Racial and Ethnic Representation Reporting in Orthopaedic Research

Authors: Ahmad A. Alnasser, Tim Kobes, Marilyn Heng

Abstract

Introduction:

Orthopaedic research informs clinical practice, making it critical to collect data that reflect diverse patient demographics, including race and ethnicity. Despite the importance of these variables in addressing healthcare disparities, race and ethnicity are often underreported in orthopaedic studies. The aim of this study was to evaluate the frequency of race and ethnicity reporting in high-impact orthopaedic journals and identify trends based on study characteristics.

Methods:

Original research published between January 1, 2018, and December 31, 2020, in Clinical Orthopaedics and Related Research (CORR), Journal of Bone and Joint Surgery, and Journal of Orthopaedic Trauma (JOT) was compiled for analysis. Data on publication date, study type, level of evidence, country, and demographic variables were collected. Reporting rates were compared with the US Census 2020 diversity index and the Historical Index for Ethnic Fractionalization for global reporting, creating a performance ratio (PR) to quantify underreporting or overreporting based on geographic diversity.

Results:

Among 1452 studies, 89.2% reported sex and 86.5% reported age, but only 18.9% included race and 6.1% included ethnicity. Reporting was highest in Level II studies (29.3% race, 10.2% ethnicity) and cross-sectional studies (31.0% race, 11.2% ethnicity). United States–based studies reported race and ethnicity more frequently (28.0%), with the highest PR in Hawaii, New Hampshire, and Washington DC. Internationally, Japan, Thailand, and Finland had the highest PR. Reporting varied by journal, with CORR leading in race reporting (23.9%) and JOT in ethnicity (8.0%).

Discussion:

Race and ethnicity are markedly underreported in orthopaedic research. Addressing this disparity through standardized guidelines and enforcement mechanisms is essential to improve research quality and identify healthcare inequities. Additional exploration of these variables can better characterize diverse patient populations and inform clinical practice.

In 1993, the National Institutes of Health (NIH) mandated reporting of race and ethnicity for all NIH-funded research, dubbed the NIH Revitalization Act.^1^ As a result, journals have enacted guidelines on reporting race and ethnicity among other demographic factors. The Oxford English Dictionary defines race as “a group of people connected by common descent or origin” or “any of the (putative) major groupings of mankind, usually defined in terms of distinct physical features or shared ethnicity.”^2^ In the United States, common race categorizations are White or Caucasian, Black or African American, American Indian or Alaska Native, Asian, and Native Hawaiian or Pacific Islander.^3^ Ethnicity is defined as “membership of a group regarded as ultimately of common descent, or having a common national or cultural tradition,” and in the United States, it is commonly denoted as “of Hispanic or Latino origin” or “not of Hispanic or Latino origin.”^4^ Although race and ethnicity do not reflect inherent biological differences, their social meanings are essential for interpreting demographic data.^5^

Race and ethnicity have been correlated with disparities and outcomes in health care.^6,7^ Factors such as spatial segregation, poverty, and other social determinants of health contribute to inequities in medical care and may lead to worse outcomes.^8^ Previous studies in orthopaedics have shown that different outcomes can be related to a patient's race and demonstrated that non-White and Black orthopaedic trauma patients are less likely to receive surgery for acute injuries.^9^ Similar studies conducted in spine^7^ and arthroplasty^10^ have demonstrated that Black patients have a markedly higher risk of unfavorable outcomes after surgery compared with White patients. The International Committee of Medical Journal editors recommend improving the frequency and quality of race reporting; nevertheless, race and ethnicity remain underreported in the literature, particularly in orthopaedics.^11,12^ Orthopaedic research informs clinical practice, and thus, it is vital to collect data from a patient population that accurately represents the racial and ethnic demographics of patients who receive care.^13^

Therefore, the primary purpose of this study was to determine the rates of reporting of race and ethnicity in orthopaedic journals for all clinical studies. Our secondary purpose was to compare the locations where the studies were conducted, in both the United States and other continents, to assess the effect of race-reporting rates depending on the region and evaluate those with respect to existing diversity indexes.

