Low Rates of Reporting and Analyzing Race and Ethnicity in Hand Surgery Randomized Controlled Trials: A Systematic Review

Purpose: Sociodemographic factors have been shown to influence musculoskeletal health. However, little is known regarding the frequency of reporting and analysis of certain sociodemographic variables (e.g., age, sex, height, weight, body mass index (BMI), race, and ethnicity) in randomized clinical trials (RCTs) pertaining to hand surgery. The purpose of this study was to assess the rate of reporting and analysis of these variables in RCTs published in the Journal of Hand Surgery (JHS). Methods: A systematic review was conducted of RCTs published in JHS between 2015 and 2021. For each study, we determined whether the following sociodemographic variables were reported and/or analyzed: age, sex, height, weight, BMI, race, and ethnicity. Frequencies were reported by year and as a cumulative total. Studies were evaluated using the Cochrane risk-of-bias tool. Results: A total of 45 RCTs met inclusion criteria, with about half (53.3%) originating from the United States. Age (97.8%) and sex (91.1%) were the most frequently reported sociodemographic variables, followed by race (17.8%), BMI (11.1%), and ethnicity (8.9%). Age (17.8%) was the most frequently analyzed variable, followed by sex (13.3%), and race (4.4%); the remaining variables were not analyzed in any study. Conclusions: While age and sex are reported at a high rate, only about 1 in 4 RCTs published in JHS report either race or ethnicity. All sociodemographic variables were infrequently included as part of statistical analysis. The significance of these findings should be recognized and used to interpret and enhance the methodology of future RCTs.


Introduction
The influence of race and ethnicity on healthcare disparity is well-recognized within the medical community and has become a growing focus within the orthopedic literature [1][2][3][4][5][6][7] . The influence of race and ethnicity on postoperative outcomes following joint and spine surgery has been particularly elucidated. Adelani et al. retrospectively reviewed postoperative complications in 585,269 patients who underwent hip and knee arthroplasty. Within this study, Black patients experienced increased rates of surgical site infection (SSI), deep vein thrombosis (DVT), pulmonary embolism (PE), myocardial infarction, stroke, and death, even when controlling for medical comorbidities 5 . In a separate retrospective review of 4,803 patients, Sanford et al. found that Native American race was found to be an independent risk factor for SSI following cervical fusion and decompression laminectomy, whereas African American race was found to be an independent risk factor for SSI and PE after decompression laminectomy and DVT after lumbar fusion 2 . Alosh et al. screened over 100 million hospital discharge records from 1992 and 2005 and found 965,600 anterior cervical spine procedures. The authors similarly found that minorities had lower rates of surgery and that Black patients had significantly higher odds of dying while in the hospital 6 . Racial and ethnic differences in outcomes, decision-making, and other aspects in hand surgery itself further illustrate the health disparities within the field [8][9][10][11][12][13] .
Despite racial differences in health outcomes within the orthopedic literature, many randomized controlled trials (RCTs) fail to report race and ethnicity 14,15 . Several reasons for this observation have been postulated, such as the belief that reporting these factors is not clinically relevant and a lack of emphasis to report by medical journals 15 . However, according to the Consolidated Standards of Reporting Trials (CONSORT) guidelines for transparent reporting of clinical trials, all sociodemographic information should be provided in the initial description of a study population 16 . Similarly, the National Institutes of Health (NIH) guidelines require that minority patients be included in NIH-funded research and suggest that race and ethnicity be collected even in cases where previous research has demonstrated no effect of these variables on the outcomes of an intervention 15 .
It is important to identify racial and ethnic differences within orthopedic studies so that further analyses may elucidate the underlying causes of differential health outcomes. While some studies have assessed the reporting and analysis of sociodemographic variables across orthopedic subspecialty areas, none have focused on hand surgery specifically. The Journal of Hand Surgery (American Volume) (JHS) is a premier journal in this field with the largest number of RCTs on PubMed search relative to similar journals. For this reason, the purpose of our study was to assess the rate of reporting and analysis of sociodemographic variables (e.g., age, sex, height, weight, body mass index (BMI), race, and ethnicity) in RCTs published in JHS from 2015 to 2021. We hypothesize that age and sex will be reported and analyzed at the highest frequency compared to the other sociodemographic variables.

