Diabetes is a chronic metabolic condition that affects more than 537 million adults worldwide, with numbers projected to rise sharply over the next two decades. Clinical research remains the cornerstone of evidence-based management, driving innovations in pharmacotherapy, lifestyle interventions, and preventative care. However, the outcomes of such research are often influenced by race and ethnicity — complex social constructs that intersect with genetics, socioeconomic status, access to healthcare, and cultural behaviors. To develop truly equitable treatment strategies, it is imperative to understand how these factors shape research findings and ultimately patient outcomes.

The Role of Race and Ethnicity in Diabetes Research

Race and ethnicity are not biological determinants but rather proxies for a range of environmental, genetic, and social variables that can affect disease susceptibility, progression, and response to treatment. In diabetes research, failing to account for these influences can lead to biased results, missed therapeutic opportunities, and widening health disparities. Recognizing and systematically incorporating racial and ethnic factors into study design, analysis, and interpretation is essential for producing generalizable and equitable evidence.

Genetic and Biological Factors

While race is not a clean genetic category, certain ancestral populations carry genetic variants that influence diabetes pathophysiology. For example, individuals of African descent have a higher prevalence of insulin resistance and lower rates of beta-cell function, which may contribute to an earlier onset of type 2 diabetes. East Asian populations often exhibit a distinct phenotype characterized by lower body mass index but greater visceral adiposity and impaired insulin secretion. Certain ethnic groups also harbor protective variants, such as the AMY1 gene copy number variations in some Native American populations that correlate with reduced risk. A landmark study published in Nature Medicine identified over 600 genetic loci linked to type 2 diabetes, many with differential effects across ancestries, underscoring the need for ancestry-informed approaches in clinical trials.

Sociocultural and Economic Determinants

Socioeconomic status, health literacy, dietary habits, physical activity patterns, and cultural beliefs about illness and medication profoundly affect diabetes management. For instance, Hispanic/Latino communities in the United States often face language barriers and limited access to diabetes education, while African American populations experience higher rates of food insecurity and neighborhood deprivation. These social determinants interact with genetic susceptibilities to produce stark disparities in glycemic control, complication rates, and mortality. A report from the CDC shows that age-adjusted diabetes mortality rates are more than twice as high for non-Hispanic Black adults compared to their White counterparts, a gap that cannot be explained by biology alone.

The Historical Context of Underrepresentation

Historically, clinical trials in diabetes have overwhelmingly enrolled White participants, particularly those of European ancestry. This lack of diversity has two major consequences: first, it limits the external validity of research findings, and second, it perpetuates a cycle of inequity where treatments proven effective in homogeneous populations may be less effective or even unsafe in minority groups. The infamous case of the sulfonylurea drug troglitazone (Rezulin) highlighted how pharmacogenomic differences — in that case, a higher rate of hepatotoxicity in women of African and Hispanic descent — can be missed when trials lack diversity. Regulatory bodies have since responded with updated guidance. The FDA’s 2020 draft guidance mandates diversity plans for sponsors, emphasizing the need for race, ethnicity, and sex subgroup analyses in pivotal trials.

The Current State of Representation

Despite these efforts, progress remains slow. An analysis of diabetes clinical trials registered on ClinicalTrials.gov between 2000 and 2020 found that Black or African American participants accounted for only 8–12% of enrollment, while Hispanic or Latino participants comprised 5–8%, far below their proportional disease burden. Moreover, nearly 70% of trials did not report race or ethnicity data at all. This statistical invisibility perpetuates blind spots in efficacy and safety profiles. For example, the cardiovascular outcomes trials for SGLT2 inhibitors and GLP-1 receptor agonists have shown consistent benefits across racial groups, but subgroup sizes are often too small to detect meaningful differences in adverse events or adherence barriers.

Impact on Clinical Research Outcomes and Health Disparities

When research does not adequately capture the diversity of affected populations, the resulting evidence base can inadvertently reinforce disparities. This occurs through several mechanisms:

Biased Risk Stratification

Risk prediction algorithms for diabetes complications — such as the ASCVD risk estimator or the UK Prospective Diabetes Study (UKPDS) risk engine — were developed primarily in White cohorts. Applying these tools to non-White populations can lead to miscalibration, either overestimating risk in some groups or underestimating it in others. For instance, the commonly used HbA1c diagnostic threshold may reflect different average glucose levels across ancestries due to variations in red blood cell lifespan and glycation rates, potentially causing missed or delayed diagnoses in African and Asian populations.

