Resource • Article
Leveraging Real-World Data and AI to Broaden Patient Diversity in Oncology Clinical Trials
Source: Anosheh Afghahi, MD, MPH, Medical Director, Flatiron Health
Summary: Technological advances combined with FDA “Diversity Action Plans (DAPs)” intersect to enhance clinical trial design and foster more inclusive outcomes. The convergence of recent technological advancements and DAPs holds great promise for creating more inclusive clinical trials, benefiting all communities and advancing medical research.
A New Landscape for Clinical Trial Design
On June 26, 2024, the U.S. Food and Drug Administration (FDA) released draft guidance for the clinical trial industry titled “Diversity Action Plans to Improve Enrollment of Participants from Underrepresented Populations in Clinical Studies.” Although currently a non-binding recommendation, this guidance outlines expectations for clinical studies requiring a Diversity Action Plan (DAP), including the format, content, and submission process.[1] The purpose of DAPs is to increase enrollment of historically underrepresented populations in clinical trials, providing more inclusive data for patients who may use the medical product.[2]
The Impact of Recent Technological Advancements on Clinical Trial Design is Undeniable
In the context of clinical trials, technology continues to be leveraged for new solutions in patient recruitment efforts and treatment. The emergence of Artificial intelligence (AI) has also assumed a pivotal role in data collection, aggregation, and analysis. More than ever, sponsors are actively utilizing data sets, electronic health records (EHRs), algorithms, and natural language translation to enrich clinical trial design.
| Examples of areas where technological advancements have impacted clinical trial design include[3]: |
| 1. Patient Recruitment and Retention |
| 2. Data Management and Analysis |
| 3. Remote Monitoring and Decentralized Trials |
| 4. Regulatory Agencies Embracing Digital Solutions |
Using Real-World Data to Inform Clinical Trial Design
Real-world data (RWD) encompasses observational data collected beyond controlled clinical research settings. This data includes information from sources like EHRs, medical claims, and patient surveys. Researchers use RWD to study outcomes in diverse patient populations within real-world contexts. These insights can then better inform healthcare research, outcomes, and decision-making.[4]
Innovative trial design, such as optimized trials, prioritizes both the patient experience and accessibility. By reimagining how clinical trials are conducted, optimized trials can break down barriers that historically limited minority participation in research. This approach considers cultural nuances, language preferences, and community engagement, fostering trust and participation – all with the aim to increase enrollment of underrepresented populations. Inclusion leads to more accurate and representative outcomes, ensuring that medical interventions work effectively across diverse patient populations.
Leading the Way
In the dynamic field of oncology, Flatiron Health, a healthcare technology company dedicated to improving cancer treatment and advancing research has made it its mission to improve quality of care by using RWD to improve patient outcomes, inform policy, and advance research through applications of technology.[5]
Dr. Anosheh Afghahi is a medical director at Flatiron Health, leading protocol optimization studies and initiatives centered around improving clinical research eligibility requirements for oncology studies.
“Flatiron is known for its high-quality real-world data,” said Dr. Afghahi. “At Flatiron, we believe learning from the experience of every person with cancer is an imperative – it is the key to accelerating research and continuing to improve the quality of care. We offer clinical research services to help address real-world evidence gaps, combined with point of care software, to deliver more efficient and inclusive clinical trials to our network of community oncology clinics.”
This approach distinguishes itself from traditional academic or biopharma methods in several significant ways:
Community Oncology Focus:
- Leveraging an extensive network of community oncology sites.
- With the majority of cancer patients receiving treatment in community settings, Flatiron optimizes data from electronic health records (EHRs) consisting of approximately 3.7 million patients treated in these community environments.
- While community oncology centers already support robust clinical trial portfolios, there remain opportunities for sponsors (biopharma and contract research organizations) to enhance study availability within the community.
Integrated Technology:
- Flatiron’s EHR system is seamlessly integrated into community oncology sites.
- This integration allows for efficient data gathering from their own EHR and facilitates clinical trial assessments.
- Additionally, Flatiron offers a tool called Flatiron Clinical Pipe™, an EHR-to-electronic data capture (EDC) connector, streamlining data transfer to reduce manual monitoring and data entry burdens for clinical research personnel.
By minimizing infrastructure requirements, this approach encourages more community oncology sites to participate in studies, benefiting diverse patient populations. And for Dr. Afghahi, that is the real goal – reaching more patients, especially the underserved.
Data Indicates that Improved Trial Design Leads to More Inclusive Outcomes
Prospective research benefits from the inclusion of retrospective research in several ways. Retrospective studies, which analyze existing data, provide valuable insights that complement prospective study designs. They help generate hypotheses, identify associations, and validate findings. By combining both approaches, researchers gain a more comprehensive understanding of research questions, enhance efficiency, and strengthen overall scientific knowledge.