Search Strategy
The Preferred Items for Systematic Reviews and Meta-Analysis (PRISMA) guidelines were used to conduct this systematic review 17 (Supplementary Table 1). This review was not prospectively registered. An advanced search on PubMed was conducted to identify all RCTs published in JHS from 2015 to 2021. Search terms included "randomized control trial" and "randomized controlled trial." Search results were screened to confirm use of an RCT design.

Inclusion and Exclusion Criteria
All RCTs published in JHS between 2015 and 2021 on human subjects were included. We chose this span of years to reflect continuity with previous orthopedic studies on this topic, which included data from 2015 to their respective dates of publication 14,15 . Longitudinal analysis and previous follow-ups of clinical trials published prior to 2015 were included. Exclusion criteria were non-RCTs, basic science studies, meeting abstracts, responses to authors, letters to the editor, and withdrawn studies.

Data Collection
Eligible studies were assessed independently by two reviewers to determine whether the following sociodemographic variables were reported and/or analyzed: age, sex, height, weight, BMI, race, and ethnicity. Any discrepancy was resolved by consensus agreement with a senior author.
Data collection was based on the methodology reported by Griffin et al. 14 . A variable was considered reported if the mean or median with or without standard deviation/ quartiles was provided for continuous variables (age, height, weight, BMI) or if a percentage of the study population was provided for categorical variables (sex, race, ethnicity). Comparing baseline demographics between treatment groups or between treatment and control groups was considered reporting but not analysis. A statement that all patients were of one race or ethnicity was considered adequate for reporting.
A variable was considered analyzed if statistical analysis was performed on the variable relative to the study's outcomes of interest. The evaluation of outcomes based on sociodemographic subgroups was considered analysis. If a variable was found not to be analyzed throughout the included papers, this indicated that no figure or table was included with the variable as part of a sub-analysis, no mention of an analysis with respect to the variable was found throughout the Methods or other sections within the study, nor were any conclusions drawn regarding the sociodemographic variable's impact on analyzed outcomes.
Race was defined using the following categories: White, African-American/Black, Asian/Pacific Islander, Native American, or other/unknown race. Ethnicity, which is defined as a subset of race, was defined as Hispanic or non-Hispanic 18 . Sex and gender were assessed as the same variable for the purpose of this study, as distinction between these terms is often interchangeable within the literature 19 . Each study's country of origin was also recorded, with the institution of the senior author considered the country of origin when authors from multiple countries contributed to a study.

Risk-of-Bias Assessment
The revised Cochrane risk-of-bias tool for randomized trials (RoB 2.0) was used to assess each included study as "high risk," "low risk," or with "some concerns," of bias. This tool evaluates the methodology of each study by scoring the following categories: randomization process, deviations from intended interventions, missing outcome data, measurement of the outcome, selection of the reported result, and overall bias 20 . Studies were assessed based on adherence to intention-to-treat analysis as this statistical method allows for optimal comparison between treatment groups and minimizes sources of bias 21 .

Search Results
A total of 10,380 studies published in JHS were initially screened. 179 RCTs were identified, of which 45 met inclusion criteria and were included in the final review ( Figure 1). Studies originating from 15 countries were included, the most common of which were the United States (53.3%), Denmark (6.7%), and Sweden (6.7%) ( Figure 2). Most RCTs included in this review were published in 2017 (26.7%) and fewest were published in 2016 (4.4%) ( Figure  3). Supplementary Table 2 compiles all included studies and illustrates data collected from each article.

