Reporting of Sociodemographic Data in Cochlear Implant Clinical Trials: A Systematic Review

Abstract

OBJECTIVE:

The purpose of this study was to systematically evaluate the literature on the frequency of reporting of sociodemographic data (gender, race, ethnicity, education status, health insurance status, geographic location of residence, and socioeconomic status) among interventional clinical trials involving cochlear implant patients.

DATABASES REVIEWED:

A systematic search was performed in PubMed, Cochrane Database of Systematic Reviews, Web of Science, and SCOPUS, to identify peer reviewed research.

METHODS:

A systematic review was performed which included original prospective clinical trial research studies involving cochlear implantation and/or interventional trials involving cochlear implant patients. Collected data included funding type, level of evidence, race reporting, ethnicity reporting, socioeconomic status reporting, education level reporting, type of insurance, geographic location, and gender of patients.

RESULTS:

A total of 644 articles were included for review. Gender was the most reported sociodemographic factor (70% of included studies). Reporting of other data among included studies was low: educational level (6%), socioeconomic status (2%), race (1%), ethnicity (1%), insurance status (0.3%), and geography (1%). The odds of reporting gender (OR 1.51), education (OR 1.81), geography (OR 2.72) increased with each subsequent publication date decade; however, this trend was not seen for reporting of race, ethnicity, socioeconomic status, or insurance. The reporting of gender was less likely to be reported in studies with the pediatric participants (OR 0.62), level II evidence (OR 0.14), and device programming interventional studies (OR 0.26).

CONCLUSION:

Reporting of sociodemographic data, other than gender, is low among prospective clinical trials involving cochlear implant patients. The lack of reporting of this key data may limit research rigor and generalizability. Clinical researchers are advised to prospectively collect this data to promote equity in cochlear implant research and clinical care.

INTRODUCTION

Cochlear implants have provided immense benefit for populations on a global scale over the past 4 decades.¹ As of 2019, nearly 750,000 people worldwide have received an implant.² Unfortunately, disparities in hearing healthcare access and utilization are persistent and pervasive.³ A wide range of racial and geographic disparities persist in hearing healthcare. In order to provide the greatest public health benefit, it is necessary to evaluate the effectiveness and clinical outcomes of cochlear implant recipients across diverse populations. Equitable recruiting and comprehensive reporting of sociodemographic data are essential elements to analyzing and interpreting results from clinical trials. Despite equity-promotion efforts by public funding agencies, regulatory agencies, and medical journals, there is evidence of inequitable representation among hearing related clinical trials.⁴ Collecting sociodemographic data in clinical trials can promote external validity of research and can be used to promote equity in hearing healthcare. A deeper understanding of socioeconomically associated variations in clinical outcomes between subgroups can improve overall healthcare access and quality.

Domains such as economic stability, physical environment, education, social and community context, and healthcare access and quality are commonly known as social determinants of health and are primary drivers of health outcomes and intimately connected with the sociodemographic data of cochlear implant patients.⁵ Addressing social determinants of health is the way to obtaining health equity. Within the cochlear implantation field, several sociodemographic factors have been linked to health outcomes across the domains of social determinants of health. Non-white infants and infants of urban low-income neighborhoods have a higher incidence of failed newborn hearing screening and confirmed hearing loss.^3,6 Non-white children are more likely to be delayed in cochlear implantation when compared with white children.⁷ Children from rural regions also have a higher incidence of hearing loss⁸ and face significant delays in access and utilization of hearing healthcare, including cochlear implantation.⁹ Similar patterns are seen in the adult population, as non-white patients are less likely to receive a cochlear implant and may face significant delays in cochlear implantation.^10,11 Rural adults are more likely to be delayed in cochlear implantation when compared to urban adults.¹² Additional studies have identified many different risk factors associated with greater odds of hearing loss such as low socioeconomic status, ethnicity and educational attainment.¹³ Overall, these studies demonstrate that non-biological factors may influence health and healthcare outcomes and further attention is needed to investigate the sociodemographic factors that impact cochlear implant recipients. The objective of this research is to systematically evaluate the literature on the frequency of reporting of sociodemographic data (race, ethnicity, gender, location of residence, socioeconomic status, education, and insurance type) in interventional cochlear implant studies. We hypothesize that 1) very few prospective clinical trials involving cochlear implants report sociodemographic data and 2) reporting of this data will be associated with the level of evidence, funding status and timing of publication.

