Methodology for promoting equity-informed research in sport and exercise medicine: recommendations from the AMSSM Collaborative Research Network

Conducting equitable, diverse and inclusive SEM research along the research continuum

1. Forming the research team

Assembling a diverse research team is critical for representing diverse perspectives in research; however, diversity alone will not ensure equity and inclusion within the team.14 Involving collaborators with different backgrounds and attributes (eg, age, gender, race/ethnicity, geographic location, disability status and/or sexual orientation) from the outset will allow the research team to draw on a variety of lived experiences to conduct more equitable and inclusive research ideally incorporating perspectives of the study’s target population. Diverse teams strengthen trust and increase participation rates when research is conducted with under-represented populations.15

An inclusive research team also actively incorporates individuals who represent the target population to ensure appropriate communication, participation and relevance to the community.16 17 These stakeholders may take an active role in the research process by identifying important areas of study, suggesting interventions and building trust between the population being studied and the research team.18–20 A commonly used phrase coming from an American Indian community describes this concept best as ‘no research about us, without us’.19 A study by Pedersen et al
17 discussed how the research team was developed with an academic-community partnership to understand physical activity behaviours of rural American Indian older adults. Community-based participatory research (CBPR), an inclusive research method, involves community members in all steps of the research process to create equitable partnerships between researchers and the community.18 19

To ensure equity on the research team, all members should have their perspectives heard and valued. During research team formation, study leaders should guide the team in reflexivity practices (eg, discussion of implicit bias, structural competency and positionality) to understand power dynamics and proper management strategies.14 21 Consistent dialogue and reflection should occur throughout the research process. Study teams should be aware that inclusion, without ongoing critical reflection of power dynamics within the study team, may inadvertently perpetuate existing inequities and harm on traditionally under-represented team members. Superficial diversity, such as adding a junior researcher from an under-represented background without valuing their input, will not lead to high-quality equity-informed or equity-focused research. Instead, senior investigators should seek opportunities for mentorship and sponsorship in coordination with these team members.14 The voices of all research team members must have equal weight and those in the typical majority must be open to changing their top-down approach and long-held practices.14 Similarly, allies within the research team need to speak up if the perspectives of under-represented groups are being ignored or dismissed.14 The burden of ensuring equity and inclusion within a research team should fall on those with more power than those with less and should be considered throughout the research process.

Recommendation 1: create an equitable, diverse and inclusive research team, ensuring the meaningful inclusion of community members representing the population of interest, with the goal of coproducing research, improving the quality of research and elevating the experience of all team members.22

1. Consider equity-related, theoretical frameworks

Prior to finalising a research question, teams should consider the problem through the lens of an intersectional, multilevel, theoretical framework (eg, equity-focused and/or social-focused). Theoretical frameworks are often considered in tandem with lived experience (including clinical expertise in the case of SEM clinicians) and the existing scientific literature to create a conceptual framework.23 Equity-related frameworks involve a logical connection of concepts and premises developed from existing theories (eg, feminist, intersectionality, queer, disability justice).24 In qualitative and quantitative study designs, conceptual frameworks can include narrative descriptions or visual displays (eg, directed acyclic graphs) to demonstrate how important concepts are related and inform a research question. While the appropriate theoretical framework will vary based on the research question, several robust, multilevel frameworks are widely used and can be instructive in developing research questions, selecting measures and finalising analytic plans. The WHO’s Social Determinants of Health (SDOH) Research Framework demonstrates how non-medical factors influence health outcomes and lead to inequities, that is, “the conditions in which people are born, grow, work, live, and age, and the wider set of forces and systems shaping the conditions of daily life”.25 26 This framework focuses on five domains of SDOH: (1) economic stability, (2) education access and quality, (3) healthcare access and quality, (4) neighbourhood and built environment, and (5) social and community context.27 In figure 2, we have included an adapted figure of the WHO’s SDOH framework with SEM-specific examples that align with each domain.

Figure 2

SDOH framework graphic.

Another commonly used equity-focused framework is the NIMHD Research Framework which specifies domains of influence on health outcomes (individual behaviour, physical environments, sociocultural environments and the healthcare system) and levels of influence (individual, interpersonal, community/organisational, and societal).28 In figure 3, we adapted the NIMHD framework with specific examples to illustrate applications to SEM. Both the WHO and NIMHD frameworks challenge researchers to situate their research in a contextualised understanding of health disparities, where individual differences are rooted in exposures that precede individual differences in behaviour. Other theories situate individuals in their context, such as ecological systems theory29 and ecosocial theory,30 which also may be useful to consider for certain research questions.

