Using data science to understand how racial discrimination impacts mental health

In this research deep dive, Dr Patsy Irizar argues that policy changes are needed to tackle ethnic inequalities at all levels, including in education, employment, housing, and healthcare. She calls for collective action across organisations and institutions to combat racial discrimination in all forms, and that further training in mental health care is needed for mental health practitioners to recognise and consider the impact of racial discrimination on mental health. Finally, Patsy calls for mental health practitioners to adopt evidence-based interventions for racial stress and trauma.

Key Findings

• I found a dose-response relationship between racial discrimination and poor mental health (effect on mental health increased with increasing number of times and areas of life in which racism was experienced)

• Recent experiences of racial discrimination (within the past 5 years only) had a stronger effect on mental health compared to experiences of racial discrimination that happened over 5 years ago

• Chronic experiences of racial discrimination (both past and recent experiences) had the strongest effect on mental health

• Racial discrimination indirectly contributed to poor mental health, through a greater likelihood of testing positive for COVID-19, financial insecurity, feelings of loneliness, and a reduced sense of belonging

Data science can be a useful tool for building a more inclusive society. Politicians and practitioners often rely on statistics to understand health inequalities, yet when looking at ethnic inequalities in health, a lot of datasets are limited. This is because they tend to recruit samples which are representative of the United Kingdom (UK), meaning they collect data from smaller numbers of people from minoritised ethnic groups, which reduces statistical power when looking at ethnic differences. The Evidence for Equality National Survey (EVENS) was set up during the COVID-19 pandemic, by the Centre on the Dynamics of Ethnicity (CoDE) – in partnership with voluntary, community, and social enterprise organisations – to address these limitations. EVENS is the largest and most comprehensive survey of people from 20 minoritised ethnic groups.

I first used EVENS to identify the level of ethnic inequalities in mental health during the pandemic, finding much higher levels of anxiety among people identifying as Arab, Mixed White and Black Caribbean, any other Black ethnicity, and any other Mixed ethnicity, compared to White British people[1]. This blog focuses on my research which used EVENS to understand the ways in which racial discrimination contributes to poor mental health among minoritised ethnic groups, both directly and indirectly.

I aimed to identify the relationship between experiences of racial discrimination across the life course and mental health (during the COVID-19 pandemic), focussing both on the timing and type of racial discrimination experienced. I also aimed to understand how experiences of racial discrimination indirectly contribute to poor mental health, through health, economic, and social inequalities.  

Why is this research needed?

It is well-known that racial discrimination contributes to poor mental health among minoritised ethnic groups[2]. Racial discrimination can be directly harmful to mental health, as these experiences, or even fear of these experiences, create a stress response in the body which leads to chronically elevated cortisol levels – known indicators of depression and anxiety[3]. However, most existing studies have only focused on experiences of racial discrimination at any point in someone’s life, or only recent experiences. Yet, racism can be a chronic and pervasive part of life for minoritised ethnic groups and there is a need to understand the accumulation of exposure to racism and the impact this has on mental health.

Racism also occurs at structural and institutional levels, through discriminatory policies and practices that create and maintain unequal access to resources. This means that minoritised ethnic groups are more likely to experience socioeconomic disadvantage, which is known to be related to physical and mental health inequalities. The COVID-19 pandemic highlighted and worsened exiting inequalities for minoritised ethnic groups. It is now widely known that minoritised ethnic groups suffered disproportionately during the COVID-19 pandemic, in terms of both health and economic inequalities[4] [5]. I argue that racism is the fundamental driver of these inequalities.

I developed a theoretical model, hypothesising the ways in which racial discrimination contributes to poor mental health, through direct and indirect pathways. I then tested these theorised pathways using data from EVENS. I theorised that there would be a direct dose-response relationship between experiences of racial discrimination and poor mental health, with the size of the effect increasing with increasing exposure to racism. I also theorised that recent experiences of racial discrimination would have a stronger association with experiences of racial discrimination that happened over 5 years ago (which may have negatively impacted mental health at the time), and that racial discrimination would be indirectly associated with poor mental health, through health (COVID-19 infection), economic (financial worries), and social (loneliness and a low sense of belonging) inequalities. 

