance of continuing to examine intersectional social determinants of health -particularly in the face of political efforts to silence such work – and highlight MAIHDA’s potential to support equity-focused research at both individual and population levels.

Click here to read the paper.

An analysis of intersectional disparities in alcohol consumption in the US

Bright, S., Buckley, C., Leckie, G., Bell, A., Mulia, N., Kilian, C. and Purshouse, R. Social Science and Medicine, 2024.

Alcohol is one of the leading causes of preventable deaths in the United States (US). Prior research has demonstrated that alcohol consumption and related mortality are socially patterned; however, no study has investigated intersectional disparities in alcohol consumption, i.e., attending to how social positions overlap and interact. In this study, we used an innovative intersectional approach (Multilevel Analysis of Individual Heterogeneity and Discriminatory Accuracy, MAIHDA) and data from a large nationally representative survey (the National Health Interview Survey, 2000-2018) to quantify inter-categorical disparities in alcohol consumption in the US (proportion of current drinkers, and average consumption amongst drinkers), along dimensions of sex, race and ethnicity, age, and level of education. Our analysis revealed significant intersectional disparities in both the prevalence of drinking and the average consumption by drinkers. Young, highly educated White men were the most likely to be current drinkers and consumed the highest amounts of alcohol on average, whilst racially and ethnically minoritized women with lower education were the least likely to drink and had the lowest levels of alcohol consumption, across all age categories. Notably, we found significant interaction effects for many intersectional strata, with much higher consumption estimated for some groups than traditional additive approaches would suggest. By identifying specific understudied groups with high consumption, such as young American Indian or Alaska Native (AI/AN) men, adult Black men with low education, and older White women with high education, this analysis has important implications for future research, policy, and praxis. This is the first known application of MAIHDA to account for a skewed outcome, highlighting and addressing critical methodological considerations.

Click here to read the paper.

Sociodemographic Inequalities in Student Achievement: An Intersectional Multilevel Analysis of Individual Heterogeneity and Discriminatory Accuracy (MAIHDA)

Prior, L; Evans, C: Leckie, G; Sage Journals

Sociodemographic inequalities in student achievement are a persistent concern for education systems and are increasingly recognized to be intersectional. Intersectionality considers the multidimensional nature of disadvantage, appreciating the interlocking social determinants which shape individual experience. Intersectional multilevel analysis of individual heterogeneity and discriminatory accuracy (MAIHDA) is a new approach developed in population health but new to educational research. In this study, we introduce and apply this approach to study sociodemographic inequalities in student achievement across two cohorts of students in London, England. We define 144 intersectional strata arising from combinations of student age, gender, free school meal status, special educational needs, and ethnicity. We find substantial stratum-level variation in achievement composed primarily by additive rather than interactive effects with results stubbornly consistent across the two cohorts. We conclude that policymakers should pay greater attention to multiply marginalized students and intersectional MAIHDA provides a useful approach to study their experiences.

Click here to read the paper.

Extending intersectional multilevel analysis of individual heterogeneity and discriminatory accuracy (MAIHDA) to study individual longitudinal trajectories, with application to mental health in the UK

Bell, A., Evans, C.R., Holman, D. and Leckie, G. Social Science and Medicine, 2024.

The Multilevel Analysis of Individual Heterogeneity and Discriminatory Accuracy (MAIHDA) approach is gaining recognition for identifying intersectional inequalities in various outcomes. Despite the availability of longitudinal data, this approach has not been applied to such data. Drawing on intersectionality and life course theories, the development of a longitudinal version of the intersectional MAIHDA approach is discussed. This approach enables the analysis of life course trajectories and intersectional inequalities, with a focus on changeable intersectional groups, generational differences, and the age-period-cohort identification problem. The article illustrates the approach with a study of mental health using United Kingdom Household Longitudinal Study data (2009–2021), revealing important differences in trajectories between generations and intersectional strata. The results show that trajectories are partly multiplicative but mostly additive in their intersectional inequalities. The paper provides a significant methodological contribution for rigorous quantitative, longitudinal, intersectional analyses in social epidemiology and beyond.