Methods

For our investigation, we selected three high-impact factor journals in Clinical Orthopaedic and Related Research (CORR), Journal of Bone and Joint Surgery (JBJS), and Journal of Orthopaedic Trauma (JOT). Titles and abstracts of articles published from January 1, 2018, to December 31, 2020, were manually screened from the journals' websites, followed by full-text screening. All studies were included in the qualitative synthesis, except for cadaver studies, reviews, and nonclinical studies; these studies were excluded because of not including living patients. This study was conducted in concordance with the Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) guidelines. Institutional review/ethics board approval was not required because it does not involve living patients.

Full texts were assessed to collect and aggregate data on publication year, type of study, level of evidence, orthopaedic subspecialty, country of study, study period, age, sex, racial categories included, and ethnic categories included.

Age and sex served as frequently reported controls. We considered the reporting of demographic variables as the standard with which to compare the reporting of race and sex. We used the following race American Indian or Alaska Native, Asian, Black, or African American, Native Hawaiian or Pacific Islander, and White or Caucasian. We used Hispanic or Latino/Not Hispanic or Latino as the ethnic categories for studies conducted in the United States. Any racial or ethnic distinction in studies conducted outside the United States was grouped, given the different racial and ethnic designations for each country.

To assess how well studies accounted for race and ethnicity relative to the diversity of the populations studied, we calculated a performance ratio (PR), using already-existing diversity index data for reference. This ratio was defined as the observed proportion of studies reporting race or ethnicity divided by the expected proportion, based on the state or country's diversity index. The US Census Bureau's Diversity Index from the 2020 Census measures the probability that two randomly chosen people from a given state are from different racial or ethnic groups.^14^ For countries in the world, we used the Historical Index of Ethnic Fractionalization,^15^ which comprised ethnic fractionalization indices for each assessed country in 2013, the most recent publicly available year. Like the US Census Diversity index, the Historical Index for Ethnic Fractionalization data set also measured the probability that two randomly chosen people are from different ethnic groups. By comparing the observed reporting with the expected values derived from these indices, we evaluated whether specific regions or countries overperformed or underperformed in their reporting of race and ethnicity.

The outcomes of interest were the frequency of reporting race and ethnicity compared with the other demographic variables (age and sex), alongside location (including city, state, or country) depending on where the study was conducted. We further analyzed these findings by overlaying the performance ratios for each region onto heat maps to visualize geographic trends in reporting relative to population diversity. Demographic variables, location, and subspecialty category were collected on RedCap. Descriptive statistics, calculation of PR, and creation of subsequent plots and images were conducted using Microsoft Excel and BioRender.

Results

A total of 3,726 studies published from January 1, 2018, to December 31, 2020, across the three journals were screened for inclusion. A total of 1,452 studies were included in the final 431 from CORR, 413 studies from JOT, and 608 from JBJS (Figure 1). Of the included studies, 899 studies (61.9%) were conducted in the United States. A total of 503 studies (38.1%) were conducted outside the United States.

Reporting of Race by Geographic Location

A total of 274 studies (18.9%) reported any racial distinction and 89 studies (6.1%) reported ethnicity (Table 1). In the United States, 252 of the studies conducted (28.0%) reported race. The highest rates of reporting race were observed in Hawaii, New Hampshire, Louisiana, and New Jersey (Figure 2A). When evaluated by PR, New Hampshire, Hawaii, Washington DC, and Vermont reported race at the highest rates compared with their ethnic diversity (Figure 2B). A full breakdown of each state's reporting and PR value is given in Supplemental Table 1, http://links.lww.com/JG9/A494.

Globally, the distribution of countries reporting race was 4.0%, with the three countries evaluating the most studies being the United States, Canada, and South Korea. The countries with the highest rates of reporting race were Belgium, Thailand, the United States, and Canada (Figure 3A). When evaluated by PR, performance relative to diversity was the highest in Japan, Thailand, and Finland (Figure 3B). A full breakdown of each country's reporting and PR value is provided in Supplemental Table 2 (http://links.lww.com/JG9/A495.

Orthopaedic Journals

Racial distinctions were reported in all three journals, CORR (23.9%), JOT (17.4%), and JBJS (16.6%), in order of decreasing reporting. Ethnic distinctions were also reported in all three JOT (8.0%), CORR (6.5%), and JBJS (4.6%). Only JOT had featured studies with solely ethnic distinctions (n = 2), while the remainder of the studies with ethnic distinction were paired with racial distinction.

