Stakeholders of RWE (including regulatory agencies, HTAs, payors, providers, patients) are increasing their expectations on the rigor we put into generating meaningful evidence from RWD. Therefore, we need to increase our efforts around demonstrating the data we select are truly fit for the "purposes" of the evidence we are generating.

The increasing availability of real-world data (RWD), including electronic health records, claims data, patient registries, and digital health sources, has broadened the applications for causal inference in biomedical and public health research. Artificial intelligence (AI) and machine learning (ML) methods are increasingly adopted in these areas, providing approaches for high-dimensional data integration, feature selection, bias reduction, and modeling treatment-outcome relationships. Current research topics include deep learning and ensemble methods for confounding adjustment and counterfactual prediction; combining heterogeneous RWD sources to enhance generalizability; developing interpretable ML techniques for transparent causal analysis; and using AI-driven simulations to evaluate causal assumptions. Issues being addressed include data quality, algorithmic bias, reproducibility, and the relationship between ML-based causal inference and conventional epidemiologic frameworks. The presentation will examine these developments and emphasize the importance of multidisciplinary collaboration among industry, academia, and regulatory agencies.

This presentation discusses the calibration of real-world data (RWD) trials and the application of External Control Arms (ECA). It highlights methods to improve trial accuracy, reduce costs, and enhance regulatory acceptance by effectively integrating RWD and ECA for robust clinical evidence.

RCTs provide critical evidence of the efficacy of medications but often leave questions unanswered about their effectiveness in diverse patient populations not included in trials or different outcomes. While additional trials answering these questions would be preferred, they may not be available for various reasons or will only become available with delay. In such cases, RWE can offer timely, cost-effective insights.

Real-world data is vast and expanding, but fragmented. By applying AI techniques, we can bridge gaps, reduce complexity, and create actionable real-world evidence more efficiently. Drawing on a pharmaceutical RWD/RWE program, this talk covers implementation of AI tools across key workflows: programming assistants (SQL, R/SAS/Python), protocol and report authoring, and systematic literature review. The session also describes the design of fit‑for‑purpose knowledge bases to ground LLMs in organizational standards.

Novo Nordisk is building cross-country collaborations with healthcare providers to maximize the value of real-world data. This session will highlight case studies from the US, UK, Denmark, and Sweden, showing how Pharma–HCP partnerships advance evidence generation, enable broader patient insights, and address practical challenges in governance, data access, and alignment, delivering scalable models for global research.

GSK has developed innovative collaborations with healthcare providers in the US and UK to unlock the value of real-world data. This session will share case studies demonstrating how Pharma–HCP partnerships enhance evidence generation, streamline data access, and address key challenges—highlighting lessons learned, mutual benefits, and emerging solutions for advancing research beyond traditional trials.

Pragmatic trials have arguably been characterized more by promise than actual progress. New developments are poised to help shift the balance. The use of pragmatic elements in clinical trials—including electronic health records, patient reported outcomes, and wearables—has been a popular topic of conversation, but also a source of great hesitation. While these options create new opportunities for regulatory decision-making, they also present challenges around sponsor oversight and selective safety data reporting. To address these challenges, TransCelerate, in collaboration with FDA’s CDER Center for Clinical Trial Innovation, conducted tabletop exercises to test trial designs under real-world conditions. This session will share key insights and approaches to support more effective design and execution of pragmatic trials. This session will also touch on a new TransCelerate effort to shed light on the use of pragmatic designs to improve safety data collection in connection with ICH E19.

As patients seek deeper engagement with healthcare, RWE is reshaping how data informs drug development and patient experience. Since June 2022, Walgreens Clinical Trials has led innovation by using pharmacy-rooted registries to drive evidence generation and patient recruitment. This has enabled solutions like in-pharmacy biospecimen collection and helped address key research challenges through its extensive pharmacy network.

As real-world evidence becomes central to clinical and regulatory strategies, too much reliance remains on fragmented, decontextualized data. This session will share how Academic Medical Centers (AMCs), through integrated research platforms like MGB HIRO, provide traceable, audit-ready, clinically rich data at the point of care. Attendees will learn why AMCs are essential partners in generating credible evidence—and how to unlock speed, quality, and trust across the R&D lifecycle.

This session presents a proven case-based training approach to build internal capacity and literacy in real-world evidence (RWE) generation and interpretation. Using pragmatic case studies covering patient-reported outcomes, comparative effectiveness, registries, and post-marketing surveillance, this scalable model equips attendees with foundational and intermediate RWE skills. Designed for cross-functional teams, the session will foster collaboration and better decision-making using RWD across the product lifecycle.

Real-world evidence (RWE) has been increasingly used to answer scientific and regulatory questions. The Prescription Drug User Fee Act (PDUFA) VII for fiscal years 2023-27 mandated the U.S. Food and Drug Administration (FDA) to develop negative control methodology to support causal inference when RWE are used for studying safety and effectiveness. In this talk, we review the current landscape and use cases of using negative controls to inform regulatory decision-making.

Unmeasured confounding is a major threat to the validity of real-world evidence. Methods via negative control have been proposed to detect and, more importantly, adjust for unmeasured confounders. In this talk, we review the latest methodology development, including their advantages, limitations, and recommend best practices in leveraging negative control during the design, analysis of real-world studies.

Too often, clinical trial protocols are held together by a patchwork of disconnected platforms and manual workarounds, a system built on repeated data entry, mismatched formats, and limited data consistency. It’s inefficient, error-prone, and a major barrier to accelerating study start-up. This session will explore how biopharmaceutical companies are changing that by digitalizing protocol design and authoring. Using an open-source, vendor-agnostic data model, organizations are automating data flow, improving interoperability across systems, and gaining insight to inform better study design. Join us to see how this approach can unlock new use cases from registry submission to statistical analysis; automate data flow and interoperability between systems; reduce study start-up effort, cycle time, and complexity; improve quality and compliance​.

At Amgen, we are transforming our analytics ecosystem with AI-driven tools that enable near real-time, self-service insights for Global Development stakeholders. Our next-generation analytics engine features a rebuilt metrics platform, centralized data governance, and a streamlined dashboard landscape leveraging Power BI with Co-Pilot and custom AI interfaces. By implementing AI/ML bots and natural language queries, we are shifting from static reporting to a dynamic, AI-driven experience.