“We now have Clinical Research Services that are prospective, while the real-world data that had been collected over the years was primarily retrospective,” said Dr. Afghahi. “Traditionally, with patients who have been seen over a span of years, it’s more a case of looking backwards to learn about their treatments and the patterns of care. But how could we do things differently, prospectively? The answer lies in mixing both the retrospective data as well as applying new point of care technologies that can help us deliver more efficient and inclusive clinical trials to our network of community oncology clinics.”
“Ultimately, we want new drugs to be helpful, successful and safe… and not for just one group of people, but for the representation of patients across the United States and the globe.”
– Dr. Anosheh Afghahi
Building the Case for Broadening Trial Eligibility
Dr. Afghahi builds upon this research in a Flatiron article entitled; “Broadening Eligibility Criteria and Diversity among Patients for Cancer Clinical Trials.” In this study, the Flatiron Health team analyzed clinical trial eligibility criteria across 22 cancer types. The researchers used RWD to evaluate the impact of broadening eligibility criteria on patient diversity.[6]
The study utilized a nationwide EHR derived de-identified database to identify retrospective cohorts of patients diagnosed with 22 different types of cancer who received systemic therapy over a span of nine years. Strict versus broad eligibility criteria were compared (based on traditional clinical trials’ inclusion/exclusion criteria) in terms of liver, kidney, and hematologic function around the time of the patient’s first line of therapy in the advanced or metastatic setting. The study emphasizes the importance of considering diverse patient populations in clinical trials to improve generalizability and reduce health disparities. Broadening eligibility criteria is one approach to inclusive clinical research and achieving enrollment goals.8
Key findings included:
- Eligibility Criteria and Patient Exclusion:
- Applying common strict cutoffs for eligibility criteria, only 48% of patients were eligible for clinical trials.
- Certain groups were more likely to be excluded using strict criteria:
- Female patients (odds ratio: 1.30)
- Older patients (age 75+ vs. 18-49 years old: odds ratio: 3.04)
- Latinx patients (odds ratio: 1.46)
- Non-Latinx Black patients (odds ratio: 1.11)
- Lower socioeconomic status (SES) patients.
- Broadening criteria increased the number of eligible patients by 78%, with the strongest impact for older, female, non-Latinx Black, and lower-SES patients.
- Real-World Overall Survival (rwOS):
- Patients who met only broadened criteria had worse rwOS compared to those with strict criteria (hazard ratio: 1.31).
- Conclusion:
- Data-driven evaluation of clinical trial eligibility criteria can optimize inclusion of historically underrepresented groups and promote access to more inclusive trials.8
Dr. Afghahi summarized this work: “We learned that when applying the traditional strict cut offs for eligibility criteria to patients, 48% were eligible. Broadening the criteria increased the number of eligible patients by 78%, with the strongest impact being for older, female, non-Latinx Black, and lower SES patients.
“Our study provides evidence to show that broadening eligibility criteria could allow investigators to potentially increase the proportion of eligible patients from historically underrepresented groups, leading to more inclusive trial results. These findings can also be used to model the effect of adjusting eligibility criteria on trial endpoints, such as real-world overall survival.”
New Guidance and Technology: Obstacle or Opportunity?
With the FDA supporting broadened eligibility criteria, and the influx of rapidly changing technology, Dr. Afghahi sees an opportunity to improve clinical trial design in oncology to include a broader spectrum of patients.
“The FDA has issued a draft guidance that outlines diversity action plans including for clinical studies,” said Dr. Afghahi. “Ideally, we would like for the clinical trial population to be reflective of the real-world population that will be taking these medicines to understand their efficacy, their potential toxicities and of course to ensure that everyone has access to potentially lifesaving medicines.”
Optimizing Trail Design for a Broader Population
Optimizing trial design to enhance diversity and use retrospective data for prospective studies is a potential pathway to adjusting clinical trial criteria to be more inclusive of all patient populations.
Consider these basic steps to broaden trial design and include more patient populations:
- Trial Optimization Phase: Before a study begins, make a close examination of the study, understanding the disease and reviewing the inclusion and exclusion criteria.
- Broadening Eligibility Criteria: Consider how changes to lab variables might impact the sociodemographic diversity of eligible patients.
- Using Retrospective Data: Retrospective data can be used to predict how adjustments to the inclusion and exclusion criteria might affect diversity metrics.
- Observing Changes: By manipulating the criteria, researchers can observe shifts in demographics, such as age, gender, and race, among eligible participants.
There is always a hesitancy around change in the clinical trial setting. And while Dr. Afghahi openly identifies some drawbacks, she sees the larger goal of inclusivity as being paramount.
“Perhaps some of the endpoints will be of a shorter duration (such as real-world PFS or real- world OS) if patients with more comorbidities or with lower performance status are eligible. However, it remains a primary imperative to understand the efficacy of these drugs for a broad patient population.
“The convergence of technological advancements and DAPs holds great promise for creating more inclusive clinical trials. By embracing diversity, we can advance medical research, improve patient outcomes, and build a more equitable healthcare system for all.”
