In a short historical review of deaths of Black men restrained in psychiatric care, Meju reminds us of the ‘Big, Black and Dangerous’ stereotype. Reference Meju1 Citing studies of biases in diagnosis and judgements about risk, she advocates for mandatory training on biases in clinical practice, alongside challenge by peers to expose and mitigate biases. There have been calls for better cultural competency training for many years. Reference Lok2–Reference Braybrook, Shephard, Wall and Binder4 The evidence base on take-up and impacts is less convincing.5,6 More recent advances propose cultural formulation and cultural consultation as ways of improving practice. Reference Aggarwal7,Reference Lam, Lewis-Fernandez and Aggarwal8,Reference Bhui, Owiti, Palinski, Ascoli, De Jongh and Archer9
Racism and prejudice are population-level risk factors for poor mental health, alongside adverse childhood experiences (ACEs) and adverse life events. Reference Francis-Crossley, Hudson, Harris, Onwumere and Kirkbride10–Reference Varchmin, Montag, Treusch, Kaminski and Heinz12 In addition, social position, poverty, unemployment, unsafe places and unregulated emotional distress if there is an absence of parental or carer support can all combine to produce trajectories of poor mental health. Reference Kirkbride, Anglin, Colman, Dykxhoorn, Jones and Patalay13 These broader eco-social-cultural models of causation require preventive actions that respond to dynamic and multiple levels of poor health and inequalities. Reference Gomez-Carrillo, Kirmayer, Aggarwal, Bhui, Fung and Kohrt14
The argument often goes that minority ethnic groups, racialised populations and migrants carry a greater cumulative burden of lifetime adverse experiences, making them more likely to develop poor mental health, and then they find that the interventions offered do not adequately address the complexity of their experiences. Consequently, in the absence of any other care and the presence of stereotyped concerns about risk, they receive more coercive care. Furthermore, persistent social disadvantages undermine trust and take-up and delivery of effective care. These structural drivers need understanding, preventive efforts and structurally competent care systems and policies.
Debates about translating structural disadvantage interventions into clinical interventions risk ecological and structural fallacies, as structures generate affordances within which there is marked individual and group-level variation of symptoms. Each person (or subgroup) will have specific experiences over their lifetime and present related symptoms that require a tailored response. It is at this clinical and individual and ethnic group level, Meju argues, that biases may enter assessment and management in crisis, leading to excessive restraint and insufficient safeguards to protect life and ensure humane and effective care.
Minority ethnic and marginalised patients less often receive and engage with psychotherapies and more often receive medication or coercive care, yet interestingly, for treatment-resistant schizophrenia (for Black patients at least), they are less likely to be offered clozapine, despite it being the treatment of choice in such situations. Reference de Freitas, Patel, Kadra-Scalzo, Pritchard, Shetty and Broadbent15 In this context, numerous inquiries investigate the deaths of Black men under restraint, as well as homicides by Black men who appear to not have received adequate care. There appears to be razor’s edge for decisions leading either to over-restraint and coercion risking harms to the patient when in the health system, or instances in which concerns about risk and safety to the public and patient are overlooked or somehow dismissed, potentially leading to poor health outcomes, premature mortality or even homicide.
Bias and stereotypes
Stereotypes of risk and dangerousness can lead to more coercive care, yet do they also lead to neglect and avoidance of responsibility, confused imprecise risk assessment, or even fears about accusations of racism? In the currently volatile political climate on anything to do with mental health and race, in both the USA and UK, it is hard to take in evidence and understand the pathways to poor outcomes without the debate becoming racialised. This manifests as binary explanations: either deaths in care are proposed to be due to racism, or reverse racism (overcorrection for anticipated racism) leads to neglect of good clinical practice.
Bias operating in both directions (more coercive care/patient fatalities versus neglect/public fatalities) is hard to reconcile, leading commentators to take up one or other position depending on their pre-existing expectations and beliefs. As for the suggestion that all the anti-racist and educational efforts (mostly ineffective at reducing inequalities) over decades are now so powerful that they lead to neglect of well-established clinical risk markers on grounds of race is also hard to credit as a potential explanation, not least because it goes against training, standards of practice and regulatory requirements of clinicians.
