Pharm Tech

How AI Expands the Reach of Clinical Trials


How AI Expands the Reach of Clinical Trials

Clinical trials of drugs are an essential part of determining whether or not a drug is useful and safe enough to release to the public. It takes a long time to reach that point, and many new drugs are never even approved to go to trials. AI in clinical trials is changing how things are done.

The Trouble With Clinical Trials

Clinical trials tend to be heavily biased. This is mainly due to the fact that scientists figured anything that worked for a white male test subject would be equally useful for everyone, disregarding their gender or race. This means that most drugs are tested mainly on white males and then approved for everyone’s use. Despite the fact that around 40% of Americans are in a minority group, up to 90% of participants in clinical trials are white.

More recent studies have shown that ethnicity and race, as well as gender, can have a major effect on how medicine and treatments work. For example, certain asthma medications work very well on white people but are ineffective when used by African Americans, who tend to have more severe asthma. The anti-platelet drug, clopidogrel, is useful for many people, but Pacific Islanders rarely get any benefit from it. It’s become quite obvious that medicine needs to be tested on a variety of people to truly study its effects on everyone. Diversity is key in clinical trials, but it is severely lacking.

Not only do clinical studies need to include more diversity, but they would also benefit from being skewed in favor of the groups that are most likely to need the medication. For example, Mexican Americans and Puerto Ricans tend to have a higher risk of developing diabetes, so it makes sense to include more of them in a trial of treatments for diabetes.

The idea of making trials more diverse has been around for some time but isn’t implemented as much as it should be. The National Institutes of Health Revitalization Act required researchers to bump up the number of women and people of color in their trials and research. It did have some impact, but not enough. In fact, clinical trials began to use more women, but the vast majority were white women, and various ethnicities were still not included.

In 2014, a study showed that out of 10,000 cancer clinical trials, just 2% of them were concentrated on ethnic groups. This is despite certain types of cancer being more predominant in specific racial groups.

Why Clinical Trials Are Biased

clinicaltrial, AIalgorithm, artificialintelligence, aiinclinicaltrials,  medicaltrials

There are several factors when it comes to the reason there are so many white people in clinical trials. First, it’s theorized that many minorities are not in a position where they have the time or money to participate in a trial, as they have to work or care for children.

Perhaps one of the main reasons frequently cited is the fact that minorities have previously received poor treatment when it comes to clinical trials. For example, during the Tuskegee Syphilis Study in 1932, African American men with syphilis and were intentionally not treated so researchers could study the disease’s progression. The unethical study, which lasted until 1972, resulted in a severe distrust of clinical trials for many people of color.

This is just the most famous example of the exploitation of a minority in the world of medicine. It’s not surprising that there is a distinct distrust of the medical system. Many African Americans don’t believe that the researchers would fully explain the trial risks and that they would likely encourage participation even if it was possible to experience harm. Strong historical backgrounds in the abuse of minorities contribute greatly to the problems in these studies.

While expanding the search pattern is part of the solution, there is also a need for building trust and providing better medical care to people of color. Minorities have historically been given lower quality medical care, even for very basic situations, such as childbirth. This also fosters distrust and has affected participation in clinical trials.

Another factor is the language barrier. If a Spanish speaker isn’t sure what the trial consists of, they’re far less likely to participate. Having translators is a big part of ensuring everyone is comfortable with the trial and understands what is required of them.


clinicaltrial, AIalgorithm, artificialintelligence, aiinclinicaltrials,  medicaltrials

How AI Changes the Narrative

AI may not be able to eliminate the distrust out there, but it can certainly give researchers a larger pool to choose from. The more potential participants there are, the more likely some of them will be willing to take part in a medical trial.

How can using AI in clinical trials make things more diverse? Stanford and Genentech put together a team of researchers, led by James Zou, to see if they could find a better way to expand the clinical trial pool.

The team focused on using artificial intelligence to determine eligibility rules for the trials. The result was an AI algorithm called Trial Pathfinder. It aids researchers in designing appropriate eligibility rules for clinical trials and then sorts patients via their electronic health records. The process is designed to expand the pool of potential participants, and it worked. Trial Pathfinder doubled the number of people that could be invited to participate in the trial and also created more diversity in gender, race, and age.

In the past, it would take far too long to go through all of this information and process it to select the best candidates for the trial. However, with AI, the process is quite rapid and provides a list of eligible participants long before humans could sort through it all.

Not only does AI give you set standards by which to create trial criteria, it helps encourage consistency. Previously, each trial had its own criteria for whether the patient could be included. The constant shifting of information made it more difficult to determine who was eligible or not. The AI algorithm sets exact boundaries and ensures that each trial is similar.


Previous Article

7 Ways That Personalized Medicine Will Revolutionize Healthcare and Drug Research

Next Article

How Those With Impaired Limb Movement Can Communicate On-Screen

Related Posts