Discussion
Our findings corroborate those of other ophthalmological CT reviews, which have shown the majority of patients to be white and from high-income countries, with under-representation of racial and ethnic minorities.14–16 To address this longstanding problem, new FDA and NIH initiatives, for example, Bridge2AI and AIM AHEAD, are driving more diverse and representative datasets in medical research.17 18
In ophthalmology, it is not usual to validate and reproduce CTs in distinct countries and populations. However, there are numerous examples of treatments in other medical areas that have been found to be beneficial in studies conducted in one country, but without benefit or even harmful in another country. For example, the Surviving Sepsis Guidelines for paediatric intensive care include rapid and early fluid resuscitation based on US-based studies; nevertheless, in a CT with a sub-Saharan African population, this gold standard treatment increased the risk of death, raising concern about implementing guidelines without local validation.19
More diverse cohorts and studies are needed to determine the exact relationship between diseases and treatment, especially in minority groups, where the gap in medical care access is greater.5 Inclusion of sex description is one of the NIH recommendations in clinical research, and it is reported in 99.5% of the participants in NIH-funded ophthalmologic CTs, with an acceptable gender distribution among enrolled patients.
The NIH requires the reporting of the breakdown of the race–ethnicity of trial participants, but it is reported in half of the ophthalmologic CTs in the US and in 20% of non-US CTs, although with improvement in recent years. The lack of standards in the reporting of race–ethnicity is also an issue. While the NIH has recommended racial categories, various studies used different racial classifications. For example, some trials had distinct classifications for Hispanic participants, and some had distinct categories for Asian participants.
A review from Allison et al reported a disproportionately low representation of minority participations in glaucoma CTs,20 a similar finding to what we identified across all ophthalmologic specialties. Our study shows that in ophthalmological CTs, white patients are predominant, with lower representativeness of blacks, Asians and other groups. This raises a concern in the applicability of the findings to patients who are not white and from low-and-middle-income countries (LMIC).
In ophthalmological CTs, only 69 countries out of 195 total countries in the world are represented. Most of the ophthalmological CTs are from populations of high-income and upper-middle-income countries, with only 0.25% from low-income countries where the burden of disease is actually highest. Several barriers to better representation of marginalised communities in ophthalmological CTs have been outlined previously. With substantially lower rates of medical insurance and access to care, patients from racial/ethnic minorities are less likely to present to health systems and consequently less likely to be recruited to trials.21 Even when access to insurance and healthcare is available, trial participation requires extraneous costs such as additional transport and additional/prolonged clinical visits, which curb participation.22 English proficiency is usually required for trial participation. Mistrust in physicians and research institutions further compound difficulties in recruiting underserved populations.22
Although the reasons for (and consequences of) poor recruitment and representation have been well documented, to the authors’ knowledge little has been done by ophthalmological research bodies to prioritise more representative inclusion. A recent global review published in the Lancet highlights this, demonstrating that of 66 recent diabetic retinopathy RCTs, none were performed in LMICs.23 A review of trials analysing how to improve access to cataract services showed only two RCTs were from LMICs.24 And despite the disproportionate prevalence of glaucoma in black populations, major trials examining the effect of medications on glaucoma have not included non-white racial groups altogether.25 Overall, almost three-quarters of published ophthalmological studies were undertaken in HICs—several regions in Southeast Asia, sub-Saharan Africa and Latin America had especially low representation.26
Strategies to improve representation in ophthalmological trials have been outlined elsewhere. They stress the importance of educating researchers in HICs on recruiting those from marginalised populations into trials.15 16 Collaboration with racial and ethnic minority clinicians and clinics that primarily serve minority communities, and establishing research facilities in LMICs, means research can be performed by those caring for those populations.15 16 Engaging members of marginalised communities as stakeholders in trials is crucial in fostering trust in the medical research community.15 27
The collection of demographic information in CTs is improving, with almost all studies including gender description and an increasing number with race and ethnicity descriptions. However, updated classifications to better represent groups are necessary, such as more detailed information among Hispanic and Asian groups.
Race and ethnicity distribution varies among countries, and reporting in CTs is fundamental to verify the generalisability and ensure external validity of findings.
There is little incentive to perform CTs, and research in general, in resource-limited countries, perpetuating the global knowledge gap, which in turn fuels the divide between rich and poor countries. Advocacy from within the research community, governments and non-government players such as the WHO, is necessary to increase representativeness and improve applicability of medical research to those who are most burdened by disease.