How did I test my theoretical model?

I used data from the novel Evidence for Equality National Survey (EVENS)[6], which recruited a greater proportion of people from minoritised ethnic groups than any other UK survey, as well as White British people. Data were collected between February and November 2021, during the second year of the COVID-19 pandemic. Participants could complete the survey online, on the phone, or face-to-face, in various languages. My analysis included people from minoritised ethnic groups, including white minoritised groups (e.g., Gypsy/Roma groups), aged 18 to 60 years old.

EVENS is unique in the way that it collected data, as the survey did not aim to be representative of the British populations, but instead over-sampled people from minoritised ethnic groups to ensure large enough sample sizes to look at differences across ethnic groups. The team monitored responses to the survey every day to check they had enough responses for each target group. The team also worked closely with voluntary, community, faith, and social enterprise organisations, who helped with recruitment, particularly where certain groups were under-represented in the survey. Statisticians then created statistical weights that can be applied when analysing the data (giving more weight to those with similar characteristics of the population), so that the data can be used in a way that is said to be representative of the British population.

One of the main strengths of EVENS is the comprehensive measures of racial discrimination, through 60 questions which captured experiences in various aspects of life and at several timepoints. I combined responses to these 60 questions to measure the accumulation of racial discrimination and differences depending on the timing and type of these experiences.

I first grouped the different types of experiences of racial discrimination into three domains:

• Hate crimes (e.g., property damaged, physically attacked)

• Interpersonal racial discrimination (e.g., being insulted, treated unfairly by friends/family/partner)

• Racial discrimination in institutional settings (e.g., treated unfairly in education, treated unfairly by the police).

  
  

I then grouped the timing of these experiences into the following categories:

• No reported experiences of racial discrimination

• Past experiences of racial discrimination only (over 5 years ago and not within the last 5 years)

• Recent experiences of racial discrimination only (within the last 5 years and not over 5 years ago)

• Chronic experiences of racial discrimination (both past and recent)

In my analysis, I examined the timing of any experience of racial discrimination, as well as looking at the timing of each domain of racial discrimination separately.

I also created an additional variable which captured the build-up of experiences of racial discrimination across time points and domains:

• No experiences of racial discrimination

• Racial discrimination in 1 domain at 1 or 2 time points

• Racial discrimination in 1 domain at 3 or 4 time points

• Racial discrimination in 2 domains at 1 or 2 time points

• Racial discrimination in 2 domains at 3 or 4 time points

• Racial discrimination in 3 domains at 1 or 2 time points

• Racial discrimination in 3 domains at 3 or 4 time points

My outcome of interest was self-reported symptoms of depression and anxiety, measured using widely used screening instruments. Participants who scored above the validated cut-offs for at least one measure were identified as having a mental health problem (depression and/or anxiety).

EVENS also included several questions about the health, economic, and social impacts of the COVID-19 pandemic. I included the following measures as potential pathways: previously tested positive for COVID-19 infection, changes in household income, worries about future financial security, feelings of loneliness and feelings of belonging to local area (and whether these feelings changed since the pandemic began).

To analyse the data, I first used regression models to look at the associations between my different measures of racial discrimination with mental health. I then used a statistical approach called Structural Equation Modelling (SEM) to test the direct and indirect effects (through the potential pathways) of racial discrimination on mental health.

What did I find?

Of the 8,897 participants (52% female) included in the analysis, 45% met the criteria for either depression or anxiety. Less than 30% reported no experiences of racial discrimination, with 16% reporting past experiences of racial discrimination (over 5 years ago) without recent experiences, 25% reported recent experiences of racial discrimination (within the last 5 years) without past experiences, and almost 30% reporting chronic experiences of racial discrimination (both past and recent experiences) over time.

When looking at cumulative exposure to racial discrimination, I found evidence of a dose-response relationship. This means that the likelihood of suffering from poor mental health increased with an increasing number of domains and time points that racial discrimination was experienced. Those who had experienced racial discrimination in all three domains and all time points were almost 5 times more likely to suffer from poor mental health than people who reported no experiences of racial discrimination.