METHODS

As this study involves deidentified data from published publicly available research, this study was exempt from Institutional Review Board approval. The Preferred Reporting Items for Systematic Reviews and Meta-Analysis checklist was used to guide the reporting of the methods of this systematic review.¹⁴ A specific outcome measure was not required for inclusion in this review. The specific inclusion criteria included (1) published in English, (2) published after 1980, (3) prospective in design, (4) clinical trial involving cochlear implantation and/or interventional trial involving cochlear implant patients, (5) reporting of clinical outcome, and (6) involvement of either children, adults, or both. Exclusion criteria included (1) retrospective observational studies, (2) non-original research, and (3) animal studies. The PICOTS statement for this review includes the following: (P) cochlear implant recipients of any age, (I) clinical trial involving a device, audiological/rehabilitative, or behavioral intervention, (C) including intervention and control comparison groups, (O) outcome of reporting of sociodemographic data, (T) no specified timeframe, (S) involving prospective study designs.

Search Strategy

To perform a systematic literature review of sociodemographic reporting in cochlear implantation studies, a search string was designed to include studies that addressed both (1) cochlear implantation and (2) prospective interventional trials. A search strategy was developed using The National Library of Medicine’s (NLM) Medical Subject Heading (MeSH) browser in expanded concept view to identify MeSH indexed search terms. To capture articles pertaining to the study objective, MeSH terms associated with “cochlea” were used to ensure that pertinent cochlear implant studies were not excluded. To identify all relevant articles meeting study design criteria, several study types were included such as: clinical trial, controlled trial, cohort studies, longitudinal studies, and prospective studies. An initial search was performed in PubMed. The title abstract field code was selected to ensure that articles that mentioned the pertinent terms in any form would be captured. The search string was then translated for Web of Science, Scopus, and Cochrane Library (Wiley); with duplicates from the searches being removed using EndNote (Clarivate, Philadelphia, USA). The search was performed in March 2021. Figure 1 lists terms utilized in the search string as well as the algorithm for inclusion/exclusion. The PRISMA Search Extension (https://www.prisma-statement.org//Extensions/Searching) was utilized for this study and the database details are described in Supplemental Appendix A.

FIG. 1.

Preferred reporting items for systematic review and meta-analysis algorithm for the reporting of demographics in cochlear implant trials.

Data Extraction and Quality Assessment

Article titles and abstracts were reviewed independently by five reviewers and were selected or removed based on the inclusion and exclusion criteria. In the event of disagreement over inclusion, the article was included for full-text review to be more inclusive. Three reviewers completed a full-text review of all eligible articles independently. These reviewers collected data on each article included for full text review. The information obtained included (1) study design, (2) number of participants, (3) patient population: adult, pediatric, or both, (4) funding type, (5) trial registration, (6) reporting standard, (7) type of hearing intervention, (8) intervention and control group details, and (9) sociodemographic reporting detail: race, ethnicity, socioeconomic status, education, insurance, geography, and gender. Education included but was not limited to the level of education for the parents of pediatric patients, patient IQ status, patient reading level or comprehension, and level of education obtained by the patient. Geography inclusion pertained to any information detailing the zip code, city, county, or rurality of a given patient population. The level of evidence of each article was also assessed according to the Oxford Centre for Evidence-based Medicine guidelines.¹⁵ Due to the nature of this review to evaluate reporting of sociodemographic data among clinical trials with no planned critical analysis of study methodology or meta-analysis of outcomes; we determined a priori not to formally evaluate study risk of bias/methodologic quality.

