Enhancing diversity in clinical trials with AI and human expertise

Achieving diversity and inclusion in clinical trials is both a moral imperative and a scientific and regulatory necessity. Diverse participant populations ensure that study findings are applicable to a broader range of patients, leading to more effective and equitable healthcare solutions. Despite efforts to improve diversity, many trials still struggle to recruit representative patient cohorts. Improving diversity in clinical trials requires a multifaceted approach. By harnessing the power of artificial intelligence (AI) augmented by human expertise, we can unlock new opportunities to enhance patient diversity in clinical trials through data-driven site selection.

The regulatory and scientific necessity for diversity

Regulatory bodies have increasingly emphasised the importance of diversity in clinical trials. For example, the FDA's guidance on diversity in clinical trial populations highlights the need for including underrepresented demographic groups to ensure the safety and efficacy of medical products. The FDA also requires diversity action plans to ensure concrete steps are taken to improve diversity in a trial and that it is strategically incorporated into planning as early as possible.

Diverse clinical trials are scientifically crucial for several reasons:

Generalisability of results: When trials include diverse populations, the findings are more likely to be applicable to a wide range of patients. This helps in developing treatments that are effective across different demographic groups, reducing health disparities.
Understanding differential responses: Different populations can respond differently to treatments due to genetic, environmental, and lifestyle factors including social determinants of health. Including diverse participants in trials helps identify these variations and tailor therapies accordingly.
Safety and efficacy: Ensuring a treatment’s safety and efficacy across all target populations requires testing in those segments. Without diverse participation, some groups may face higher risks or reduced efficacy from treatments developed based on homogenous data.

Demonstrating robust safety and efficacy data across appropriately representative populations is key for registrational studies, regulatory approval and commercialisation. Selecting sites that improve access to target communities can significantly benefit a trial by getting the right data the first time, avoiding costly delays or additional trials.

Integrating diversity data in decision-making

Accurately measuring diversity data requires more nuance than assigning an overall diversity score. It involves understanding the diversity within the disease population and the therapy’s target population. Leveraging data like prevalence to participation ratios (PPR) and participation consumption ratios (PCR) helps identify the appropriate target populations, overlooked or underrepresented groups, and groups with highest unmet need.

Human-enabled AI can analyse vast datasets with precision and efficiency. One Search, for example, can integrate multiple datasets like our Symphony Health Integrated Dataverse®, claims data, census data, and other third-party data to provide researchers with clear insights into enhancing diversity. Understanding which patient populations have the highest need, where they are located, and their unique patient journeys will help optimise site selection decision-making along with other operational and feasibility criteria.

Identifying diverse connections in site selection

Specialised AI like ICON’s One Search allows us to assess the connectedness of sites, institutions and investigators and their access to patients. Implementing human-enabled AI to map these connections, has shown that increased connectedness correlates to higher enrolment performance. AI can rank the quality of these connections by overlaying additional layers of data, including evaluations of site performance and core clinical fit, ensuring a more comprehensive and effective recruitment strategy.

Using One Search, we can select sites connected to underrepresented communities to improve clinical trial access within key geographies and communities. Access to clinical trials is a major hurdle to participation for diverse patient populations, and using a novel lens to identify sites opens possibilities to build trusting relationships with communities, clinicians and patients that have historically not been included.¹

The role of investigators

Gathering additional population data and demographic breakdowns of the institutions’ patient populations and examining what patients the principal investigators are treating is crucial.

Diversity is also a challenge within healthcare professions. Clinicians who share a cultural alignment with the target populations are more likely to successfully enrol those patients into trialsi due in part to the inherent cultural competency of shared culture or community, enhanced trust building, and a deeper understanding of patient motivations and barriers for participation. For example, there was a reported 50% increase in the proportion of female participants in heart failure trials when the investigators were female.²

Choosing connected investigators can boost enrolment and support expedient recruitment solutions. On the other hand, ineffective investigators can cause costly delays or failure to meet recruitment goals, jeopardising the feasibility of a trial. Fortunately, the right AI can leverage demographics data, past performance, and patient data to evaluate and predict investigator effectiveness in recruitment. The result is a refined, country-specific list of investigators that feasibility analysts can incorporate into site assessments. This data-driven personnel analysis contributes to accelerated site identification timelines, enhances investigator engagement, and increases the likelihood of reaching overall diversity and patient recruitment goals.

Conclusion

Diversity in clinical trials is essential for ensuring that clinical research and the resulting treatments are effective for broader populations. By combining AI with human expertise, tools like One Search offer a promising path forward and enable sponsors to better address the complex factors that influence patient participation in a study.

This human-enabled AI process integrates the identification and recruitment of diverse participant populations into site selection, ensuring that clinical trials are both inclusive and effective to meet scientific rigor and regulatory requirements. Embracing these innovative strategies will not only enhance the quality of clinical research but also contribute to a more equitable healthcare system and meaningful therapies for patients in need.

To learn more about unlocking the benefits of AI-enabled site selection, read our whitepaper or connect with us today.

References

¹ Zhu, J, D'Angelo, F, Miranda, J. et al. Incorporating Cultural Competence and Cultural Humility in Cardiovascular Clinical Trials to Increase Diversity Among Participants. J Am Coll Cardiol. 2022 Jul, 80 (1) 89–92. https://doi.org/10.1016/j.jacc.2022.05.001
² Reza N, Tahhan AS, Mahmud N, DeFilippis EM, Alrohaibani A, Vaduganathan M, Greene SJ, Ho AH, Fonarow GC, Butler J, O’Connor C, Fiuzat M, Vardeny O, Piña IL, Lindenfeld J, Jessup M. Representation of Women Authors in International Heart Failure Guidelines and Contemporary Clinical Trials. Circ Heart Fail. 13 (2020) e006605.

The regulatory and scientific necessity for diversity

Integrating diversity data in decision-making

Identifying diverse connections in site selection

The role of investigators

Conclusion

In this section

Connect with us