Feasibility High Impact Low to medium Time horizon Medium (2–4 years) Barriers Integrating and cleaning data from multiple sources and in multiple formats, while ensuring data security and privacy. Complex data analysis is not fast enough to inform real-time decision-making. Statistical programming for custom analyses is complex, time-consuming and error-prone. Biggest unlock Automated systems for collecting, cleaning and processing sensitive health data. Automated code generation tools that integrate with existing statistical analysis workflows and software. Data analysis for clinical trials today Given the increasingly sophisticated methods for collecting trial data, development teams face growing challenges related to data quality and standardization. Data analysis relies on time- intensive processes and analytical methods that, while not entirely manual, still demand substantial time and specialized expertise. Data analysts typically work with fragmented data sources, including EHRs, lab results and patient-reported outcomes, which are often not easily integrated. This necessitates extensive data cleaning, transformation and validation before meaningful analysis can begin. Coding tasks, such as writing scripts for data processing and statistical analysis, are also partly manual, making them time- consuming, costly and prone to human error. Data analysis for clinical trials enhanced with Gen AI Gen AI can serve as a co-pilot in data analysis by automating routine tasks, such as writing code and generating tables for submission. Gen AI can also help unlock insights throughout the development life cycle. Co-pilots can integrate data from disparate sources – patient engagement data from chatbots and mobile apps as well as sensor data from wearables – clean it and prepare it for analysis significantly faster than programmers using predictive AI methods, reducing the time between data collection and informed decision-making.Evolution, not revolution: Data analysis 4 Global pharmaceutical company Eisai recently teamed up with Medidata to implement Medidata Clinical Data Studio,17 an AI-powered platform designed to optimize the management and analysis of clinical trial data. By integrating multiple data sources, both internal and external, Clinical Data Studio breaks down traditional data silos and accelerates data review by up to 80%. This enhanced data control enables Eisai to scale the complexity of its clinical trials while maintaining high data quality and integrity. With streamlined data import and automatic validation capabilities, the platform offers a comprehensive, real-time view of patient data, enabling faster, more accurate decision-making. This innovative approach helps pharmaceutical companies such as Eisai execute complex clinical trials more efficiently.USE CASE Eisai and MedidataIn terms of statistical analysis, AI can help optimize clinical trial designs and statistical power – for instance, in randomized clinical trials – by using covariate adjustments and other techniques to reduce noise and improve the design without altering the trial itself. Eric Durand, Chief Data Science Officer, Owkin Intelligent Clinical Trials: Using Generative AI to Fast-Track Therapeutic Innovations 10

Intelligent Clinical Trials 2024