Factors Associated with Reporting

We evaluated the association between frequency of reporting of sociodemographic factors (race, ethnicity, gender, geographic location of residence (i.e. – urban versus rural), socioeconomic status, education, and insurance type) and a number of study characteristics such as: decade of publication (1980–1989, 1990–1999, 2000–2009, 2010–2019, and 2020–2021), age group of participants (adult, pediatric, or both), funding status/type, study type, trial registration (such as clinicaltrials.gov), level of evidence, and type of intervention within the trial (CI device, other hearing related device, device programming, rehabilitative strategy). Univariate and multivariate logistic regression analysis was performed to generate odds ratios of sociodemographic data reporting along with 95% confidence intervals. Each of the sociodemographic factors were analyzed as a dependent variable in the analyses and the independent variables included: decade of publication, age group of participants, funding status, funding type, trial registration, level of evidence, and type of trial intervention. Results were considered statistically significant if the resulting P-value was equal to or less than 0.05. Statistical analysis was performed using STATA 12.1 (College Station, TX, USA).

RESULTS

Search Results

The initial search of all three databases yielded 22,032 articles. After duplicates were removed, 15,664 articles remained. The titles and abstracts of these articles were scanned to determine if they met the study objectives; 14,352 articles were removed through this process, leaving 1,312 articles included for full text review. 644 of these articles (Supplemental Appendix B) met the inclusion criteria and were eligible for the systematic review. The reporting of demographics by decade, study type, participant age, and funding type is recorded in Table 1.

Table 1.

Results from systematic review

			Demographic Reporting
		Number of Articles, N	Gender Reported, N (%)	Race Reported, N (%)	Ethnicity Reported, N (%)	Socioeconomic Status, N (%)	Education, N (%)	Insurance, N (%)	Geography, N (%)	1 or more Reported, N (%)	2 or more Reported, N (%)
Decade	1981–1989	16	12 (75%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	12 (75%)	0 (0%)
	1990–1999	65	27 (42%)	0 (0%)	0 (0%)	1 (2%)	2 (3%)	0 (0%)	0 (0%)	29 (45%)	1 (2%)
	2000–2009	147	82 (56%)	0 (0%)	1 (1%)	3 (2%)	7 (5%)	1 (1%)	1 (1%)	86 (59%)	7 (5%)
	2010–2019	329	254 (77%)	7 (2%)	3 (1%)	7 (2%)	20 (6%)	0 (0%)	4 (1%)	258 (78%)	22 (7%)
	2020–2021	87	74 (85%)	0 (0%)	0 (0%)	3 (3%)	8 (9%)	1 (1%)	3 (3%)	75 (86%)	9 (10%)
Study Type	Prospective Cohort Study	582	405 (70%)	7 (1%)	4 (1%)	14 (2%)	34 (6%)	1 (0%)	6 (1%)	416 (71%)	34 (6%)
	Randomized Clinical Trial	48	35 (73%)	0 (0%)	0 (0%)	0 (0%)	3 (6%)	0 (0%)	2 (4%)	35 (73%)	4 (8%)
	Other	14	9 (64%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)	1 (7%)	0 (0%)	9 (64%)	1 (7%)
Participant Age	Adult	359	263 (73%)	2 (1%)	2 (1%)	3 (1%)	10 (3%)	1 (0%)	2 (1%)	265 (74%)	13 (4%)
	Pediatric	211	138 (65%)	4 (2%)	2 (1%)	10 (5%)	26 (12%)	1 (0%)	5 (2%)	147 (70%)	24 (11%)
	Both	74	48 (65%)	1 (1%)	0 (0%)	1 (1%)	1 (1%)	0 (0%)	1 (1%)	48 (65%)	2 (3%)
Funding Type	Foundation/ non-profit	48	37 (77%)	2 (4%)	2 (4%)	1 (2%)	4 (8%)	0 (0%)	0 (0%)	38 (79%)	3 (6%)
	Government	53	37 (70%)	3 (6%)	1 (2%)	4 (8%)	7 (13%)	1 (2%)	1 (2%)	38 (72%)	9 (17%)
	Industry	103	81 (79%)	2 (2%)	1 (1%)	1 (1%)	5 (5%)	0 (0%)	2 (2%)	81 (79%)	7 (7%)
	No funding or not listed	440	294 (67%)	0 (0%)	0 (0%)	8 (2%)	21 (5%)	1 (0%)	5 (1%)	303 (69%)	20 (5%)
	Total	644	449 (70%)	7 (1%)	4 (1%)	14 (2%)	37 (6%)	2 (0.3%)	8 (1%)	460 (71%)	39 (6%)