Reporting of Sociodemographic Variables
Of the 45 included studies, 44 reported age (97.8%), 41 reported sex (91.1%), eight reported race (17.8%), five reported BMI (11.1%), four reported ethnicity (8.9%), one reported height (2.2%), and one reported weight (2.2%) ( Figure 4).   Sociodemographic reporting trended upward with time ( Figure 5). Age and sex were reported at a high rate for each year of the study period, with both variables being reported in all studies since 2019. The reporting of race was variable year to year but trended upward with time, from 11.1% in 2015 to 25% in 2021. Ethnicity was only reported in 2017 (16.7%), 2020 (16.7%), and most frequently in 2021 (25%), and no specific trends could be observed. The proportions of sociodemographic reporting by year can be found in Supplementary Table 3.

Analysis of Sociodemographic Variables
The most analyzed sociodemographic variable was age (n=8; 17.8%), followed by sex (n=6; 13.3%) and race (n=2; 4.4%) ( Figure 6). No studies analyzed weight, height, BMI, or ethnicity. No specific trends could be observed for the analysis of the variables when separated by year of publication ( Figure 7). The proportions of sociodemographic analysis by year can be found in Supplementary Table 4. Figure data is also summarized in Supplementary Table 5 to allow for accurate, additional interpretation of the provided figures.

Risk of Bias
"Selection of the reported result" had the least bias, with 95.6% of studies in this category classified as low risk ( Figure 8). Most studies (>84%) had low levels of bias with respect to "deviations from intended outcomes," "missing outcome data," and their "randomization process." "Measurement of the outcome" revealed some concerns for bias in 31.1% of studies.