Figure 3

Research framework chart.

Framework-driven research is particularly important for ensuring health disparities research is not perpetuating unfounded and potentially stigmatising explanations for differences in health outcomes.31 Boyd and colleagues caution that race is often incorrectly identified as a risk factor in research involving health disparities,32 rather than specifically examining the pathways through which racism and other forms of oppression are impacting the outcomes of interest.32 They emphasise that research measuring racial differences should specifically use the word racism, and note the type of racism (eg, systemic, interpersonal, internalised) that is theorised. Ideally, such potential pathways are tested analytically. For example, rather than solely describing differences in subspecialty concussion care by ethnicity, Copley et al
33 demonstrated that these differences were in part driven by English proficiency and insurance status. Knowing that differences between two racial or ethnic groups exist due to modifiable experiences in a healthcare setting because of differences in access to healthcare services provides important insights for future action. Researchers should also acknowledge there are multiple potential forms of oppression that may impact health outcomes by considering intersectional frameworks that incorporate lenses of discrimination based on race, sex, sexual orientation, gender identity, disability and other characteristics of historically marginalised populations.34

Recommendation 2: identify the theoretical framework informing your research question, hypotheses, study design and interpretation of results. To advance equity-informed and equity-focused SEM research, consider using a multilevel, intersectional framework.35

1. Choosing a study design

When planning equity-informed or equity-focused research, it is important to design studies that avoid unintentional bias towards the study population who may be at higher risk of marginalisation and the impact of health disparities. Importantly, inclusivity in study design involves intentional reflection on traditional and Western ways of knowing (eg, authority, rationalism, scientific method) and embracing other ways of knowing (eg, indigenous ways of knowing) when deciding on the best approach to answering a research question.36–38 As previously discussed, involving members of the researched community throughout the research process is critical as they provide invaluable input to all aspects of research, including study design. Incorporating CBPR should be considered for any study design, from observational studies to randomised control trials (RCTs).

Observational research (cohort, case-control and cross-sectional studies) is common in SEM.39 40 These designs allow researchers to evaluate exposures and outcomes that occur within specific populations or at certain timepoints. Cross-sectional study designs assess exposures and outcomes at the same time and may be a good option for a relatively quick study that seeks to gather baseline data for a future, larger study.40 41 However, not all data can be quantified, and qualitative data are uniquely capable of providing contextualisation for lived experiences of study participants as they relate to the research question, particularly in equity-focused research or when little previous research exists on a specific topic.

Qualitative and mixed-methods approaches in health disparities research could reveal contextual information and participant-centred views and perspectives that can be used to promote equity and reduce disparities in SEM.42 43 Additionally, qualitative methodological approaches are well equipped to reveal insights from participants with intersectional identities. For example, a qualitative study from Couch et al
44 used semistructured interviews with Olympic female African American sprinters to better understand how race, gender and athletic identity impacted transitions during their competitive careers. To provide additional guidance, health equity researchers recently developed analytic guidance and considerations for conducting equity-informed qualitative analysis.45 Mixed-methods research involves using both qualitative and quantitative study designs either simultaneously or sequentially.41 A study by Eberman et al
46 provides a helpful example of a mixed-methods study using a cross-sectional survey and follow-up semistructured interviews to investigate athletic trainers’ knowledge and experiences regarding the healthcare needs of transgender student-athletes.

RCTs, by design, lack generalisability, as researchers may exclude important populations (eg, excluding non-English speakers from a study given that one of the survey tools is only validated in English). These exclusions may explain—in part—why these populations have been historically understudied.47 Additionally, randomisation is often not feasible or ethical in health disparities research.41 Observational, qualitative or mixed-methods study designs may be more appropriate and reveal important findings to inform future research priorities.

Recommendation 3: when appropriate, use qualitative or mixed-methods (both quantitative and qualitative) study designs to answer research questions. Ensure potential inherent biases are taken into consideration and addressed in the research team’s choice of study design.

