Justice In Human Participant Research: Reinterpreting Belmont For Contemporary Clinical Trials

By Stephanie Pyle, MFA, CIP

When the National Commission for the Protection of Human Subjects of Biomedical and Behavioral Research published the Belmont Report in 1976, it became a foundational document for many policies, regulations, and guidance for research involving human participants.

However, the world has changed quite a bit since Belmont. New technologies, therapeutic modalities, and cultural perspectives have dramatically altered research design and conduct today. For example, expanded capabilities for data collection and analysis provide researchers with more information than ever before, data volumes that require increased effort to sift out meaningful details. Also, specialized study designs, such as platform and umbrella trials, and the evolution of personalized medicine provide new ways to help patients and simultaneously increase study conduct complexity.

Today, the Belmont principles remain the bedrock for safe and ethical clinical trials, even as clinical trials continue to grow in complexity. This article revisits the research community’s Belmont roots to consider what the report’s principle of “justice” means for research studies developed and conducted today.

Historical And Regulatory Context Of Justice In Clinical Research

Justice emerged as a corrective to mid-20th-century abuses wherein researchers used disadvantaged populations in research that principally benefited other groups. To avoid exploitation and ensure fair opportunity to benefit from the research, the National Commission codified justice as one of the three pillars of ethical human subjects research in the report.

In the context of clinical research participation, justice means that no one group of people should bear an unfair share of risk. It also requires fair distribution of treatment by avoiding discrimination between classes of people or among members of the same class. (It’s also worth noting that, when drafting Belmont, the National Commission considered including “justice, equal protection and equal opportunity; and equity.” While it simplified the wording in its final version, this consideration provides useful insight into the commission’s intentions.)

Today, the Federal Policy for the Protection of Human Subjects found at 45 CFR 46 (which is based on the National Commission’s work) requires most research involving human participants be reviewed by an IRB; the Belmont principles ensure sponsor companies minimize and inform the regulations governing IRB review.  

For the principle of justice in particular, IRB approval criteria requires confirmation that also participant selection is equitable. In a tip sheet to IRB members, NIH explains the goal of this criteria is to ensure that “both benefits and burdens of research are distributed fairly, and that no specific population is being unfairly targeted or excluded” without valid scientific or ethical rationale.

Contemporary Challenges To Equitable Participant Selection

Lawmakers ostensibly designed federal regulatory provisions to balance participant protection against undue exclusion. In practice, however, such protective mechanisms can result in systematic omission of the very populations that should be represented in research. Below, we discuss some aspects of clinical research that may contribute to a miscarriage of justice.

Evolving Clinical Trial Complexity

Clinical research has grown more complex, with adaptive designs, precision therapeutics, decentralized or remote elements, and multinational recruitment strategies introducing operational and analytic considerations that can unintentionally skew whether participant enrollment fully represents the intended study population. Strict eligibility filters, convenience-based site selection, and reliance on existing investigator networks all favor recruitment from readily accessible populations, not necessarily those most affected by the disease under study.

Overprotection and Unjust Exclusion

Researchers sometimes use regulatory safeguards intended to reduce risk — such as special protections for pregnant people, children, and cognitively impaired individuals —to exclude these groups categorically, rather than facilitate their safe inclusion. Historical responses (e.g., restrictive post-thalidomide policies toward women of childbearing potential) demonstrate how overcorrection can persist across decades and limit evidence for significant patient populations.

Operational Barriers and Accessibility

Practical constraints in clinical trial development and execution can also disproportionately limit access for racial and ethnic minorities, older adults, non-English speakers, and people of lower socioeconomic class. For example, language barriers and a lack of translated consent materials may preclude sites from enrolling non-English speakers. This excludes all non-English speaking patients from the research, even if they meet all enrollment criteria, and yields non-representative study data. Transportation and scheduling burdens might similarly limit participation to only participants with sufficient time and resources.

Technological requirements for participation in research can also exclude certain groups; for example, requiring participants have access to a stable internet or cell service connection may exclude people in rural geographies where coverage is spotty or certain economically disadvantaged people. Such practical constraints compound analytic risks created by non-representative sampling.

Emerging Tools

AI and machine learning tools can improve protocol design, study populations, and site selection. AI might also be used for data analysis, hypothesis generation, and automation of routine tasks.

Inherent with any transformative tool, there are impacts, considerations, and the need to have a human in the loop.

An AI tool trained on incomplete data or data from biased sources may yield biased and unjust participant recruitment strategies. In dermatology, AI can exacerbate skin tone biases. When AI algorithms for skin cancer diagnosis are developed using datasets that underrepresent darker skin types, diagnosis for people of color may be significantly delayed. Image contrast issues with darker skin can also impact AI’s accuracy.

Validating AI tools to detect possible hallucinations, data privacy concerns, or bias is important to manage the risk of algorithmic approaches amplifying existing inequities under the guise of efficiency or other stated benefits.

Consequences Of Inadequate Representation

While not all exclusionary practices occur intentionally or explicitly, they all yield real consequences. Consider the following:

• Scientific Validity: Trials that underrepresent subgroups produce results with limited external validity. Important subgroup differences in pharmacokinetics, drug metabolism, efficacy, and adverse-event profiles remain undetected, reducing the reliability and generalizability of evidence for diverse clinical populations.
• Harms to Clinical and Public Health: Clinicians relying on incomplete trial evidence may make decisions that expose patients to inappropriate dosing, unexpected adverse events, or ineffective treatments. The Ambien (zolpidem) dosing revisions for women issued decades after approval exemplify how sex-based exclusion or inadequate subgroup assessment can produce long-term patient harm.
• Regulatory and Commercial Impacts: Regulators may constrain labeling and promotional claims where safety and efficacy data lack subgroup representation, forcing sponsors to conduct supplemental studies or restrict marketing. These outcomes increase time and cost to market and can also damage credibility for sponsors, institutions, individual healthcare providers, and the life sciences more broadly.