Structural and historical influences
The Royal College of Psychiatrists and many charities, as well as current ongoing inquiries, conclude that care systems in the UK (the level of bed provision, the mixture of service types and links between them, continuity of care) are inadequate and lessons have not been learned from previous inquiries. A minimum level of service adequacy, with specified models of effective evidenced care, is not difficult to define or recommend. We have much evidence of what works. Many helpful models have been de-implemented on grounds of cost or complexity or concerns of potential harms. Such reviews, not uncommonly, occur when new governments come into office, or a new health minister emerges, or the financial and political context shifts to force services to function on less resources. Changes to services and economic decisions should have ethical, clinical, equalities and risk impact assessments.
Financial and policy reviews in the UK have, over decades, eroded our belief in and commitment to effective care models – for example, assertive outreach or more intensive psychologically informed care environments. A lack of continuity between different service types – in-patient/out-patient, primary/secondary – and even between similar tier services in geographically distinct parts of the country, are failing the public and patients. Such experiences and decisions also place undue ethical and professional strain on practitioners, and this can lead to bias in decisions or incompetent decisions. We might reasonably argue that practitioners fear repercussions while working with a fragmented, understaffed and under-resourced services. In such instances, their ability to act think, plan, act are compromised. This might lead to bias in decisions, in either direction. More research on this is warranted.
Amid the climate of economic constraints and a revolution in digital and artificial intelligence (AI) prototypes, we should revisit how to improve care and make it more efficient and less costly, yet such innovations may well enact the same bias as humans if they are trained on existing practice. If neighbourhood care is to effectively prevent progression of illness and crises, we need to rapidly evolve and invest in community models to reduce the systems burden of acute and crisis care. We need an adequate and appropriate mix of skilled staff and connected services. For now more investment is needed for a minimally safe service.
One of the risks of focusing only on training to address racism is the simplification and bureaucratisation of cultural competency and equalities policies when scaled. This risks loss of the more nuanced skills of investigating the role of culture, weighing up different options and considered decision-making for individual people. Consequently, it can lead to an overreliance on stereotypes. Reference Johnson, Hickey, Switzer, Miller, Winger and Nguyen16,Reference Puddifoot17 Race biases in assessment and diagnosis and care are well-evidenced. Reference Teplin, Abram and Luna18,Reference Londono Tobon, Flores, Taylor, Johnson, Landeros-Weisenberger and Aboiralor19 An intriguing perspective on human decision-making is the work of Karl Friston on active inference, a natural predictive process deployed to minimise cognitive effort, maximise efficiency and avoid surprise. Reference Parr, Pezzulo and Friston20 Stereotypes, according to this analysis, are the norm. Prediction errors then should be rapidly updated, requiring more effort, resetting to a more accurate prediction process.
Prediction of rare events is especially difficult as risk assessment tools do not predict well. Reference Fazel, Singh, Doll and Grann21 Furthermore, clinical risk decisions, made many times a day, are cognitively demanding tasks and require expertise, judgement, service capacity, and effective and high team functioning. Reference Bhui22 We need to shift the emphasis back towards more intensive and rigorous training in decision-making and adequacy in safe service design and function.
About the author
Kamaldeep S. Bhui CBE, is Professor of Psychiatry in the Department of Psychiatry and Senior Research Fellow at Wadham College, University of Oxford, UK. He is also Honorary Consultant Psychiatrist at East London NHS Foundation Trust, London, UK, and Oxford NHS Foundation Trust, Oxford, UK.
Acknowledgement
Artificial Intelligence use: No artificial intelligence tools have been used in the research or writing process for this manuscript.
Funding
This work received no specific grant from any funding agency, commercial or not-for-profit sectors.
Declaration of interest
Bhui is Deputy Editor of BJPsychOpen and former Editor in Chief of BJPsych. He is Director of Research and Implementation in the Public Mental Health Implementation Centre, RCPsych, UK.
Open access
eLetters
No eLetters have been published for this article.