Discussion
The primary purpose of this study was to evaluate the frequency of reporting and analyzing of several sociodemographic variables in hand surgery RCTs published in JHS between 2015 and 2021. Age and sex were the most reported demographics and were presented in nearly every study. Interestingly, race was reported in about one of six studies, while ethnicity was reported half as often. Height and weight were most infrequently reported but included through BMI in more than 10% of studies. Despite high rates of reporting, age was included within the statistical analysis of only 17.8% of studies, followed by sex in 13.3%, and race in 4.4%. There was no analysis of ethnicity, height, weight, or BMI throughout the seven-year period.
A recent systematic review investigated the rates at which randomized controlled trials published in 10 orthopedic journals between 2015 and 2019 reported and analyzed these same sociodemographic variables 15 . Of 482 total articles, only 7.3% reported race and 3.1% reported ethnicity. Analysis by race (1.2%) and ethnicity (0.2%) were much less frequent. Within this same study, articles were further subdivided into subspecialty categories. Of the 12 articles pertaining to hand surgery, 8.3% reported race and no studies reported ethnicity; each of these studies failed to perform analysis of these demographics. Interestingly, we found that RCTs published in JHS over a similar period reported race (17.8%) and ethnicity (8.9%) at much higher frequencies. These findings imply that rates of sociodemographic reporting may vary amongst RCTs of different orthopedic journals even if they pertain to the same orthopedic domain. Further study of specialtyspecific journals should be performed to increase the power of these results and more accurately identify the rate of reporting and analysis of sociodemographic variables 14 .
In reviewing the racial and ethnic demographics of included publications within our study, we found there to be great diversity but inconsistent reporting of included patients [22][23][24][25][26][27][28][29] . For example, of the eight RCTs that reported race, half provided data on the number of White versus non-White patients 22,[24][25][26] only, while half provided data on a larger variety of subgroups including White, Black, Asian, Native American, and Pacific Islander patients 23,[27][28][29] .
Only three studies reported ethnicity by including the percentage of Hispanic patients within their study populations 23,28,29 , while one study commented on the 'ethnic homogeneity' of its study population 30 . Regarding analysis of race and ethnicity, only two of the 45 studies (4.4%) within this review analyzed outcomes based on race, and none analyzed outcomes based on ethnicity.
As discussed previously, race has been shown to impact joint and spine surgery in the field of orthopaedic surgery 2,5,6 . However, other studies have explored the effect of race and ethnicity on the outcomes, decision-making, and other aspects of hand surgery. In a retrospective review of 92,921 patients with carpal tunnel syndrome, Brodeur et al. found that Black and Asian patients were less likely to undergo surgery compared to White patients. Similarly, the authors showed that patients of Hispanic ethnicity had decreased odds of surgery compared to patients of non-Hispanic ethnicity 31 . Mahmoudi et al. retrospectively reviewed 13,129 patients with traumatic digit amputation and found that Black patients were less likely than White patients to undergo replantation procedures 32 . In a separate analysis of over 31,000 trigger finger patients by Brodeur et al., Asian, African American, and other minority patients were less likely to undergo surgery relative to White patients 9 . Squitieri et al. similarly showed that Black and Hispanic children underwent attempted reimplantation of an amputated finger at significantly lower rates than their White counterparts, even after controlling for potential confounding factors 11 . Following a brachial plexus injury, Bucknor et al. found that Black patients are more likely to be treated in the emergency department as opposed to an elective, outpatient setting and are also less likely to receive supported discharge compared to White patients 12 . Walsh et al. found that Black and Hispanic patients show worse functional outcomes and report higher levels of pain following a distal radius fracture relative to White patients 10 while a review by Khetpal et al. revealed many outcomes affected by various sociodemographic variables, including race 13 .
These studies illustrate the disparities in treatment rate and outcomes, among other factors, that are associated with the race and ethnicity of the patient. By highlighting these findings, we hope to emphasize the presence of these health inequities and stress the importance of analyzingor at least reporting-these sociodemographic variables for future RCTs. Because race and ethnicity have been shown to affect patient decision-making and outcomes and may affect access to healthcare or the biases that patients experience, study outcomes should take into account these potential moderating factors. Not only can this reveal other yet-undiscovered health inequities, but it can also lead to treatments optimized for the patient's race and prevent complications disproportionately affecting particular minorities.
While race was infrequently reported and analyzed even less often, age was the most frequently analyzed sociodemographic variable in our review. Many of the included studies show that age, like race, can impact outcomes, decision-making, and other aspects in hand surgery. In a mixed-methods study by Zhuang et al., participants were asked to choose between receiving carpal tunnel release (expensive) or orthosis wear (inexpensive) for hypothetical carpal tunnel symptoms after either receiving or not receiving cost information regarding the procedure. After stratifying the participants based on age, the younger subgroup was more inclined to choose surgery despite exposure to cost information when compared to the older subgroup 23 . Valdes et al. investigated whether there was a difference in postoperative outcomes following volar plate fixation for distal radius fractures in patients randomized to home (unsupervised) vs. therapist-supervised hand therapy. Overall, there were no statistically significant differences in self-evaluation scores, extremity motion, pain, or grip strength. However, older subjects had poorer grip and self-evaluation scores at 12 weeks and reported less pain when compared to younger subjects 33 . Finally, Chung et al. studied the predictors of outcomes 12 months following distal radius fractures. The authors found that increasing age was associated with lower Michigan Hand Questionnaire scores, implying that older patients with distal radius fractures may expect poorer outcomes when compared to those of younger age 27 .
These findings show that outcomes of hand surgery RCTs can be heavily influenced by age. Differences based on other sociodemographic variables may be elucidated with increased reporting and analysis. Researchers should focus on the identification of health disparity so that clinicians can more effectively counsel patients before and after treatment.
The conclusions of our study may be limited by the small sample size of RCTs that met the inclusion criteria. Though all RCTs between 2015 and 2021 published in JHS were included, additional years of data collection or inclusion of other study types -such as highly powered cohort studies 15 -may help to reveal racial and ethnic differences. Similarly, only papers published in JHS were included in this study; future studies should explore these sociodemographic trends in other hand surgery journals. As JHS is based in the United States, there may be some bias in the rate of publishing of studies submitted from the journal's home country, potentially skewing the geographical distribution of recent hand literature. Studies published in non-US-based hand surgery journals can be evaluated in future studies to explore this possible bias. In addition, there are a variety of socioeconomic factors associated with patient race and ethnicity that may influence study outcomes and never be adequately considered 34,35 . As such, the effect of biological versus social factors on health outcomes remains difficult. Finally, included studies may have forgone the evaluation of race or ethnicity if these differences were known to have negligible effect on the intervention. Future reviews can select studies associated with topics known to exhibit notable health differences amongst these variables.