The Role Of IRBs And Institutional Oversight

IRBs are uniquely positioned to evaluate equitable selection during protocol review. Practical IRB responsibilities include:

• assessing eligibility (i.e., inclusion and exclusion criteria) to minimize the risk to participants and ensure equitable inclusion of participants based on the purpose of the research and where the research will be conducted
• evaluating recruitment and retention strategies for their ability to reach affected populations
• ensuring consent processes and participant materials are accessible and culturally appropriate
• assessing monitoring plans and enrollment targets aligned with disease epidemiology
• documenting determinations and conditions in meeting minutes and correspondence.

Additionally, IRBs should be wary of unnecessary protective measures and instead require investigators to propose calibrated safeguards that enable safe inclusion.

Operational Strategies To Advance Justice

Each stage of clinical research development and conduct presents opportunities for stakeholders to increase justice and fairness. These strategic approaches can help avoid the consequences of inadequate representation:

• Epidemiology-Driven Eligibility in Study Design
  • Define target populations based on burden of disease prevalence, natural history, and biological considerations.
  • Narrow exclusions to those scientifically necessary and prespecify subgroup analyses and stratification strategies to detect differential effects.
• Site Selection and Recruitment
  • Use data-driven site selection and include community-based sites that serve affected populations.
  • Incorporate registries, EHR analytics, and validated AI tools to identify potential sites and participants.
  • Build relationships with community clinics and advocacy organizations to improve outreach and trust.
• Consent and Accessibility:
  • Normalize the translation of consent and study materials into languages spoken by the community.
  • Employ plain language, teach-back methods, interpreters, and multimedia tools.
  • Reduce logistical burdens through reimbursement, transportation, childcare support, and flexible scheduling. Integrate decentralized trial elements where appropriate and while ensuring digital inclusion.
• Monitoring and Adaptive Mitigation
  • Implement enrollment dashboards tracking sex, race, ethnicity, age, language, and other pertinent characteristics.
  • Establish prespecified mitigation triggers and corrective action plans, such as opening additional sites, adjusting outreach strategies, or revising eligibility criteria when enrollment deviates from target representation.
• Inclusion Safeguards Instead of Exclusion Criteria
  • For populations with uncertain safety profiles, adopt staged enrollment, sentinel cohorts, and enhanced monitoring instead of categorical exclusions.
  • Ensure consent processes and oversight mechanisms address special vulnerabilities without denying participation.
• Accountability, Metrics, and Transparency
  • Incorporate equity impact statements into scientific review and feasibility assessments.
  • Define enrollment targets tied to epidemiology, publish enrollment demographics, and report progress to sponsors, IRBs, and registries.
  • Tie resourcing and investigator performance metrics to equity objectives.
• Responsible Technology Strategies
  • AI and ML can accelerate identification of eligible cohorts and optimize site selection, but they must be applied with explicit safeguards. Validate algorithms across diverse data sets, measure fairness metrics, maintain human oversight, and monitor for disparate impact.
  • Decentralized trial technologies expand reach but require strategies to avoid digital exclusion.
• Institutional Culture and Capacity Building
  • Justice requires institutional commitment. Fund translation and outreach, provide training on implicit bias and cultural competency for investigators and IRB members, and establish community advisory boards to inform study design and retention strategies.
  • Equity should be resourced during feasibility planning, not treated as an unfunded afterthought.

Considerations For The Research Community

Stakeholders from all corners of research can encourage fairness and justice in healthcare. Whether working directly with participants, every role can contribute to better representation and better science. Consider implementing the following procedures:

• Conduct an equity impact assessment. Prespecify subgroup analyses and monitoringing plans proportional to likely heterogeneity of effect.
• Normalize translation processes as standard study activation activities and ensure culturally appropriate consent materials for populations in the study or site catchment.
• Select sites using validated, data-driven methods rather than experience alone, and include fairness checks when using AI tools.
• Implement enrollment dashboards and pre-triggered mitigation strategies.
• Favor inclusion with calibrated protections rather than reflexive exclusion.
• Include details on demographic enrollment and post-approval subgroup data in appropriate public-facing communications, where feasible.

Justice in human subject research is both an ethical imperative and a scientific necessity. Contemporary trial design and execution must move beyond procedural compliance to embed equitable selection, accessibility, and accountability throughout the research life cycle. Doing so strengthens evidence of quality, reduces patient harm, minimizes regulatory setbacks, and upholds the social compact that underpins ethical research.

IRBs can help operationalize justice through equitable review and approval of just research practices. Research sponsors can design studies with clear justice metrics, such as appropriate representation. Investigators and institutions can advance justice by ensuring adequate resources, maintaining community partnerships, and continually monitoring the research. Re-centering justice in this manner ensures that the benefits of research are distributed fairly, and that clinical practice is informed by evidence applicable to the populations it serves.

Acknowledgements: The author would like to thank Lauri Carlile, Joshua Fedewa, Jan Hewett, and Julie Ozier for their contributions and support throughout this article’s development.

About The Author:

Stephanie Pyle is a regulatory documentation manager and a certified IRB professional. With 20+ years of experience in education, communications, and IRBs and human subject protections, she specializes in transforming complicated information into useful and meaningful resources. Stephanie works with IRB members, regulatory staff, and organizational leadership to develop clear, consistent regulatory documentation for internal and external audiences. She was also the longtime moderator of a popular IRB educational webinar series. Prior to joining the research community, Stephanie taught creative writing and rhetoric, and composition at Penn State.