Level of Evidence and Study Type

When evaluated by level-of-evidence distinction, Level II evidence studies reported both race and ethnicity higher than all other study types (29.3% for race, 10.2% for ethnicity). Level I evidence studies reported race at 13.3% and ethnicity at 6.3%. Levels III and IV reported race at 22.6% and 12.0% and ethnicity at 6.8% and 3.7%, respectively. No Level V evidence studies reported race or ethnicity. When separated by study type, cross-sectional studies reported both race and ethnicity at higher rates than all other types of studies (31.0% for race, 11.2% for ethnicity). Case-control and case series designated studies had the lowest reporting rates of race (5.8%) and ethnicity (2.5%).

Discussion

The primary aim of this review was to assess how frequently race and ethnicity were reported and analyzed in orthopaedic studies published in three major journals between 2018 and 2020, in comparison with other demographic variables. Among the 1452 studies that met the inclusion criteria, data were categorized by study type, study location, level of evidence, and journal. Notably, race and ethnicity were reported far less frequently than other demographic variables, with only 18.9% of studies including race and 6.1% including ethnicity, compared with more consistently reported demographic variables such as age and sex.

Previous studies have found that only 7% to 18% of orthopaedic randomized controlled trials (RCTs) report participant race or ethnicity.^6,12,16^ In our study, we included original clinical research studies with living patients, in addition to RCTs. Cohort studies—retrospective and prospective—make up a notable portion of published orthopaedic articles, and they made up 67.1% of the studies we assessed, so including them allowed for a more representative picture of the orthopaedic literature. By level of evidence, Level II studies reported race and ethnicity most frequently. This pattern may reflect that Level I is primarily RCTs, whereas Level II spans a broader mix—RCTs, prospective comparative studies, and registries.^17^ Unlike RCTs, registries and comparative studies often collect larger data sets because they are not designed with the assumption that confounders are naturally controlled. Another factor may be the lack of collection of race and ethnicity in RCTs, which has historical and sociological bases^17^ in reluctance among researchers to engage with these populations and systemic racism as a barrier to access for patients.^18^ Furthermore, minority patients have expressed feelings of mistrust with enrolling into clinical trials, stemming from historical precedents such as the Tuskegee Syphilis Experiments.^18,19^ Moreover, 28% of US patients report hesitancy in providing race and ethnicity data, especially those who have experienced discrimination.^20,21^

In geographically assessing the difference in reporting race and ethnicity in orthopaedic research, we were able to map “hot” and “cold” zones, to assess what factors could be contributing to a lack of reporting. Our study is the first of its kind to track reporting of these factors in comparison with existing diversity indices, reporting a PR relative to an area's existing diversity level. The northeastern United States had the highest rates of race reporting, and both the Northeast and Midwest outperformed their diversity indices. Internationally, countries in North America had the highest rates, while South America, Africa, and the Middle East had the lowest, with many countries having no studies that report race or ethnicity. The difference in reporting rates for ethnicity globally might be due to some countries having more homogeneous racial and ethnic compositions, as well as an issue of access to orthopaedic journals in lower resourced areas of the world, given that some countries have not published any studies in these journals. However, even developed countries with similar or greater diversity indexes than the United States, such as Canada,^22^ had a lower percentage of reporting race and ethnicity. Likewise, generally homogeneous populations like in Japan vastly overperformed other countries, given our PR metrics. Furthermore, other countries differ in their relationships with race—researchers have previously identified a lack of data on race and ethnicity, a perceived lack of relevance compared with biological demographics, and omission of reporting because of low representation.^23^ The findings suggest multifactorial drivers of race reporting by country and state, underscoring the need for additional research to support comprehensive population reporting.

Our findings reinforce the documented racial and ethnic disparities in orthopaedic care. From physician-patient interactions in the clinic to differences in recovery outcomes, race and ethnicity play a vital role in our understanding of care. Consistent with previous literature, non-White and minority patients demonstrate worse functional outcomes in pain, length of stay, and complication rates compared with White patients,^24^ a phenomenon consistent across all orthopaedic subspecialties.^25-27^ Physicians are susceptible to implicit bias as well, as it has been shown that Black patients are less likely than their White counterparts to receive analgesics for acute pain.^28^ In an outpatient setting, Black patients were likely to be prescribed fewer at-home exercises in recovery than White patients.^29^ Given the occurrence of these disparities, reporting race and ethnicity in research can allow orthopaedics to delve into their root causes^30,31^ and our findings reveal a critical gap in that regard. Improving the inclusion and analysis of race and ethnicity allows for more actionable steps toward equitable care. Moreover, comprehensive population reporting enhances clinical judgment and facilitates applying research in practice. While race and ethnicity as genetic determinants of medical treatment are unfounded in research, their ability to provide greater insight into a group's social conditions, access, and existing disparities can aid clinicians when time with patients is limited.^32,33^ All in all, we demonstrate the lack of reporting of race and ethnicity, even in rigorous studies such as RCTs (16.0% reporting of race). Because research informs clinical practice, painting a better picture with orthopaedic research allows for better care to be performed.