Click here to read the paper.  

Clarifications on the intersectional MAIHDA approach: a conceptual guide and response to Wilkes and Karimi (2024)

Evans, C.R., Borrell, L.N., Bell, A., Holman, D., Subramanian, S.V. and Leckie, G. Social Science and Medicine, 2024.

The MAIHDA method has gained widespread acceptance as a gold standard for evaluating intersectional inequalities in the fields of health and social sciences. However, Wilkes and Karimi (2024) have expressed methodological concerns about this approach, advocating for the continued use of conventional single-level linear regression models with fixed-effects interaction parameters for intersectional analysis. In response, we systematically address and debunk these concerns, identifying them as stemming from misunderstandings of the MAIHDA approach and literature. To aid those new to MAIHDA, we provide clarifications on four key points: the level at which additive main effect variables are defined in intersectional MAIHDA models, potential issues with collinearity in MAIHDA models, the reasons behind the small Variance Partitioning Coefficient (VPC) and large Proportional Change in Variance (PCV) in MAIHDA, and the overarching goals of MAIHDA analysis.

Click here to read the paper.  

A tutorial for conducting intersectional multilevel analysis of individual heterogeneity and discriminatory accuracy (MAIHDA)

Evans, C.R., Leckie, G., Subramanian, S.V., Bell, A. and Merlo, J. Social Science and Medicine Population Health, 2024.

Intersectional multilevel analysis of individual heterogeneity and discriminatory accuracy (I-MAIHDA) is an innovative approach for investigating intersectional inequalities in various outcomes, offering conceptual and methodological advantages over conventional single-level regression analysis. It enables the study of inequalities caused by multiple intersecting systems of marginalization and oppression, making it a valuable tool in various fields such as social epidemiology, health psychology, precision medicine, public health, and environmental justice. The approach allows estimation of average differences between intersectional strata, exploration of interaction effects, and decomposition of individual variation in outcomes. Specific guidance for conducting and interpreting MAIHDA models is consolidated into an accessible tutorial covering both continuous and binary outcomes, with an emphasis on I-MAIHDA while being informative for related approaches such as multicategorical MAIHDA. The tutorial provides step-by-step analytical advice and an illustrative health application using simulated data, along with the necessary data and syntax for replication of the analyses.

Click here to read the paper.  

Geographical and sociodemographic differences in statin dispensation after acute myocardial infarction in Sweden: a register-based prospective cohort study applying analysis of individual heterogeneity and discriminatory accuracy (AIHDA) for basic comparisons of healthcare quality

Merlo, J., Öberg, J., Khalaf, K., Perez-Vicente, R. and Leckie, G. BMJ Open, 2023.

In Sweden, as in many other countries, official monitoring of healthcare quality is mostly focused on geographical disparities in relation to a desirable benchmark. However, current evaluations could be improved by considering: (1) The intersection of other relevant axes of inequity like age, sex, income and migration status; and (2) The existence of individual heterogeneity around averages. Therefore, using an established quality indicator (i.e. dispensation of statins after acute myocardial infarction, AMI), we valuate both geographical and sociodemographic inequalities and illustrate how the analysis of individual heterogeneity and discriminatory accuracy (AIHDA) enhances such evaluations

Click here to read the paper.  

Neighbourhood deprivation and intersectional inequalities in biomarkers of healthy ageing in England

Holman, D., Bell, A., Green, M. and Salway, S. Health & Place, 2022.

While social and spatial determinants of biomarkers have been reported, no previous study has examined both together within an intersectional perspective. We present a novel extension of quantitative intersectional analyses using cross-classified multilevel models to explore how intersectional positions and neighbourhood deprivation are associated with biomarkers, using baseline UK Biobank data (collected from 2006 to 2010). Our results suggest intersectional inequalities in biomarkers of healthy ageing are mostly established by age 40–49, but different intersections show different relationships with deprivation. Our study suggests that certain biosocial pathways are more strongly implicated in how neighbourhoods and intersectional positions affect healthy ageing than others.