Study Characteristics

Of the 644 articles included in this systematic review, 48 (7%) were randomized clinical trials (level of evidence 1b) and 596 (93%) were prospective cohort studies (level of evidence 2b). The studies were conducted from 1981 to 2021, and all were reported in English. The study participants were adults in 358 articles (56%), pediatric in 211 articles (33%), and both in 75 (12%) articles. Regarding funding status, 231 articles (36%) reported funding and 413 (64%) were either unfunded or had no funding listed. Of the articles reporting funding, 103 (16%) reported industry funding, 84 (13%) were government funded, and 58 (9%) were funded by a foundation or non-profit. Seven percent of articles recorded trial registration while less than one percent noted the reporting standard used.

Study Demographic Reporting

Seventy-one percent of the included articles included at least one demographic variable and six percent included two or more demographic variables. Of the seven demographic variables collected in this systematic review, gender was reported most at seventy percent. The next highest variable reported was education at six percent followed by socioeconomic status which was reported in two percent of the included articles. Race, ethnicity, and geography were all reported in one percent of the articles, while insurance status was recorded in only two of the 644 articles. After 1980–1989, there was a trend toward increased reporting of gender with subsequent decades with 77–85% of articles published between 2010 and 2021 reporting gender (Figure 2). Based on logistic regression, the odds of reporting gender were 1.81 times higher with each subsequent decade (p=0.001, 95% CI 1.49–2.19). While the number of articles has increased as well with each subsequent decade, the odds ratios regarding demographic data were calculated using percentages. This ensured the data set was controlled for total number of articles, focusing on highlighting the trends. The reporting of education level was overall low with only 6% of all included articles reporting this data. While less frequently reported, a trend was seen in the reporting of participant/parent education level over time (Figure 3). Based on logistic regression, the odds of reporting education were 1.51 times higher with each subsequent decade (p=0.05, 95% CI 1.00–2.27). The reporting of geography was overall low with only 1% of all included articles reporting this data (Figure 3). The odds of reporting geography were 2.72 times higher with each subsequent decade (p=0.05, 95% CI 1.00–2.27). Reporting of race, ethnicity, socioeconomic status, and insurance was overall low and ranged between 1–2% of all included articles and there was no statistically significant evidence of increased reporting over the past 4 decades (Figure 3).

FIG. 2.

Reporting of gender by decade.

FIG. 3.

Reporting of race, ethnicity, socioeconomic status, education, insurance and geography by decade.

