AI has proven its worth as a research tool and now it’s time to tap its potential for clinical trials, argues Tero Laulajainen. That places more, rather than less, onus on the human element.
In the ever-evolving landscape of pharmaceutical research and development, artificial intelligence not only acts as a powerful tool, but also as an innovation catalyst in drug discovery and development. The advancements made to date have been remarkable and it stands to play a pivotal role in our ambition to develop more precise and effective, better tolerated and better tailored approaches to an individual patients’ profile and health characteristics.
Examples of how AI has changed the healthcare landscape include:
Identifying novel drugs: AI facilitates the identification of novel drug targets by analysing complex biological datasets and uncovering new avenues for research into more complex disease areas.
Improved success rates: The drug development process is notoriously risky, with many potential drugs failing in clinical trials. AI’s predictive models could mitigate risk by identifying likely outcomes and risk/benefit ratios based on existing data.
Patient recruitment and screening: AI has enabled rapid screening of vast databases to identify investigator sites with access to right patient populations and to identify suitable participants for clinical trials. This makes patient recruitment more efficient and accurate.
Predicting drug effects: AI can be used to predict the effects of drugs on the human body, including potential side effects. This can help researchers to identify and mitigate safety issues earlier in the drug development process and predict what is the likely benefit/risk ratio for a drug in development.
Although AI has significantly impacted numerous facets of research and discovery, its influence on randomised clinical trials (RCTs) has been relatively minimal. So far, the integration of AI into clinical development has focused on improving operations and speeding up processes, but as AI continues to expand its capabilities, we can expect to see a significant shift in the way clinical trials are conducted. Leveraging these advancements should see us design more efficient, more precise trials with targeted recruitment strategies, leading to better patient engagement/retention and greater success rates.
The need for trust
There are major benefits of the use of AI in clinical trial development. For example, using predictive AI models, analytical tools and realworld data has accelerated our understanding of disease. It also allows us to find suitable patients and key investigators, quickly and efficiently. However, as we incorporate new technologies like AI, we face several challenges:
Trust: A major obstacle to integrating AI among patients and healthcare professionals. They often struggle to understand how AI makes complex decisions and assume there is no human input. The key is that AI does not make those decisions, humans do, and they do so with the help of AI that can analyse large and complex sets of data.
Data quality and quantity: AI algorithms rely heavily on high-quality, diverse datasets for training and validation. In the context of clinical trials, ensuring the availability of representative data poses a significant challenge in rare disease populations where patient population numbers are low.
Generalisability and transferability: AI models trained on one dataset or patient population may not generalise well to diverse real-world scenarios or external validation cohorts, therefore an increased level of validation and calibration across different real-world scenarios is required. This can slow down dataset processing and increase the cost needed for these validations across different patient populations.
Interpretability: The nature of the current AI model relies heavily on an in-depth brief and understanding of the task before it can develop the output. Therefore, the quality of the input directly affects the quality of the outcome. If the input is of poor quality, the output will be poor.
When navigating the complexities of AI, a ‘need’ gap emerges, the need for human input and expertise. Human input is crucial in understanding the unique needs and perspectives of diverse populations, fostering trust, and addressing human nuances that AI may not have the capabilities to address.
While AI algorithms bring quicker and faster processes, human clinicians and researchers bring invaluable clinical judgment and contextual knowledge to navigate informed decisions. Therefore, there is a need for the human touch right from the beginning of a clinical trial. Establishing collaborative foundations between humans and AI from the outset ensures patients and peers feel atease, knowing that the human touch remains integrated within sensitive processes, such as clinical trials.
The importance of collaboration: Getting ahead of the curve I strongly believe a collaborative approach between individuals and organisations is also a key to success and getting ahead of the curve. It is not about individuals or individual companies anymore, it is about collaboration and building partnerships with experts in their field. At UCB, we cannot solve the world’s health problems alone; we must drive innovation with collaboration and choose suitable partnerships to ensure we address the unmet needs quicker and faster.
For example, we have partnered with technology giants such as Microsoft, to develop an inter-disciplinary approach across various teams and cross-industry expertise. This approach not only fosters innovation but allows us to understand various qualities from a different lens, bringing fresh insights and diverse skill sets to our perspective. This partnership aims to create a holistic, datadriven perspective on patient populations, accelerating the discovery and development of medicines for individuals with severe diseases.
Although AI has significantly impacted numerous facets of research and discovery, its influence on randomised clinical trials (RCTs) has been relatively minimal… but as AI continues to expand its capabilities, we can expect to see a significant shift in the way clinical trials are conducted
The collaboration underscores the potential of AI technology to synergise with scientists and data specialists, revealing essential correlations and patterns that drive innovation. By fostering interdisciplinary collaboration between our data scientists, clinicians, regulatory experts and patients, we can leverage our understanding of clinical trials and development from all angles and enhance the efficiency and precision of clinical research for unmet needs in complex health conditions.
The future of AI in clinical development
Looking ahead, I believe we can bolster the use of AI in the pharmaceutical industry and develop pioneering work through collaborations and partnerships. Beyond expanding our clinical research and development, being able to predict the effects of drugs on the human body, including potential side effects. This can help researchers to identify and mitigate risks earlier in the drug development process and drive decision-making forward.
Again, we can’t do this alone, partnerships and collaborations between human expertise and AI is integral to developing groundbreaking research in the new landscape of clinical research. In my vision of the future, AI seamlessly merges with human expertise, propelling the pharmaceutical landscape towards efficiency, personalisation and patient-centricity, dynamically creating an environment where innovation can thrive.
Tero Laulajainen is vice president – head of global clinical science and operations at UCB