1. Selecting study measures

Not only is the use of valid and reliable measures necessary to appropriately answer research questions in quantitative studies,48 it is also important for the comparability of research findings across studies. This helps to support future meta-analyses and systematic reviews. The National Institutes of Health has created a PhenX Toolkit, which may be useful for SEM researchers.49 This web-based resource contains a repository of reliable and valid common data elements related to social determinants of health, divided by level of exposure (eg, individual, structural). The current sports medicine literature includes the use of many non-validated survey items and patient reported outcome measures, or those which are validated in specific populations—often adult, able-bodied men. This can lead to inequitable health outcomes in certain marginalised groups within SEM, including people with disabilities, youth, people with obesity and pregnant people.9

The selection of specific measures depends on the research group’s underlying theoretical framework and research question. For example, if neighbourhood characteristics are the outcome of interest, this could be measured at different levels. The Area Deprivation Index (ADI) is a well-established, US census-based measure that gives a synthesised view of socioeconomic disadvantage within census block groups. It has previously been used to demonstrate the relationship between neighbourhood context and health outcomes, for example, higher readmission rates among Medicare patients living in the most disadvantaged neighbourhoods.50 51 Within the SEM context, the ADI was used to demonstrate that black athletes with sudden cardiac arrest come from neighbourhoods with greater socioeconomic deprivation than white athletes or athletes of other race with sudden cardiac arrest.52 These findings suggest that racialised disparities in SDOH, via structural racism in the USA, may influence cardiovascular risk in young athletes. Another composite measure of area-level deprivation is the Social Deprivation Index developed based on seven demographic characteristics collected in the American Community Survey and used to quantify levels of socioeconomic disadvantage and their association with health outcomes.53 Other examples are tools that evaluate the impact of racism at multiple levels and in different ways. This may include measuring individual experiences of racism or racial trauma with the use of a validated self-report scale54 55 or using a community-level proxy for systemic racism such as racial inequities in policing or unemployment.

Alternatively, in a qualitative context, theoretical frameworks and existing theories are useful to guide the development of interview or discussion guides. For example, Pedersen et al
17 used an ecological model of health behaviour to investigate individual, social/cultural, and environmental influences on physical activity behaviour among rural American Indian older adults, a group prone to health disparities.

Recommendation 4: to perform high-quality equity-informed SEM research, select reliable and valid study measures based on the research question(s), study design and underlying research framework. Quantitative measures and/or qualitative inquiry of socioeconomic deprivation or discrimination should be considered to explore social determinants and health inequities.

1. Recruiting and retaining participants

Recruitment and retention of participants are essential parts of many research studies. The inclusion of a diverse study population is necessary to ensure an accurate and comprehensive assessment of the research question. Equitable, diverse and inclusive recruitment and retention strategies must begin from the inception of a research study. At this stage, research teams should once again engage in reflexivity practices to evaluate unconscious biases in participant recruitment and retention practices, to approach this process with a more inclusive lens.

An essential component of successful recruitment and retention is building trust and connection with the target community from the beginning of the research process and regularly examining this relationship throughout the study.56–59 Common practices in participatory research approaches (eg, Action Research, Community-Based Participatory Research, Participatory Action Research) prioritise genuine and meaningful participation among stakeholders.60 This can include empowering and training community members to be part of the research team and assist with participant recruitment. Recruitment of diverse populations can be challenging due to the legacy of unethical research involving under-represented communities, including abuse and deception as well as withholding treatment for research purposes.61 This has resulted in mistrust, apprehension and hesitancy to participate in research.56 Research teams must actively work to address barriers related to research mistrust by engaging directly with the community, including community members as participatory and active members of the research team, and facilitating open communication and collaboration. This collaboration must include opportunities for feedback and an open-mindedness to modify the study design, if indicated, based on input from the community.57 62 Inclusive practices to promote retention of diverse participants could include accommodations for caregiving responsibilities, evening and weekend data collection, and addressing accessibility of study materials for people with disabilities.

Recruitment approaches should be multimodal62 and inclusive of both proactive and reactive strategies.56 Proactive strategies include direct contact with potential study participants in order to share study information and to request their participation. This may include face-to-face interactions with community members at events such as community centres, fitness facilities, sports practices, interscholastic events and street fairs.56 63 These interactions provide opportunities for community members to engage in dialogue, discuss concerns for potential harm, address cultural values, express expectations, ask questions, and build connection and trust with the research team.56

Reactive strategies involve indirect contact with potential participants. This may include indirect contact between the research team and participants through collaboration with key community leaders, snowball or word of mouth, printed and broadcasted material, or referrals from healthcare providers.56 Encouragement from trusted community members, family and friends helps build confidence and ease apprehension and distrust.56 59 64 65 Mindful development of recruitment materials is imperative to ensure they are culturally sensitive, available in different languages and strategically placed in a variety of locations to help reach a diverse population.56 59 63 Broadcast recruitment strategies should focus on locations with high accessibility to ethnically diverse populations to capture a wide audience of participants.

