Publishing research on equity in healthcare

BMJ Quality and Safety has a long history of interest in publishing research on equity, diversity and inclusion (EDI) within healthcare. For example, in 1993 we published a review of how health and healthcare differed among ethnic groups in Britain.5 This concluded with an important recommendation to avoid categorising people simply as ‘ethnic minorities’, an almost useless term given the wide diversity of peoples, experiences and healthcare needs that may be represented by this and similar terms. This conclusion foreshadowed more recent guidance on the reporting of race and ethnicity in medical journals.6 While we have published additional papers exploring EDI since then, we suspect that the volume of papers we have published on this topic has been lower than it should have been, given the scope and severity of inequitable healthcare. For example, a recent systematic review on the impact of providing patients access to electronic health records mapped study findings to six dimensions of quality of care; the authors found that none of the 20 included studies addressed the dimension of equity.7

Nevertheless, there have been some important papers published. These include studies that reduced disparities in breast cancer screening for Arab women in Israel,8 decreased disparities in maternal morbidity between black and white women in the USA,9 demonstrated that homeless women were less likely to be hospitalised or visit an emergency department after giving birth than low-income housed women,10 explored the effect of income on medication non-adherence11 and found that voluntary reporting systems underdetect safety events in vulnerable patients.12 An important editorial,3 commenting on this last paper, also presents a framework for advancing health equity, pitfalls to avoid and recommendations for the patient safety field to advance health equity.

Going forward, we are therefore interested in publishing research that demonstrates ways to address inequitable care, drawing on such recommendations. We are also interested in studies that advance our understanding of how to identify and understand inequities in healthcare,13 that explore the perspectives of those affected, or study why inequities occur or may be worsened.14 Besides equity as the focus of research, we also encourage researchers to incorporate EDI principles more broadly in their research and quality improvement initiatives,15 such as in study team members, authors and research participants, study populations, and in involvement of suitably diverse patient and public partners to shape their work. This should help reduce the likelihood that certain groups will be excluded or unheard, or that quality initiatives will inadvertently worsen inequities.

Improving our editorial practices

We believe that greater diversity among our authors, editorialists, editors and reviewers will lead to wider perspectives being reflected in published papers, which we hope will shift the scientific narrative from traditionally privileged groups towards a more balanced understanding of how to improve quality and safety for all. However, there is growing awareness that editorial policies determining what gets published are also affected by systemic biases.16 This requires publishers and editors to explicitly consider the diversity of editorial teams and reviewers, scrutinise how decisions are made, and take actions to mitigate the problems. For example, reporting standards developed by Western researchers and unconscious bias among reviewers and editors can affect decisions about the types of papers that get published, potentially affecting what then gets implemented in practice. This may also lead to fewer papers being published from underrepresented author groups, limiting authors’ career progression and further opportunities to conduct high quality research, and thus exacerbate a cycle in which some groups of potential authors face systemic barriers.

To start our journey towards addressing these issues, we have set up an EDI working group, comprising four of our associate editors plus our editors-in-chief, to identify practices that may introduce bias and suggest ways to reduce inequities in the publishing process. We chose to have a group rather than one editor with responsibility for this area, to bring more diverse perspectives and to encourage a culture in which EDI must be considered by everyone and not just a nominated person within the organisation. The group identified several initial aims relating to potential authors, editorialists, reviewers and editors.

First, we identified an aim to support submission of potentially publishable research papers from authors from underrepresented groups. As a first step, we have run some online teaching sessions on writing quality improvement reports and research papers within geographical areas that are currently underrepresented among our published papers. We are also exploring ways of mentoring researchers from groups that typically face systemic barriers to academic progression.

A second aim relates to increasing diversity among editorialists when we commission editorials to accompany published research papers. Our first objective in this area was for gender parity among these editorialists. The editors started recording data on editorialist gender some time ago and while we initially had some male bias, we found that a more deliberate focus on identifying female editorialists helped to address this. More recently, to encourage greater diversity in all its forms, we amended the invitation letter sent to potential editorialists to include a request that they explicitly consider diversity and representation when inviting any coauthors.

Third, initial data suggest that our reviewers are less diverse in relation to ethnicity and geography than our authors. We suspect that this may be at least partly due to the demographics and networks of our editors. While we have good gender diversity among our editors, our teams are less diverse than we would like in relation to ethnicity and geography. With an aim of increasing diversity among our editors (and therefore reviewers), our most recent round of associate editor recruitment specified that we encouraged applications from underrepresented groups and geographical areas, and included this as part of our decision-making process. This resulted in a wider range of geographical representation among our editors; we now wish to actively recruit high quality editors from a much wider range of backgrounds who will introduce complementary perspectives and wider networks of potential reviewers.

Finally, we identified and committed to following a number of good editorial practices. For example, we adhere to equity policies and other guidelines of the BMJ group as a whole.17 These include encouraging our editors not to be part of male-only conference panels and enabling transgender authors and others to easily change their names on their published work with our author name change policy. We are also now asking authors of research papers to list the categories for demographic data (such as gender, race and ethnicity) alphabetically, rather than in any other order that could inadvertently imply a sense of hierarchy. To increase awareness of potential sources of bias in the publication process, we also focused on raising awareness of how our own biases may affect us as editors at our annual editors’ meeting in 2022.

Using data to identify inequities in our processes and monitor progress

As a journal focusing on research and quality improvement, we understand the importance and power of using data to drive improvement, and we aim for good quality data to underpin our commitment to improving EDI. Until recently, however, we had no information on the diversity of our authors and reviewers and were unable to identify areas of potential inequity or to monitor the effect of any interventions to address them. We were therefore delighted to be one of three journals in the BMJ group that piloted capture of data on self-reported gender and ethnicity within our manuscript handling system last year. Authors and reviewers may have noticed being asked for this information when interacting with the system. While gender and ethnicity are just two of many aspects of diversity, this is an important starting point, and we are pleased that following this pilot, data collection is now being rolled out across all journals within the BMJ group. Our hope is that this will help us to monitor aggregated data on gender and ethnicity for all submitting and accepted authors, as well as for our reviewers and editors, providing an opportunity to measure the impact of initiatives to address any disparities identified. Individual-level data are not visible to editors or reviewers, and so this information cannot inadvertently affect our editorial or peer review processes for submitted papers. BMJ Quality and Safety is also one of a group of medical journals taking part in a larger research study of gender bias in publishing, which will explore any influence of first author gender on paper acceptance.

Moving forward

We recognise that these are early days in our learning and development as a journal in relation to EDI. The steps we have taken so far are just the beginning. In the spirit of continuous improvement, we have set out objectives in each of the above domains (table 1) and will use a variety of approaches, including those based on improvement science, to achieve these.

We welcome feedback and suggestions from our readers as to how we can do more to advance this important area and as we seek to understand different perspectives in relation to diversity and inclusion. Together, we aspire to end the cycle of injustice, bias and prejudice in scientific research publication and healthcare delivery.