Increasing Diversity in Clinical Trials Using Deep 6 AI

Increasing Diversity in Clinical Trials Using Deep 6 AI

Clinical research is a necessary component to developing new medical treatments, serving as the method of determining their safety and efficacy in humans. Clinical trial participants should reflect the demographic breakdown of the patients that will need those novel therapies. However, while many diseases and disorders disproportionately affect ethnic minorities, or affect people differently based on genetic makeup, minority groups are majorly underrepresented in medical research for new treatments. There are a variety of reasons for this. Principal investigators (PIs) often limit their search for study participants to their own patients that they treat regularly, rather than widening their search to reach the entirety of their health system’s patient population. Implicit bias is also an issue that still exists in healthcare, even in medical professionals, which can result in minorities not often being presented with the option to join active studies. Many other factors come into play as well such as language and cultural barriers hindering one’s ability to understand medical practices like clinical trials, lack of resources in underserved neighborhoods, limited outreach tailored to people in these communities, and so on.


One of our goals at Deep 6 AI is to help overcome this substantial lack of diversity in clinical trials. Using artificial intelligence (AI) concepts like natural language processing (NLP) and machine learning (ML), we can decrease human biases and extend trial participation opportunities to other communities to find the best-matching patients for trials. These methods can help increase diversity, equity, and inclusion to allow for more access to innovative life-saving treatments for anyone who might benefit from them. 

The Origins of Non-Equitable Research

Because most well-known ailments were first studied long ago with the advent of modern medical and clinical research practices, trials focused largely on white, male subjects. Historically white communities were provided with more resources, better access to medical care, and more attention from healthcare providers. In addition, there is a level of distrust in the medical and research communities from minority groups, stemming from the United States’ history of medical mistreatment of minorities (e.g., the Tuskegee Experiment). This mistreatment and distrust, and thus lack of diversity in research, has affected people of different races, ethnicities, genders, religions, sexual orientations, socioeconomic status, etc.

While ethical guidelines like the Belmont Report in 1976 and laws like the NIH Revitalization Act of 1993 have been established, clinical research has been slow to reflect North America’s overall growing diversity. As a result, even today, trial participation is not quite reflective of the demographic breakdown of the greater population. This lack of representation in research has led to negative ramifications for our current understanding of how different diseases affect different types of people, and how effectively those people are treated. Current ongoing efforts to combat this issue include the FDA’s Office of Minority Health and Health Equity’s “Diversity in Clinical Trials Initiative” and the Biden Administration’s Cancer Moonshot Initiative, which includes a focus on increasing diversity in trials.

How Technology Helps Make Research More Inclusive

As technology has developed over time, we can reach new populations who had not had prior access to certain medical care, including clinical trials. We’re also able to use it to counteract the effects of the human biases that have led to today’s gaps in trial participant diversity. AI can carry out functions that were previously only done manually, which significantly decreases the effects of inequitable practices and limited outreach to wider populations.

Deep 6 AI is the leader in precision research software, connecting all research stakeholders in an AI-powered, real-time, data-driven, collaborative ecosystem. The software works by using ML and NLP to search through both the structured and unstructured data in a health system’s electronic medical records (EMR), thereby massively accelerating and improving the patient recruitment process. Outdated tools and processes would require researchers to manually sort through unstructured data, like scattered physician’s notes, often from multiple sources and different systems. Using Deep 6 AI, experts can widen eligibility criteria as a way to potentially increase trial diversity for comorbidities that disproportionately affect diverse patient populations.

Within our platform you can easily edit the exclusion and inclusion criteria as you go and get real time feasibility counts to make sure it’s necessary to have certain criteria based on what patient populations are eliminated with each one. Otherwise, some criteria may inadvertently exclude groups of people which may be affected by the study results. Once a trial’s criteria are input, a list of potential participants is presented, leaving the researcher only to have to easily validate these matches before reaching out to patients to join. 

Typically, many doctors would only tell their patients about a trial whenever they come in for their next appointment, or PIs would rely on referrals to reach patients they don’t personally treat. The Deep 6 AI platform is implemented on a health system’s entire EMR. So, when a search is conducted through the software, the list of potential matching patients spans all the hospitals and doctors’ offices that are in that health system, including those who receive treatment in community clinics.

The sponsors we work in our ecosystem use the platform to find the best sites to run their trials. One of the biggest causes of a failed trial is a site’s inability to meet its recruitment goals, largely due to a misunderstanding of the makeup of their patient population. With a better understanding of the types of patients required for a study, more sites and sponsors can easily meet their recruitment goals, including diversity goals. Patients in community hospitals can be approached for studies more often. Often, patients with cases that require more novel treatments because the standard of care is not good enough, such as those with terminal illnesses, go looking for trials on their own. Because of so many intricacies in clinical trial criteria, they are often not matches for the studies they try to join.  

When physicians have access to better tools, they can have confidence in bringing the best new treatments to their patients. As technology continues to evolve, we hope to keep playing a part in the efforts toward more inclusive and equitable treatment options for all people. 



Nour Malki, Communications and Content Specialist