Recommendation 5: to recruit and retain diverse study participants, investigators should cultivate a diverse and inclusive study team, carefully consider the study design, build trust with the community, have them involved in the research from the initiation of the study, and use multimodal recruitment strategies.

1. Collecting data

For researchers to identify causes of health disparities, the collection of representative and accurate data is paramount. Without appropriate measurement of factors influencing health outcomes, researchers cannot examine why those outcomes occur, or hypothesise ways to intervene.66 Many current standards of care are rooted in norms established from data of white, English-speaking, non-disabled, cis-gendered men.66 While there has been recent dialogue on how to improve EDI in research, data instruments have not kept pace with societal and demographic changes in the USA and globally.67 In quantitative studies, selection biases, information biases and measurement errors can lead to biased results and conclusions.68 Determining ways to avoid and minimise these potential biases prior to data collection is critical.

To improve the breadth and depth of demographic data collected, SEM researchers should consider collecting data on the following characteristics: race, ethnicity, preferred/primary language, gender, disability or functional status, geographic location, sexual orientation and gender identity, among others.69 Moreover, considering the intersection of these characteristics may reveal important associations and outcomes in equity-focused research. Qualitative methods are well equipped to examine the intersectional dimensions of human experience as mentioned previously in the study by Couch et al.44 The SDOH are another important component of data collection. Examples of SDOH relevant to SEM are safe housing, socioeconomic deprivation, access to protective sporting equipment, access to transportation, job opportunities, exposure to pollution, and literacy skills (figure 2).

There is often a lack of standardisation in data collection methods, which can lead to missing data, making it difficult to combine and compare datasets across studies. Guidance on the collection of data on race and ethnicity has been developed and acknowledges that researchers should use inclusive language, provide comprehensive categories and subcategories for participants, and report on how these data were collected.70 The reflexivity of the researcher should be considered and discussed to examine how one’s own lived experiences and assumptions may influence the data collection and analytic processes. Appropriate data collection methods in both quantitative and qualitative studies should be prioritised, and scientific journal editors and reviewers should hold authors to these standards. Validated assessment and screening tools of social drivers of health should be identified and disseminated to researchers so that best practices are known.71–73

Addressing structural barriers to research participation in data collection is also essential. Special consideration should be given to ensure that data collection methods offer opportunities for flexible assessment times, such as evenings and weekends, and at convenient sites to reduce potential barriers to transportation, meals and/or childcare.56 58 63 Factors related to universal design and access for study participants with physical, sensory, communication or cognitive disabilities should also be considered.74 Appropriate compensation that covers expenses related to travel, loss of pay from work, childcare and participant time should be provided to promote participation.56 59 63 Ongoing communication, and proactively seeking feedback from participants, is essential to minimise participant burden and address barriers to participation as they arise.

Recommendation 6: ensure accurate and representative data are collected by minimising potential barriers to data collection; consider the unique needs of study participants and reduce barriers to their participation.

1. Developing and executing a data analysis plan

Data analysis plans are a critically important aspect of the research journey. Authors Light, Singer and Willet astutely stated, “you cannot fix by analysis what you bungled by design”.75 In short, any data analysis is only as strong as the thought put into the study design. A data analysis plan should be developed during the initial study design phase and in close alignment with the study aims and hypotheses, target study population, interventions and outcomes. Choices made during this stage will impact what outcomes can be evaluated from the collected data.

A data analysis plan should include analyses that evaluate both the impact of a specific participant characteristic/factor on the outcome of interest and (when feasible) any intersections that may lead to inequities. Each individual characteristic may have a different association with the outcome when paired with other characteristics/factors. A SEM-focused example of using an intersectional lens in data analysis is the cross-sectional survey of parent-reported youth sports participation by Hyde et al
76 which examined differences in participation by demographic factors, finding that household income level had a pronounced impact on participation among younger age groups.