Differential Treatment Efficacy

Pharmacogenomic differences mean that the same diabetes medication may produce varied glucose-lowering responses, side effect profiles, and risk of complications across racial or ethnic groups. Metformin, the first-line agent for type 2 diabetes, shows reduced efficacy in some African American patients due to a higher prevalence of MATE1 transporter variants that increase renal clearance. Similarly, the thiazolidinedione class (e.g., pioglitazone) appears more effective in South Asian populations who have greater insulin resistance, yet clinical guidelines rarely specify ancestry-specific recommendations. Without prespecified subgroup analyses powered for race or ethnicity, these nuances remain hidden in aggregate results.

Barriers to Retention and Adherence

Even when minority populations are recruited into clinical trials, retention rates are often lower due to logistical barriers (transportation, childcare, time off work), cultural mistrust of medical research (rooted in historical abuses like the Tuskegee syphilis study), and lack of culturally tailored communication. This differential attrition introduces bias, as the participants who remain may not represent the broader population. Strategies such as community-based participatory research, use of multilingual materials, and provision of trial-related support services have been shown to improve retention and data completeness among minority participants.

Addressing Disparities Through Culturally Responsive Research

To move toward health equity, the diabetes research community must adopt a multipronged approach that goes beyond simply increasing enrollment numbers. True inclusion requires intentional study design, culturally competent recruitment, and analytical methods that capture intersecting identities.

Designing Culturally Sensitive Studies

Study protocols should incorporate culturally appropriate dietary interventions, accommodate religious practices around fasting (e.g., Ramadan for Muslim participants), and consider health literacy levels when designing informed consent processes. Pragmatic trials embedded in community health centers — such as the ALLIANCE study published in JAMA — demonstrate how leveraging local infrastructure and peer educators can reduce HbA1c disparities in low-income Black and Hispanic populations.

Improving Access to Healthcare and Research Participation

Addressing social determinants of health is critical. This includes expanding health insurance coverage, providing transportation vouchers for trial visits, offering telemedicine options, and mobilizing community health workers. The National Institutes of Health (NIH) has launched the All of Us Research Program, which aims to enroll one million participants from diverse backgrounds and will collect long-term health data including diabetes outcomes. Such infrastructure promises to generate data that can power subgroup analyses across racial and ethnic categories with adequate statistical power.

Analytical Innovations: Intersectionality and Stratified Medicine

Researchers should move beyond simple Black/White comparisons and adopt intersectional frameworks that consider interactions among race, ethnicity, sex/gender, socioeconomic status, and geographic location. Advanced statistical methods such as multilevel regression, latent class analysis, and Bayesian hierarchical models can uncover differential treatment effects without resorting to crude categorization. Moreover, machine learning algorithms trained on diverse datasets can predict which patients are most likely to benefit from a specific therapy, though care must be taken to avoid algorithmic bias.

Future Directions: Policy, Education, and Collaboration

Sustaining progress will require systemic changes at multiple levels. Funding agencies should enforce diversity requirements and require inclusion of prespecified analyses by race and ethnicity. Journals should mandate transparent reporting using standardized categories (e.g., the STROBE-MI guidelines). Investigator education must include training on culturally sensitive research practices, and community engagement should be funded as a core activity rather than an afterthought.

Additionally, the rise of real-world evidence from electronic health records, payer databases, and wearable devices offers new opportunities to study diabetes outcomes in diverse populations beyond the controlled confines of clinical trials. However, these data sources also carry biases related to healthcare access and documentation that must be addressed through rigorous sensitivity analyses.

Conclusion

Race and ethnicity profoundly shape outcomes in diabetes clinical research, influencing everything from genetic predisposition and drug metabolism to trial participation and retention. Dismantling the historical legacy of underrepresentation and bias requires deliberate, transparent, and culturally competent research design. By embracing diversity not as a box to check but as a scientific imperative, the field can produce evidence that is both more robust and more equitable. Only then will the fruits of diabetes research benefit every community burdened by this devastating condition.