Accelerating Regulatory Communication with Natural Language Processing

Rebecca Vislay-Wade, Principal Data Scientist, Moderna

Rebecca Vislay-Wade, PhD is a Principal Data Scientist at Moderna, where she leads a team of scientists developing AI applications for clinical operations, regulatory science, and pharmacovigilance. Prior to Moderna, she worked as Senior Research Data Scientist at Highmark Health. Rebecca holds a PhD in biochemistry from Harvard University and did postdoctoral work in neuroscience at the NIH and Children’s National Medical Center in Washington, DC. She currently lives in the Boston area with her family.

Abstract:

Advanced AI and machine learning played a crucial role in Moderna's rapid development of one of the first mRNA COVID vaccines during the peak of the pandemic. As a pioneer in the field of "Digital Biotech," Moderna is committed to applying these advanced techniques across various aspects of our operations, spanning from pre-clinical research to clinical trial management and regulatory affairs. In this presentation, I will delve into a specific solution designed for Global Regulatory Sciences, known as "RegBot." RegBot harnesses the power of natural language processing (NLP) to streamline our interactions with health authority organizations. Since its deployment in September 2022, RegBot has successfully processed over 1,000 emails from health authorities, archived more than 5,000 health authority questions (HAQs), and significantly reduced the manual effort required to handle HAQs by more than sixfold. A recent significant update introduced a similarity search feature, enabling us to retrieve the most relevant HAQs that Moderna has previously addressed. This represents the first step in assisting our teams with HAQ response research and drafting. Our future plans include integrating generative models to further alleviate the burden of composing responses to HAQs for our regulatory and subject matter experts, thereby expediting our regulatory correspondence. By leveraging AI to enhance the efficiency of regulatory communications, we can accelerate the delivery of potent mRNA medicines to patients.

Data Science and Gen AI: Transforming the Paradigm of Medical and Safety Data Review practices

Sunil Talathi, Senior Director, Data Enginering Analytics, BeiGene

Sunil Talathi is an Inclusive Leader, technologist, and Innovator with extensive experience in Data Science, Analytics, Clinical Systems Operations, Risk-Based Monitoring (RBM) and Clinical data management across top pharma/biotech companies. He is a problem-solver and is passionate about automation, innovations, and building technological solutions. As a Director of data Analytics Engineering at Beigene, he is responsible for building fully autonomous NextGen Analytical & digital products powered by AI and he strongly believes AI technologies, Automation systems, Digital technology, and Analytics will be 4 pillars for the Organization to be much more effective, innovative and nimble.

Abstract: The integration of AI in Data Analytics Engineering is pivotal to the success of regulatory submissions for clinical trial data. These engineers have traditionally managed a spectrum of tasks, including the initiation of data sources, the collection of data, data profiling, the guarantee of data integrity, and the performance of analytics. Such functions are interconnected, often manual, repetitive, labor-intensive, and subject to stringent time constraints. The realm of Medical and Safety Data Reviews is under immense pressure to evolve as the industry shifts towards more sophisticated adaptive clinical trial frameworks, individualized treatments, and Decentralized Clinical Trials (DCTs) that employ a range of digital tools such as smartphones, wearable technology, sensors, and telehealth services. Simultaneously, the constant influx of data and the escalating requirement for detailed analytics are placing a growing strain on resources that are already expected to deliver results more rapidly than ever.

This talk will elucidate how Data Science is being applied to refine the data review process within the context of clinical trials, addressing the challenges by promoting standardization, enhancing efficiency, and fostering digital transformation. We will also illustrate how GenAI, a groundbreaking automated technology, is instrumental in decreasing review cycle times and scaling up productivity.

Biostats et al: Building an Interdisciplinary Platform for Aggregate Safety Assessment

Glen Colopy, Head of Data Science, Wildfell Software

Glen Wright Colopy is the Head of Data Science & Statistics at Wildfell, a startup specializing in custom software and data science solutions for the biotech and life science industries. Glen's day-to-day work is as a hands-on full stack data science contributor, focused on helping companies (typically startups) succeed in their data science & AI initiatives. His focus lies in statistical machine learning, with a particular interest in healthcare and life sciences. Glen holds a Doctorate in Engineering Science from the University of Oxford. His work includes two patents in vital sign monitoring and 16 publications, most of which are centered around medical machine learning.

Glen is an active contributor to the Statistical Learning & Data Science (SLDS) Section of the American Statistical Association. He also hosts the Data & Science podcast and serves as a section editor for the Journal of Data Science. His dedication and contributions to the field were recognized with the SLDS 2023 Outstanding Service Award.

Abstract:

The science of safety monitoring differs significantly from the more familiar and established approaches to efficacy assessment. Namely, safety monitoring cannot rely solely on pre-specified statistical hypothesis testing to deliver sound and timely safety inference. Instead, safety monitoring grapples with open-ended questions and subjective evaluations. It integrates diverse domain experts across Biostatistics, Clinical, Regulatory, Data Science, alongside FDA and CROs. These different groups must not only bring the best of their own knowledge to bear on the problem but also integrate this knowledge with the other domains.

To achieve this, we propose an Aggregate Safety Assessment Platform (ASAP). ASAP’s primary objective is to facilitate normal business operations within individual ecosystems while building interoperability between them. This approach is akin to laying a foundation, where the end structure (whether a modest cabin versus a towering skyscraper) strongly informs the initial requirements. Following this analogy, we focus on building a foundational platform which can ultimately support both (i) the specific needs of each business unit while (ii) allowing collaboration.

This platform is a step towards a cohesive, interoperable environment that respects each function's expertise, ultimately contributing to safer medications and improved patient outcomes. We will show the initial steps of how biostatistics and data science's needs can be met while facilitating collaboration.