Click here to read the paper.

Intersectionality: buzzword or key to solving health inequalities?

Holman, D., Salway, S. and Bell, A. Fuse Open Science Blog, 2021.

We have published a new blog on Fuse Open Science which summarises the paper we co-produced with stakeholders on the policy potential and challenges of an intersectional health perspective (Can Intersectionality Help with Understanding and Tackling Health Inequalities? Perspectives of Professional Stakeholders). We explore some of the key policy suggestions in this area and offer suggestions for the way forward.

Take a look here!

Can intersectionality help with understanding and tackling health inequalities? Perspectives of professional stakeholders

Holman, D., Salway, S., Bell, A., Beach, B., Adebajo, A., Ali, N., and Butt, J. Health Research Policy and Systems, 2020.

The concept of “intersectionality” is being increasingly used in public health arenas , with potential to advance health inequalities research and action. However, no existing research has explored professional stakeholder perspectives on applying intersectionality to this field. To address this gap, a consultation survey and face-to-face workshop were conducted in the United Kingdom. Findings indicated a generally positive response to the concept, though various challenges were raised, especially regarding its critical and transformative nature and methodological operationalization. Policy and practice professionals’ views suggested that intersectionality has potential to counter individualistic narratives and be sensitive to subgroup inequalities, but it still has far to go to gain traction in the United Kingdom.

Click here to read the paper.

Intersectionality and the life course

Holman, D. and Walker, A. Ageing Issues Blog, 2020.

We published an accompanying blog to the below paper on unequal ageing on the British Society of Gerontology’s Ageing Issues. The blog argued that it is axiomatic that the life course perspective is fundamental to understanding unequal ageing. People move through various life stages as they age, experiencing different social roles and relationships with others, who are each doing the same. Social and cultural processes and policy encounters provide the context for these experiences, shaping what is possible, and the attendant life chances. Life course researchers have shown that ageing is unequal with respect to a number of key axes of dis/advantage such as social class, gender, and ethnicity – but what about the ways in which these axes of dis/advantage overlap and interact with each other?

Take a look here.

Understanding unequal ageing: towards a synthesis of intersectionality and life course analyses

Holman, D. and Walker, A. European Journal of Ageing, 2020.

This paper aims to illustrate how intersectionality might be synthesised with a life course perspective to deliver novel insights into unequal ageing, especially with respect to health. From the intersectionality literature, it focusses on the concepts of intersectional subgroups, discrimination, categorisation, and individual heterogeneity, and from the life course literature, roles, life stages, transitions, age/cohort, cumulative disadvantage/advantage, and trajectories. Synergies between these concepts hold exciting opportunities to bring new insights to unequal ageing.

Click here to read the paper.

Mapping intersectional inequalities in biomarkers of healthy ageing and chronic disease in older English adults

Holman, D., Bell, A. and Salway, S. Nature Scientific Reports, 2020.

This paper analyses intersectional inequalities in biomarkers of healthy ageing and chronic disease, in older adults using data pooled from the English Longitudinal Study of Ageing and the UK Household Longitudinal Study. It finds granular inequalities that vary according to biomarker. These inequalities are additive rather than multiplicative in nature and have significant clinical implications.

Click here to read the paper.

Using multilevel models to understand intersectionality: a simulation study and guide for best practice

Bell, A., Holman, D. and Jones, K. Methodology, 2019.

This paper uses simulations to test a novel multilevel approach (termed ‘MAIHDA’) to intersectionality, comparing it with conventional regression approaches which typically use interaction terms. It finds that although this new approach needs further work and development to improve its statistical properties, it is overall an improvement on conventional approaches.

Click here to read the paper.