Using multivariate analysis logistic regression, we evaluated the odds of reporting sociodemographic data within these included studies based on a variety of factors. The reporting of race was associated with the type of intervention studied in these included trials, with studies evaluating hearing aid/other device interventions more likely to report race of participants (OR 50.3, p=0.025, 95% CI 1.64–1540.42). Similarly, studies evaluating hearing aid/other device interventions were more likely to report ethnicity of participants (OR 95.77, p=0.022, 95% CI 1.92–4780.05). The reporting of socioeconomic status was low (14 of 644 studies, 2%) and was not associated with any of the study characteristics on multivariate analysis. The reporting of education status was associated with age of participants, with studies evaluating pediatric participants more likely to report education status than studies with adult participants or both adult and pediatric participants (OR 4.71, p=0.001, 95% CI 1.89–11.75). The reporting of insurance was low (2 of 644 studies, 0.3%) and was not associated with any of the study characteristics on multivariate analysis. The reporting of geography was low (8 of 644 studies, 1.2%) and was associated with the type of intervention studied in these included trials, with studies evaluating cochlear implant devices 98% less likely to report geographic residence of participants (OR 0.018, p=0.019, 95% CI 0.001–0.523). The reporting of gender was associated with the age of participants, level of evidence, and type of intervention studied. Studies involving pediatric patients were 38% less likely to report gender (OR 0.62, p=0.034, 95% CI 0.393–0.964). Compared with level 1 evidence, studies of level 2 evidence were 86% less likely to report gender (OR 0.14, p=0.026, 95% CI 0.026–0.794). Studies evaluating device programming strategies in a clinical trial were 74% less likely to report gender (OR 0.26, p=0.001, 95% CI 0.117–0.570).

Different sources of study funding also demonstrated variations in demographic reporting. Studies funded by industry had the highest percentage of gender reporting (79%) followed by foundations or non-profits (77%), government (70%), and lastly studies without funding or that did not list funding (67%). Government-funded studies had the highest percentage of race reported at six percent compared to four percent of studies funded by foundations/non-profits and two percent of industry-funded studies. Government-funded studies also had the highest percentage of reporting socioeconomic status (8%) and education (13%). Studies funded by foundations/non-profits had the second highest percentages for these demographic variables with 2% reporting socioeconomic status and 8% reporting education. As seen in Figure 4, seventeen percent of studies with government funding reported two or more demographic variables compared seven percent of industry-funded studies, six percent of studies funded by foundation/non-profit, and five percent without funding.

FIG. 4.

Percentage of articles by funding type reporting at least one or two demographic variables.

DISCUSSION

The reporting of sociodemographic data has become standard practice in conducting clinical trials. Trial registries, such as ClinicalTrials.gov, began requiring the reporting of race and ethnicity statistics during submission of trials, if this information was collected, recently in 2017.¹⁶ This requirement is a specific example of methods to promote diversity in clinical research and to facilitate increased rigor and greater external validity. Furthermore, these sociodemographic factors are connected to the social determinants of health which influence health outcomes. These demographics are important to include as they influence how we interpret research findings, mobilize resources to increase access to care, use the information to impact clinical care, adapt interventions to increase the effectiveness and implementation, and develop new clinical questions for subgroups. This systematic review demonstrates the overall low reporting among cochlear implant trials of important sociodemographic factors that influence health and healthcare outcomes.

Of the 644 cochlear implant clinical trials included in this systematic review, less than 1% of the studies reported race and/or ethnicity. It is difficult to comment on equity in cochlear implantation when essential racial and ethnic data is not collected in most cochlear implant trials. Low reporting of race and ethnicity has been demonstrated in another systematic review evaluating the reporting of race/ethnicity and sex in 125 hearing-related trials conducted in the US.⁴ In contrast, this systematic review includes studies conducted outside of the US, reports on additional sociodemographic variables, and focuses on cochlear implant therapy as an intervention rather than hearing loss as a general diagnosis. Lack of diversity in clinical trials has also been noted in the Orthopedic Surgery, Vascular Surgery, and Pulmonary Medicine fields.^17–19 It is essential to pursue robust reporting and equitable recruitment in cochlear implant trials as there is evidence of significant disparities in access to and utilization of cochlear implantation healthcare based on race and ethnicity. Among adults, patients from non-white racial groups are less likely to pursue cochlear implantation when they are eligible candidates.¹⁰ Among pediatric populations, black and Hispanic children are delayed in cochlear implantation and less likely to receive a CI compared to white children.⁷ The factors influencing these findings are poorly understood; however, unconscious bias and racial/ethnic discrimination contribute to disparities such as these.²⁰