Conclusions
The present review of RCTs published in JHS between 2015 and 2021 found that the sociodemographic variables of age and sex were reported at high rates. Conversely, race was less commonly reported, and ethnicity was rarely reported. Each variable was infrequently included as part of statistical analysis. Because outcomes of hand surgery RCTs can be heavily influenced by race and age and may be further moderated by other sociodemographic variables, both reporting and analysis of these variables is crucial to ensure accurate and comprehensive study conclusions. The significance of these findings should be recognized and used to interpret and enhance the methodology of future RCTs.

Conflict of Interest
The Authors declare that there is no conflict of interest.

Funding
This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors. Specify the methods used to decide whether a study met the inclusion criteria of the review, including how many reviewers screened each record and each report retrieved, whether they worked independently, and if applicable, details of automation tools used in the process.

94-121
Data collection process 9 Specify the methods used to collect data from reports, including how many reviewers collected data from each report, whether they worked independently, any processes for obtaining or confirming data from study investigators, and if applicable, details of automation tools used in the process.

94-121
Data items 10a List and define all outcomes for which data were sought. Specify whether all results that were compatible with each outcome domain in each study were sought (e.g. for all measures, time points, analyses), and if not, the methods used to decide which results to collect.

94-121
10b List and define all other variables for which data were sought (e.g. participant and intervention characteristics, funding sources). Describe any assumptions made about any missing or unclear information.

99-121
Study risk of bias assessment 11 Specify the methods used to assess risk of bias in the included studies, including details of the tool(s) used, how many reviewers assessed each study and whether they worked independently, and if applicable, details of automation tools used in the process.

125-131
Effect measures 12 Specify for each outcome the effect measure(s) (e.g. risk ratio, mean difference) used in the synthesis or presentation of results.

13a
Describe the processes used to decide which studies were eligible for each synthesis (e.g. tabulating the study intervention characteristics and comparing against the planned groups for each synthesis (item #5)).

94-121
13b Describe any methods required to prepare the data for presentation or synthesis, such as handling of missing summary statistics, or data conversions.

Study selection
16a Describe the results of the search and selection process, from the number of records identified in the search to the number of studies included in the review, ideally using a flow diagram.

137-138, 297
16b Cite studies that might appear to meet the inclusion criteria, but which were excluded, and explain why they were excluded. For all outcomes, present, for each study: (a) summary statistics for each group (where appropriate) and (b) an effect estimate and its precision (e.g. confidence/credible interval), ideally using structured tables or plots.

Results of syntheses
20a For each synthesis, briefly summarise the characteristics and risk of bias among contributing studies.
137-165, 308 20b Present results of all statistical syntheses conducted. If meta-analysis was done, present for each the summary estimate and its precision (e.g. confidence/credible interval) and measures of statistical heterogeneity. If comparing groups, describe the direction of the effect.  Provide registration information for the review, including register name and registration number, or state that the review was not registered. 78 24b Indicate where the review protocol can be accessed, or state that a protocol was not prepared. 78 24c Describe and explain any amendments to information provided at registration or in the protocol. 78, N/A Support 25 Describe sources of financial or non-financial support for the review, and the role of the funders or sponsors in the review.

293-294
Competing interests 26 Declare any competing interests of review authors. 292 Availability of data, code and other materials 27 Report which of the following are publicly available and where they can be found: template data collection forms; data extracted from included studies; data used for all analyses; analytic code; any other materials used in the review.