Feasibility constraints may preclude racial and ethnic categorization in some studies. Most studies surveyed from the three journals were retrospective, and the lack of available data at an institutional level would preclude the inclusion of their race or ethnicity. Likewise, studies involving research solely on genetic factors of a disease may not benefit from the inclusion of race and ethnicity, especially if the genetic factor is well established as not varying by race or ethnicity. Studies that are conducted in regions with highly homogeneous populations might lack racial and ethnic diversity, thus not deeming it necessary to report race and ethnicity as variables for analysis.

Interestingly, even after 30 years following the NIH Revitalization Act, low participation in trials and low rates of reporting persist, even among NIH-funded studies.^34,35^ Our review highlighted notable differences in reporting across the three journals. CORR had the highest percentage of studies reporting race (23.9%), followed by JOT (17.4%) and JBJS (16.3%). Notably, while JBJS did not have guidelines for inclusion of race and ethnicity reporting at the evaluated period for this review, as of January 1, 2024, it requires race and ethnicity reporting for submission.^36^ The submission guidelines for CORR encourage, but do not mandate, race and ethnicity reporting. JOT lacks specific guidelines to reporting. Thus, greater editorial emphasis, potentially including explicit publication requirements, would allow more studies to present complete demographic profiles in their results. After JBJS's decision, journals should track reporting rates closely.

Publishing criteria should require reporting race and ethnicity and encourage analyzing these variables alongside age and sex because such practices enrich study conclusions and promote inclusion of diverse patients in future research.^9,37^ For example, the Journal of the American Medical Association noted that if race or ethnicity is reported, the Methods section should include an explanation of who identified participant race, as well as the reason they were assessed.^32^ Similarly, Circulation has specified that authors provide subanalyses by sex or race/ethnicity when appropriate or indicate that there was no difference between subgroups.^38^ The proposed steps provide actionable, concrete guidance for reporting race and ethnicity in research articles, and adopting them as standards incentivizes authors to capture these variables during data collection.

Critiques of race/ethnicity mandates emphasize that definitions vary across medical contexts and that race and ethnicity represent distinct, multidimensional constructs that should not be collapsed into a single analytic variable.^5,39^ While these statements are correct and important considerations, race and ethnicity paint an important picture of how the data were collected in a given study. Although their classification does vary under different contexts, including these demographic variables can guide our assessment of treatment modalities and evaluation of research across specialties and thus a step in the right direction to improve how we characterize patient populations. Because we do not have any holistic models or variables yet to illustrate the complexity of race and ethnicity,^40^ clinicians and researchers ought to use the information currently available to allow the progression of knowledge toward these models. Given the methodology introduced within this study, journals can use the US Census every 10 years as a benchmark to assess their own PR with the publicly available diversity index, to ensure that research is keeping up with a racial and ethnic landscape set to change further, because the United States has a non-White majority in populations younger than 18.^13^

In conclusion, we present the first geographic assessment of race and ethnicity reporting in orthopaedics, incorporating non-RCT designs from the literature. Furthermore, we introduce the usage of the PR in quantifying overreporting and underreporting of race in the orthopaedic literature. Our study demonstrated that participant race and ethnicity were underreported in published orthopaedic research in three journals and were further underreported when assessed by location of study, level of evidence, and type of study and when adjusting for an area's baseline diversity. Race and ethnicity should be reported more often in an effort to make orthopaedic research more holistic and to potentially narrow race-based and ethnicity-based disparities that exist in orthopaedic research. Potential future directions involve analyzing a larger number of journals, different demographic variables related to socioeconomic condition, or multiple specialty-specific journals to compare reporting rates between orthopaedic subspecialties.

Supplementary Material