There were differences depending on when people experienced racial discrimination. People who reported recent experiences of racial discrimination (within the last 5 years) were twice as likely to report a mental health problem, whereas people who reported past experiences of racial discrimination (over 5 years ago) were 16% more likely, compared to people who reported no experiences of racial discrimination. Chronic experiences of racial discrimination showed the strongest association, increasing the likelihood of poor mental health 3-fold. These findings reflect the total effects of racial discrimination on mental health (without controlling for potential pathways).

I then controlled for the potential pathways to test the direct effects of racial discrimination on mental. The strength of the association between recent and chronic experiences of racial discrimination and mental health decreased, and past experiences of racial discrimination were no longer associated with mental health. This suggests that health, economic, and social inequalities are important for explaining how racial discrimination might impact mental health.

I also found indirect effects of racial discrimination on mental health, through these pathways. Recent and chronic experiences of racial discrimination were indirectly associated with poor mental health, through an increased likelihood of testing positive for COVID-19 infection, greater concerns about future financial security, greater feelings of loneliness and isolation, and a low sense of belonging. These findings remained the same when looking at the different types of racial discrimination experienced (hate crimes, interpersonal, institutional).

What should be done?

My research provides strong evidence of a dose-response relationship between racial discrimination and poor mental health, with the likelihood of poor mental health increasing with the more times and aspects of life in which racial discrimination is experienced. This is concerning given that over 70% of people reported experiencing racism at least once during their life.

Racism leads to unequal access to social and economic resources, which is evidenced through persisting ethnic inequalities in education, housing, employment, and health. Ethnic inequalities in health and economic outcomes widened during the pandemic, and my research now shows that racial discrimination contributes to poor mental health, through these inequalities. I also show that racial discrimination contributes to feelings of loneliness and a reduced sense of belonging, and these are important factors to address to prevent poor mental health among minoritised ethnic groups.

I argue that policy changes are needed to tackle ethnic inequalities at all levels, including in education, employment, housing, and healthcare. Collective action is required across organisations and institutions to combat racial discrimination in all forms. In terms of changes to mental healthcare practice, I argue that further training is needed so that mental health practitioners recognise and consider the impact of racial discrimination on mental health, such as adopting evidence-based interventions for racial stress and trauma.

Read the whole paper here.

[1] Irizar, P., Taylor, H., Kapadia, D., Pierce, M., Bécares, L., Goodwin, L., ... & Nazroo, J. (2024). The prevalence of common mental disorders across 18 ethnic groups in Britain during the COVID-19 pandemic: Evidence for Equality National Survey (EVENS). Journal of Affective Disorders, 358, 42-51.

[2] Paradies, Y., Ben, J., Denson, N., Elias, A., Priest, N., Pieterse, A., ... & Gee, G. (2015). Racism as a determinant of health: a systematic review and meta-analysis. PloS one, 10(9), e0138511.

[3] Dziurkowska, E., & Wesolowski, M. (2021). Cortisol as a biomarker of mental disorder severity. Journal of Clinical Medicine, 10(21), 5204.

[4] Irizar, P., Pan, D., Kapadia, D., Bécares, L., Sze, S., Taylor, H., ... & Pareek, M. (2023). Ethnic inequalities in COVID-19 infection, hospitalisation, intensive care admission, and death: a global systematic review and meta-analysis of over 200 million study participants. EClinicalMedicine, 57.

[5] Platt, L., & Warwick, R. (2020). COVID‐19 and ethnic inequalities in England and Wales. Fiscal Studies, 41(2), 259-289.

[6] https://www.ethnicity.ac.uk/research/projects/evens/

Patsy is a Senior Lecturer in the School of Psychology, at Liverpool John Moores University, with expertise in research relating to mental health and health inequalities. During the COVID-19 pandemic, Patsy investigated the level of ethnic inequalities in COVID-19 health outcomes and explored the reasons for these inequalities. She has recently completed a research fellowship, where she examined the impact of the COVID-19 pandemic on the mental health of ethnic minority people. Her research is centred around the role of racism in contributing to ethnic inequalities in physical and mental health.

p.irizar@ljmu.ac.uk

@patsy_irizar