Article Text

Download PDFPDF

Moving upstream to address diagnostic disparities
  1. Denise M Connor1,2,
  2. Gurpreet Dhaliwal1,2
  1. 1 Department of Medicine, University of California San Francisco, San Francisco, California, USA
  2. 2 Medical Service, San Francisco VA Medical Center, San Francisco, California, USA
  1. Correspondence to Dr Denise M Connor, Department of Medicine, University of California San Francisco, San Francisco, California, USA; Denise.Connor{at}

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Twenty years after the National Academy of Medicine report, Unequal Treatment, called attention to inequity in the diagnostic process, many communities continue to receive less accurate and less timely diagnosis.1 2 In this issue of BMJ Quality & Safety, Herasevich and colleagues examined 20 studies encompassing more than 7 million patients presenting to the emergency department with cardiovascular or neurologic symptoms.3 Among six studies looking at cardiovascular diagnoses, the authors found that Black patients were more likely than White patients to have their acute coronary syndrome diagnosis missed; there was no consistent diagnostic discrepancy identified among patients with neurologic symptoms. While individual studies of both cardiovascular and neurologic symptoms demonstrated diagnostic disparities linked with ethnicity or insurance status, in pooled analysis, no consistent associations were identified. These findings add to the growing evidence base of diagnostic disparities across many clinical conditions such as delayed appendicitis diagnosis in Black children, overdiagnosis of schizophrenia in Black individuals, and delayed dementia diagnosis in Asian and Hispanic/Latinx individuals.4–9

Herasevich and colleagues highlight the variability in diagnostic error definitions (eg, an incorrect diagnostic label vs an extended time before diagnosis), approaches to measuring diagnostic error (eg, ‘look-forward’ vs ‘look-back’10) and grouping of demographics (eg, combining vs separating race and ethnicity) among the studies they analysed. This heterogeneity made it impossible to calculate pooled effect estimates, leading the authors to shift from a meta-analysis to a narrative synthesis-style review. Without consistent definitions and methodological approaches among researchers, it will be challenging to adequately address the complex and entrenched problem of diagnostic disparities.

This study also provides an opportunity to consider how problems of equity and disparities are framed in the literature. In science, language shapes the research questions we ask, the places we look for solutions and the areas where we invest our resources. Being intentional and precise in our language will improve the quality of our research efforts and interventions while also respecting the individuals and communities most impacted by disparities.

The authors, like many in the field of healthcare disparities research and policy, invoke ‘lack of trust in healthcare institutions’ as one causative factor in diagnostic inequity. This common formulation risks unintentionally locating the problem within individuals and communities who do not trust the system instead of within institutions who are untrustworthy.11 During the COVID-19 pandemic, limited vaccination uptake was often attributed to lack of trust, implying a need to ‘fix’ individuals and communities. When discussing factors that put populations at higher risk of diagnostic error, authors and commentators should consider focusing on the impact of ‘untrustworthy institutions’ rather than individuals’‘lack of trust.’ Institutions must make investments to earn trust, such as ensuring equitable access to care and improving informed consent processes.11

Many studies and scholars also use the term ‘vulnerable populations’ when speaking about groups that experience healthcare disparities. While the positive intention of this framing is to call out the urgent need to address inequities in healthcare, the term risks misattributing the nidus of the problem within populations and communities, rather than explicitly naming the social and structural determinants causing disparities. Dr Camara Jones argues that naming racism is the first critical step to working against it12; we can extend this principle to all forms of oppression. The Centers for Disease Control and Prevention’s health equity style guide advises against using terms like ‘vulnerable’, ‘marginalised’ and ‘high-risk’ as adjectives.13 Rather than invoking ‘high-risk populations’, we can describe populations disproportionately impacted by structural racism. If short-hand labels like vulnerable populations are used, ensuring that we have first clearly described and located the underlying issues behind that vulnerability can help prevent unintended messages that imply deficits inherent to communities.

True antiracism practice in healthcare involves moving beyond documenting disparities towards developing solutions. Researchers and healthcare leaders have an opportunity to learn from disability rights activists who championed the phrase ‘nothing about us without us’.14 Individuals and communities most impacted by oppressive forces must be represented as we develop a research agenda to eliminate diagnostic disparities. Community-based participatory research (CBPR) provides a roadmap for how to equitably partner with and learn from communities in research efforts.15 There are many examples of CBPR to guide us, from exploring the community impact of a contaminated water supply to partnering with school systems to tackle diabetes prevention.16 17 Without taking approaches that centre community stakeholders, healthcare leaders risk major missteps in understanding and addressing the complex causes of diagnostic disparities.

The findings shown by Herasevich and colleagues should also catalyse change in how we act as clinicians and educators. Active strategies to support equity should be embedded in our diagnostic processes.18 Diagnostic communication can be negatively impacted by bias and power differentials linked with social identities such as race, gender and economic class. To flatten the hierarchy and build strong connections with patients, clinicians can commit to practising relationship-centred communication skills.19 20 Individuation, the deliberate effort to learn about each patient as an individual, can facilitate diagnostic evaluation based on personal, not stereotyped, attributes.21 Additionally, clinicians taking an intentional pause in the diagnostic process to see if they would favour an alternative diagnosis if the patient had a different social identity can counteract biased reasoning.22

Clinicians can also work to influence systems-level changes such as advocating for healthcare institutions to collect and stratify data by racial and ethnic groups.23 It can be helpful to look at subgroups as well; for example, rather than having one category for Asian patients, analysing data about different Asian groups, such as Filipino and Vietnamese communities, can yield much more actionable insights. Extending this practice across multiple social groups and identities (eg, LGBTQ+ individuals) is critical to recognising and addressing diagnostic disparities.24 25

As educators, we can model antioppressive thinking and strategies for trainees. We must ensure that our teaching avoids suggesting that individuals, populations and communities are the source of healthcare disparities. Instead, a focus on the structural determinants of health will enable learners to grasp the root causes of diagnostic disparities. For example, when discussing risk factors for lung cancer, rather than exclusively focusing on individual behaviours such as smoking, we can discuss the ways that cigarette companies intentionally target certain communities in advertising and distribution.26 We can highlight the inequities built into the clinical calculators and tools we use to make diagnoses (eg, kidney function estimates and pulse oximetry measurements) that increase the risks of diagnostic disparities.27 28

How we frame problems determines whether we can successfully address them. Consistently naming the social and structural determinants of health which create diagnostic disparities is critical to our efforts. We can make equity in the diagnostic process central to our practice by advancing our communication skills, advocating for change and modelling these strategies with learners. Researchers, clinicians and educators—in partnership with patients and communities—have important roles to play in moving towards a healthcare system where all patients can be assured of timely and accurate diagnoses.

Ethics statements

Patient consent for publication

Ethics approval

Not applicable.



  • Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

  • Competing interests DMC is a contributor to a forthcoming professional development module on diagnostic disparities from the American College of Physicians. GD is a board member of the Society to Improve Diagnosis in Medicine.

  • Provenance and peer review Commissioned; internally peer reviewed.

Linked Articles