Understanding the associations of all intersecting characteristics with the outcome is often not practical; however, ensuring diversity of enrolled participants may allow testing meaningful interactions of subgroups (ie, the intersections of individual characteristics). Importantly, the absence of evidence that subgroup outcomes differ should not be equated to evidence of an absence of differences. Relatively novel statistical approaches to the field of SEM, such as Bayesian hierarchal methods, could be considered to assess the probabilities of meaningful differences between subgroups when sample sizes are small; however, these methods often require strong statistical knowledge. We strongly recommend partnering with statisticians/epidemiologists/data scientists/qualitative researchers early in the research process to ensure proactive and intentional use of analyses that support equity-informed and equity-focused methods. Analyses of qualitative or mixed-methods data should involve an expert in these methodologies. Importantly, qualitative analysis involves researchers transcribing interviews and focus groups, coding the data and mapping the data to develop rich interpretations. Given its iterative nature, sufficient time should be allocated for this part of the research process and consideration of participatory data analysis approaches, which involve community members, should be considered for analysis and interpretation.77 78

When developing a data analysis plan and reporting results, investigators should also consider disaggregating the data by various characteristics or factors when possible, while also preserving the anonymity of potentially vulnerable study participants. Oftentimes various levels of a data variable are grouped (due to small numbers) to conduct more meaningful subgroup analyses. However, this grouping makes a strong assumption that those subgroups have similar outcomes. The importance of these characteristics and their varying levels to the research question should be considered during the design phase so that sufficient data are collected to conduct meaningful analyses.

Recommendation 7: data analysis and reporting plans should be developed a priori, with attention to methods that emphasise equity. This is best done in collaboration with experts (eg, biostatisticians, epidemiologists or qualitative researchers) to select the most appropriate quantitative and/or qualitative approaches and identify opportunities to assess intersectionality and disaggregate the data when possible.

1. Publishing and disseminating study results

The availability, awareness and application of equity-informed SEM research findings are dependent upon publication and dissemination of study results to a broad audience. However, as with many aspects of research, researchers must be aware of the potential for bias that limits publishing and dissemination. Gender equity is lacking in scientific journal senior authorship for women.79–82 Although racism remains a key determinant of health and a wide body of research exists detailing the relationship between race and health, a relatively small number of published papers address racialised inequities in the USA.83 84 Despite the increase in publications focused on identifying health disparities within specific populations, there remains a paucity of publications documenting trends towards improving health equity for marginalised groups.85 SEM journals should consider the recent work by the British Journal of Sports Medicine (BJSM) to provide EDI guidance to authors and consider incorporating similar principles into their own publication guidelines.86 87

Current avenues for publishing equity-informed SEM research present their own challenges. In recent years, the open-access publication movement, which strives to make research more widely accessible, has favoured senior male researchers at prestigious academic institutions88 and researchers in higher resourced environments.89 As a result, women and authors under-represented in medicine publish less frequently, and those with little or no financial means (eg, researchers in low-income countries) may not have the same opportunity to publish in open-access journals, representing a clear example of structural bias and lack of equity and diversity within publishing houses. The National Institutes of Health (NIH) has acknowledged the role structural racism plays in the scientific community broadly, while offering useful resources towards supporting EDI within SEM and medical research in general.90 Parallel work by the NIH has also acknowledged the impact for structural ableism in scientific research.91

The last area to address regarding health equity in SEM research involves the public dissemination of research results. Given how important and challenging it can be to effectively disseminate research into communities of colour, people with disabilities, non-Anglophone individuals and other marginalised groups, alternative methods of research dissemination are important. Research teams should have conversations with communities of interest to help identify the most appropriate mechanisms for dissemination. Digitally networked technologies and the internet can be extremely valuable;92 however, consideration of additional modes of communication is equally important, including potential translation of outcomes into other languages. Over the previous decade, nearly all research journals, scientific magazines and scholarly books have become available online. In addition, through digital platforms created within social media, blogs, wikis and podcasts, health research can be communicated to new audiences formerly out of reach with traditional communication methods.93 However, dissemination through digital platforms, while clearly advantageous, also creates potential barriers for those without access to these platforms (ie, households without internet access) and those for which online platforms serve as impediments to learning. Providing multiple avenues for dissemination of SEM research, beyond digital platforms (eg, hosting community events, holding discussions with policy makers with community member involvement, interacting with local press and media), is needed to create new opportunities for diverse investigator teams and interdisciplinary collaborations to communicate their results to diverse audiences94 and to ensure accessibility of the research to the communities it is meant to support.

Recommendation 8: to enhance equity in research dissemination, SEM researchers should consider equity-informed dissemination strategies during the planning stages of their research to ensure these principles are implemented when publishing and disseminating results. Considerations should include, for example, free layperson’s summaries, low-cost or free open-access publication, and actively campaigning for a more equitable SEM publication model.