AXON Clinical Studies

Precision in practice: GenAI in clinical trials

In the dynamic world of the healthcare industry, clinical research stands at the forefront of innovation, using emerging techniques and technologies to bring generation after generation of new therapies to patients around the world.

With the buzz around generative AI (GenAI) growing at a rate of knots, it is no surprise that one of the most promising areas for the technology is in clinical trials. GenAI’s prowess in data analysis, its ability to transform arduous tasks into actionable assets in a matter of mere seconds, has the potential to empower navigating information with unprecedented speed, ushering in a new era of expedited, data-driven clinical research.

“The clinical trial space is, as we all know, an area where we have a lot of repetitive and very time-consuming tasks,” explains Silvia De Carvalho, clinical studies lead at AXON. “GenAI is really a tool that enables an acceleration in some of these tasks, where before it would be a very manual prolonged process for both sponsors and sites.”

With nearly 20 years of experience in digital communication and engagement, De Carvalho understands the multifaceted influence generative AI presents, as well as the challenges that remain with fully realising this potential. Here, she discusses the use of GenAI in clinical trials, from data analysis to patient engagement.

Navigating the uncharted waters of GenAI in clinical research

While the concept has been in development for more than a decade, the potency of GenAI has surged in recent years, with ChatGPT being perhaps the most prominent example in the public space.

In the realm of clinical development, De Carvalho explains that a primary and most immediate advantage of GenAI lies in its ability to swiftly process and summarise extensive structured and unstructured datasets, offering insights and learnings from various data sources, such as protocols within a therapeutic area. This, she says, has enormous potential in reducing the time and resources required to complete certain tasks.

“The power of GenAI has really exploded, and that’s why we now have the challenge of fully grasping its potential. One of the most immediate benefits is its ability to extract from content and summarise information, based on simple queries,” says De Carvalho. “Now, we have the capability not just to look at and understand text, but also imagery. It’s getting more sophisticated in its capacity to learn and create from that knowledge. Certainly, in the clinical development space, there are a lot of areas where that can be a beneficial tool in conjunction with specialists and analysts who can look at the data.”

To illustrate how GenAI can support decision-making, De Carvalho describes the process of protocol design. Rather than dedicate hours of precious time to scrutinising each protocol that touches the same therapeutic area, she explains that GenAI can perform the same task, quickly extracting and summarising the relevant information.

“The clinical operation lead can then look at this information and say, ‘Okay, these are the learnings’, whereas it would take hours for someone to go through all that data,” she says. “In terms of summarising existing data sets, it’s extremely powerful. The main concern that is still present is potential hallucinations (responses generated that contain false or misleading information presented as fact), which is why the process still requires human analysis and supervision to avoid creating insights from incorrect data.”

De Carvalho continues: “Another good example of its potential is its capacity to fill in for missing data, allowing researchers to work with more complete and reliable information for downstream analysis and modelling. Knowing how challenging it is for programs to gain complete data, this has the potential to really impact approval processes.”

Improving recruitment and retention

Successful recruitment and retention of trial participants are the two most challenging aspects of clinical research today. Whereas traditional – manual – approaches to identifying and enrolling candidates prove lengthy and inefficient, with the aid of GenAI, healthcare professionals can analyse large volumes of patient data to rapidly pinpoint eligible patients.

“Recruitment is so difficult and competitive,” says De Carvalho. “It’s not just finding the right patient, it’s finding the right patient ahead of the next trial, or another trial trying to recruit the same population. Being able to do that very quickly and efficiently is going to be critical in driving the speed that everyone is trying to achieve, while ensuring the right patient population is included. The other challenge is retention, because you might have initially recruited patients, but then you lose that population through the funnel.”

Many sponsors have been looking at new ways to communicate with potential patients, and help sites recruit. De Carvalho highlights two pioneers in the space: Pfizer and Novartis. Pfizer is using generative AI to personalise recruitment messages to potential patients and has created a chatbot that generates personalised messages that are tailored to the specific needs of that patient and can answer questions about the clinical trial.

Novartis is also creating personalised recruitment materials for a clinical trial for a new drug for psoriasis. The results of the campaign showed a 20% increase in the number of patients who responded to the recruitment campaign.

However, she notes, it is important that healthcare professionals remain closely involved with the identification process to prevent unintended biases from excluding demographics. Biases can be reinforced when the model learns from data sources that do not fully represent all communities. This is especially true for communities that rely on oral traditions or non-digital means of communication.

“It’s not perfect, she says. “GenAI should, by its nature, enable us to identify patients that we might not have recognised or identified at site level because now we’re looking at billions of parameters at the same time versus a couple. But, at the same time, you need that human check to make sure that we’re purposeful in the populations we’re including.”

Beyond initial recruitment, De Carvalho explains that GenAI can play an important role in improving retention rates.

Proactive communication, a hallmark of GenAI, will be instrumental in mitigating challenges and expediting the trials’ progression. As the industry transitions towards decentralised and hybrid trials, GenAI allows for increased touch points with participants who may not physically visit the trial sites, bolstering engagement and providing an opportunity to acquire data that can be used to accelerate the road to approvals, for example, through synthetic control arms.

“As patients are increasingly monitored at home, for instance with wearables, data points are being measured outside of sites and that data can be quickly seen and processed through the AI, enabling faster insights and optimisation,” she says.

Clinical trials can span months, and keeping individuals engaged requires time and resources to deliver effectively. However, today’s GenAI is sophisticated enough to help streamline and personalise communications efforts.