Going beyond the findings of limited racial/ethnicity reporting, this study also demonstrated limited reporting of all the other key sociodemographic factors within clinical trials involving CI patients. The lack of reporting of these data represents a significant limitation of clinical trials within this field and these factors influence access to and utilization of healthcare. Promotion of equity in cochlear implantation is dependent on accurate and exhaustive definition and description of inequities in access or care delivery/quality and how those inequities relate to the social determinants of health (SDOH).²¹ Socioeconomic status and insurance status are directly connected to the SDOH and these elements were seldomly reported in the included studies in this review. These data are important for cochlear implant research teams as insurance status influences cochlear implant care for many patients.²² Public insurance coverage in some states may prevent patients from receiving sequential bilateral cochlear implants compared with those who are privately insured.²³ Patients of lower socioeconomic status or with public insurance coverage are also significantly more likely to experience CI surgery complications and less likely to adhere to post-operative appointments.^23,24 The geographic residence of patients was reported in 1% of the studies and the lack of this information is also significant as geography influences cochlear implant outcomes. There are pervasive limitations in the services and the hearing healthcare workforce in rural regions, which is problematic as hearing loss may disproportionately affect rural populations.^25–26 When compared to urban populations, rural children and adults face pervasive delays in diagnosis and treatment of hearing loss (including cochlear implant care).^{9,12,,27–29} Reaching vulnerable patient populations (such as rural residents) is complex but efforts are underway through a variety of pragmatic interventional trials to improve access and timing of hearing healthcare.^30–32 These types of interventions have the potential to improve the access to and timing of cochlear implantation. Educational status of participants or parents was reported in only 6% of studies included in this review. This sociodemographic factor influences health in a variety of ways; however, there is evidence that adult speech perception outcomes following cochlear implantation are influenced by educational attainment levels.³³ Parental educational levels are associated with delays in utilization of hearing healthcare and with speech development outcomes in pediatric CI recipients.³⁴ These disparities, based on sociodemographic data, provide further justification as to why collecting and studying this data is important with cochlear implantation.

There are limitations to this review. Considering the purpose and design of this study, we are unable to comment on the methodology of each included trial or the results/outcomes of each study. The primary objective of this study was not focused on an objective clinical outcome to be evaluated across studies; thus, a meta-analysis was not conducted. However, we were able to conduct multivariate analysis on sociodemographic data reporting and various independent variables related to article/publication factors. The overall reporting of sociodemographic factors was very low across all studies, and this may account for variability in the data with wide confidence intervals for significant associations on multivariate analysis. The historical lack of standardized trial registries, such as ClinicalTrials.gov (which was made public in February 2000), may have influenced the limited reporting among clinical trials; however, only 81 of the 644 included articles were published before 2000.

Based on the impact of key factors of the social determinants of health, consistent and comprehensive reporting of sociodemographic data in cochlear implant research represents an important step forward to promote research rigor and equity. The National Institutes of Health has taken steps to promote equity in research recruitment and reporting through a variety of initiatives.^35–37 While there are systematic obstacles to recruit diverse patient populations in research, each cochlear implant program should strive to connect with communities that are underrepresented in clinical care and clinical research.³⁸ Study sponsors and journal editorial staff should consider the value of this data and should promote the inclusion of this data in clinical research. There should be a continued effort on the part of researchers in interventional studies to include the reporting of sociodemographic data.

CONCLUSIONS

Reporting of sociodemographic data, other than gender, is low among prospective clinical trials involving cochlear implant patients. There is evidence of increased reporting data over the past 4 decades; however, sociodemographic data other than gender is reported in less than 6% of all cochlear implant clinical trials. The lack of reporting of this key data may limit research rigor, generalizability, and health equity. Clinical researchers are advised to prospectively collect this data to promote equity in cochlear implant research and clinical care.

Supplementary Material

Supplemental Digital Content: Appendix A

Database Coverage Details

Supplemental Digital Content: Appendix B

References of Included Studies in the Systematic Review