AI – Championing Clinical Trials

In recognition of AI being the future of clinical research and medical innovation, Parexel, a clinical research organisation recently moved into a multi-year-long partnership with Palantir, a leading builder of AI systems. Parexel intends to leverage Palantir’s Artificial intelligence platform (AIP) to its advantage to accelerate safe and effective clinical trials for their biopharmaceutical clients. Sanjay Vyas, Executive Vice President and Managing Director, Parexel India talks about the the strategic partnership with Palantir to enhance the operational efficiency of clinical trials.

By Sonali Patranabish

Parexel already has an in-house AI-powered lab, how will this strategic partnership with Palantir further enhance the operational efficiency of clinical trials at Parexel?

Parexel has set generative AI (GenAI) at the heart of our AI strategy to deliver operational gains and we will differentiate our capabilities through the way we leverage this opportunity.

Integrating Palantir’s powerful Artificial Intelligence Platform (AIP) will enable Parexel to apply advanced AI and machine learning capabilities directly to our clinical data, enabling automation of many manual processes and extraction of deep insights from historical trial data to streamline operations and improve data quality. Most importantly, Palantir’s advanced tooling allows Parexel to move at an unprecedented pace —meaning more solutions are delivered more rapidly, enabling us to accelerate value delivery for our business and customers.

In addition, Palantir’s Foundry platform provides real-time access to validated, high-quality data for our ongoing studies. This transparency allows Parexel’s stakeholders to make more informed decisions, respond quickly to emerging trends and improve trial efficiency.

Overall, Palantir brings industry-leading expertise in data governance and AI capabilities that complement Parexel’s existing AI-powered platforms. Coupling Palantir’s platform capability our existing AI Labs delivery model provides Parexel with a significant and unique AI capability in the industry.

Shifting gears from becoming the largest global CRO to a fully digital CRO, What does this mean in terms of capacity building for Parexel?

Parexel is investing in several ways to become a fully digital CRO:

  • Building, developing and integrating cutting-edge digital platforms such as clinical data management systems, electronic data capture (EDC) tools and secure data-sharing platforms.
  • Further strengthening the foundations of AI across our business through upskilling training programs on digital technologies, data management, and digital clinical operations so we can build stronger data science teams with expertise in artificial intelligence (AI), machine learning (ML), and advanced analytics. This will better position us to drive data-driven decision-making and leverage the power of AI in clinical trials.
  • Maintaining and enhancing already robust cybersecurity measures and data privacy protocols to ensure the security and integrity of sensitive clinical trial data in a digital environment. This will be an area of continuing investment.

Modern clinical research and trials are hinged on patient-centricity, how will leveraging AI help revolutionise clinical trials and in turn better clinical outcomes for patients?

AI can help address one of the more important challenges faced by every organisation engaged in clinical trial conduct – patient recruitment and retention. Advanced algorithms can assist in identifying the most suitable patient populations, optimizing trial protocols, and ensuring efficient patient recruitment and retention strategies. By analysing patients’ experiences, preferences and behaviour while participating in clinical trials, we can also make trials more accessible for future participants. One example is by making clinical trials easier to join and less impactful in their daily lives.

Sophisticated machine learning models can process vast amounts of data from diverse sources, including electronic health records, wearable devices, and patient-reported outcomes.

This capability helps detect previously undetected patterns, identifies potential safety signals, and makes data-driven decisions in real-time, ultimately enhancing patient safety and overall clinical research quality.

Over time, we will also see virtual assistants and chatbots using GenAI to provide personalised support and guidance to patients throughout the trial process. However, there is a known tendency of GenAI to hallucinate or present incorrect information as fact, which makes this vision for the future a challenge. As the technology improves, the potential of this technology to enhance patient adherence and reduce dropout rates is clear.

The current landscape of clinical trials has been seeing a lack of translation from a research setting to a clinical setting, i.e there has been a lag in translation from a trial world to the real world. How can AI address this issue?

Clinical trials are typically conducted under highly controlled conditions with strict inclusion and exclusion criteria, which may not fully represent the heterogeneity of patient populations encountered in clinical practice. New treatments are also typically licensed based on average treatment effects. This means in practice that they do not work for everyone when

used in a real-world setting. Understanding why that is, is an area where AI can unquestionably help us. By integrating data from multiple sources, including clinical trials, observational studies, and real-world evidence, AI-driven analysis can help us find previously hidden patterns and relationships that ultimately lead to new treatments for those in whom already available treatments have failed or have been sub-optimal.

Gen AI is a force to reckon with, but comes with its set of challenges. What sort of challenges do you foresee and do you have a master plan in place for troubleshooting them?

Perhaps most critical, given the potential for GenAI to hallucinate, is the need to ensure we always have a ‘human-in-the-loop’ to check, correct where needed, and be accountable for the final outputs of GenAI-based systems. GenAI models hold great promise for clinical trials, but we need to use it with care. 

As GenAI is trained on vast amounts of data, which may include sensitive or proprietary information, it is important to establish rigorous protocols to ensure the protection of patient privacy, intellectual property rights, and compliance with data protection regulations. Related to this is the potential for bias. AI models replicate the biases from the information they are trained on, which significantly leads to providing inaccurate or unfair decisions. It’s important to establish comprehensive bias mitigation strategies. Fine-tuning on diverse and representative data sources may be helpful as will rigorous model testing, and ongoing monitoring for fairness and accountability.

Informed by these and other challenges, Parexel recently published its 6 Principles for Artificial Intelligence. This sets out a rigorous framework that requires: 

  • Thoughtful design & deployment
  • Accountability & Senior Level Governance;
  • Human Oversight & Control;
  • Transparency;
  • Regulatory Conformance & Legal Compliance; and
  • AI-informed Security & Privacy. 

We have also created a robust governance process, including representation up to C-suite level, to set strategic direction and oversee solution delivery. Finally, Parexel is also investing in training for all our colleagues and creating resources to support their engagement with the tools we’re delivering.

The evolving space of clinical research and trials has been seeing disruptive trends and a host of technological advancements. How do you perceive this space shaping up shortly?

Trying to predict the future is usually an invitation to be wrong. But we can be confident that technology, and AI in particular, will be very impactful to the future of clinical research. Already existing trends will likely continue and strengthen – including the use of technologies to simplify and unburden the experience of participating in clinical trials. For example, telehealth to reduce the number of clinic visits; health monitoring wearables to enable passive data capture around the clock; investigational products, blood draws and much more to be conducted in the patient’s home.

These advances will also allow for more geographically diverse and inclusive trials. This means researchers can tap into a wider pool of potential participants, leading to more generalisable results.

Additionally, AI-powered drug discovery and development platforms are accelerating the identification of promising therapeutic candidates, potentially reducing the time and cost associated with traditional drug development processes. Alongside this, with the help of big data and AI analysis, researchers will be able to design trials that target specific patient subgroups so there are more effective treatments with fewer side effects, as therapies can be tailored to individual needs.


Support us in our endeavor to bring you Advertisement free content.
Choose your options to donate or subscribe.


Related Articles

Back to top button