“Now, we’re able to produce initial communication that is personalised and engaging people early on,” explains De Carvalho. “That should also accelerate our ability to recruit certain populations that might not have been accessible based on just looking at all the electronic health record (EHR) data at the sites, identifying the patients, and then pushing communication to those patients.”

She continues: “If I can tell you every week or every month how you are doing, and guide you through the process, chances of accelerating the outcomes of the trial and chances of getting all the patients to stay from a retention point of view are also increased. Most sites are understaffed and struggling to effectively coordinate all that information at speed. GenAI can substantially decrease site burden as it takes over simple things, such as appointment scheduling, reminders, or even integrated notes across stakeholders to provide only relevant information to the patients.”

Addressing limitations and misconceptions

Integrating GenAI into clinical trials presents a promising frontier, yet, as De Carvalho is clear to note, it is not without its share of challenges and barriers. Of paramount concern is the need to address compliance and privacy issues associated with the integration of these tools into healthcare practices.

“That is probably still the number one challenge,” she says. “As we are identifying use cases for the technology, everybody’s trying to think, ‘What are the boundaries?’. If we’re asking the tool to look at EHR data and patient personal data, and we’re training the tool based on that information, is that compliant? Where is the line? Where is the data privacy bridge?”

The level of disclosure and consent required for patients participating in clinical trials, who may be unaware of AI leveraging their personal data and communicating with them, poses a significant ethical dilemma that demands careful consideration.

“AI is not a substitute for those relationships and those communications, and the transparency of knowing who’s communicated to you,” she says. “We don’t want to create a false sense of it’s all automated and de-humanised by leveraging too much AI. You still need to have a limit that requires people to be that check and balance of biases to make sure that you didn’t avoid an entire population because of an X-factor, for example.”

Of course, any tool – no matter how sophisticated – only performs as well as the person using it. This is particularly true in the instance of GenAI, where the comprehensive prompts and queries are essential to ensuring that the tool provides the intended responses. As such, a fundamental challenge lies in the upskilling and training of staff members to optimise the utilisation of GenAI. As the technology advances, the workforce must evolve alongside it to effectively harness its capabilities.

“For people to feel comfortable with the technology, you need to know how to use the tool. It’s going to take time to get to that level of comfort and knowledge for it to be widely used in the same way we’re seeing the barriers with digital health solutions. It’s not going to be a simple switch for sites and sponsors to leverage the technology for sure.”

Clinicians and GenAI: A dynamic duo in clinical research

Currently, the integration of generative AI into clinical trials is creating incredible opportunities that will undoubtedly transform the R&D framework in pharma. Beyond its major impact on new targets and new drug identification, it offers efficient solutions for data management, streamlined patient engagement, and reduced site burden, and has the potential to improve overall trial outcomes. However, it is imperative that we approach this technology with a focus on data privacy, compliance, and inclusive practices to avoid further dividing communities.

“There’s a lot of scepticism around AI and GenAI,” says De Carvalho. “With all the buzz around it, I think it’s good to remind ourselves that it is a very new technology that most people haven’t embraced yet, and that it’s still representing a huge transformation for the industry. I think we’re going to see the cases, and the pros and cons, as people continue to leverage the technology. The FDA, EMA, and China’s government have already introduced draft recommendations and we will continue to see the regulators intervene, as we are in a much more regulated, low-risk industry.”

As with any new technology, the key to unlocking the potential of GenAI in clinical research is embracing the transformation responsibly. But, while there are, indeed, still a variety of unknowns when it comes to the future of AI in pharma research, for Carvalho, it’s an exciting time to be working in the sector.

“It’s a fascinating era for us,” she concludes. “The speed at which GenAI integration is taking place is incredible. But we are still in the early days of the technology enhancement curve. I certainly expect that, within the next five to ten years, our traditional RCT model will be completely disrupted, and new data-driven models will have taken over in how we get drugs to the road of approval.”

About the interviewee


Silvia De Carvalho, a seasoned professional with nearly two decades of experience in healthcare marketing, currently serves as the driving force behind AXON’s Clinical Studies Practice. Formerly the senior vice president of activation strategy at Klick Health, she brings a wealth of expertise in digital communication and engagement to her role at AXON.

De Carvalho’s career path began at Sanofi, where she collaborated with clinical development teams to optimise protocols and assess product viability in the diabetes landscape. This early experience instilled in her a profound understanding of aligning clinical development with the genuine needs of patients. Now, as a key leader at AXON, De Carvalho is passionately navigating the Clinical Studies Practice through a dynamic phase of accelerated growth and innovation.

About AXON

AXON logo white

AXON is a global healthcare communications agency that ignites change in healthcare. Through the power of life-changing communications, we contribute to medical advances that improve lives. At the core of every scientific innovation and advancement, there is a simple and compelling story to be told. We know how to tell that story, how to use the right tools to reach the right audiences at the right time, to provoke meaningful change.

We specialise in advising clients on medical affairs, clinical studies, real-world evidence, marketing, advocacy, and communications, all underpinned by insights and creative strategy. AXON has an international reach, with offices in Copenhagen, London, New York, and Toronto, and a worldwide affiliate network. Established in 2002, we have a deep heritage and expertise in healthcare strategy and communication, and we continue to learn and grow in the fast-changing healthcare sphere. AXON is proudly an AVENIR GLOBAL company.

Don’t miss your complimentary subscription to Deep Dive and our newsletter

Sign up



Your name

Your e-mail

Name receiver

E-mail address receiver

Your message