PS23: Advancing the Implementation of Safety Methodologies in the Digital Era: An EFSPI/PSI Benefit Risk SIG Safety WG Outreach

This session presents the initial outputs from a dedicated working group, established last year within the EFSPI/PSI Benefit-Risk SIG, with a core focus on advancing the implementation of safety methodologies. The EFSPI/PSI sponsored WG has transatlantic presence with 15+ members across industry, regulatory and academia with established correspondence with ASA Safety WG and PHUSE Safety WG. Recognizing the persistent challenges in translating innovative safety science into routine practice, this session brings together three distinct but complementary perspectives on enhancing drug safety across the pharmaceutical development lifecycle. Firstly, we will explore the broader strategic and practical hurdles in accelerating the adoption of quantitative drug safety inventions into impactful innovations, examining regulatory, organizational, and technical factors. Following this, we'll delve into the statistical and operational aspects of aggregate safety reporting, particularly for anticipated serious adverse events in investigational trials, offering practical strategies for regulatory compliance and trial integrity. Finally, we will provide an update on the working group's efforts to develop a practical 'Toolkit' for Integrated Summary of Safety (ISS) development, aiming to standardize best practices for pooling, grouping, and adjusting safety data. Three talks are planned: 1. Quantitative Evolution of Drug Safety: Adapting Inventions to Innovations Ensuring drug safety is a cornerstone of pharmaceutical development, yet the adoption of advanced analytical tools and methods often lags behind scientific invention by years. This work examines the strategic and practical challenges of transforming innovative safety approaches into widely adopted solutions. We review the current regulatory landscape (FDA, EMA) and highlight limitations in standard safety analyses, such as reliance on incidence rates and aggregate reporting. Our strategic analysis explores the value of advanced statistical techniques, the importance of regulatory acceptance, and the organizational factors-such as infrastructure, training, and stakeholder engagement-required for successful implementation. We emphasize the need for open-source tools and collaborative development to accelerate uptake. To ground these concepts, we present a series of real-world case studies, including Bayesian analytics, systematic safety surveillance, and interactive safety graphics. These examples illustrate both the opportunities and barriers encountered during implementation, offering insights into what enables inventions to become impactful innovations. We conclude by identifying persistent gaps and proposing strategies to bridge them, aiming to foster a culture of innovation in drug safety. Our findings provide actionable guidance for researchers, industry leaders, and regulators seeking to translate scientific advances into improved patient outcomes. 2. Beyond Individual Cases: Operationalizing FDA IND Safety Reporting for Anticipated Serious Adverse Events The U.S. Food and Drug Administration (FDA) mandates rigorous safety surveillance and expedited reporting of suspected unexpected, serious adverse reactions (SUSARs) in investigational drug trials. However, operationalizing these requirements - especially for anticipated serious adverse events (SAEs) - remains challenging due to unclear guidance and numerous potential implementation approaches while balancing regulatory compliance with trial integrity. This talk presents practical strategies for statisticians to implement these requirements, with a focus on aggregate safety reporting of anticipated SAEs. Anticipated SAEs - events expected due to underlying disease, study population characteristics, or background therapies - require unblinded analysis to detect clinically meaningful imbalances across treatment arms during trial conduct. We explore two operational models for conducting unblinded safety assessments: outsourcing to external contract research organizations (CROs) and establishing internal, firewalled independent Statistical Analysis Teams (iSATs). Each model is evaluated for its governance structure, data handling protocols, and operational efficiency. Key considerations include regulatory expectations, including the use of Safety Surveillance Plans (SSPs), blinded trigger methodologies, and the maintenance of trial integrity during unblinded reviews. We discuss trigger-based and periodic unblinding strategies, highlighting their advantages and challenges, and provide recommendations for best practices in planning, execution and documentation. Practical examples illustrate common challenges such as data transfer logistics, validation of unblinded outputs, and maintaining independence within sponsor organizations and discuss mitigation strategies. Recommendations are provided to help sponsors select the most appropriate model based on organizational capacity, regulatory risk, and trial complexity. By translating regulatory guidance into actionable steps, this session aims to equip clinical trial sponsors with actionable insights to enhance regulatory compliance, protect patient safety, and uphold scientific rigor in the context of FDA IND safety reporting. 3. Pooling, Grouping, Adjusting: A Statistician's Toolkit for Integrated Summary of Safety development The Operationalizing ISS reporting sub-team was formed to recommend best practices for safety data analysis in Integrated Summary of Safety (ISS) projects. Noting variability in company approaches, the team seeks to develop standardized recommendations for methods and data presentation and potentially provide code to facilitate analysis. The team has identified topics of interest for ISS development (e.g. exposure adjustment, study-size adjustment, grouping of terms, pooling strategy) and collated guidelines, literature, and examples for each topic. While considerable published information is available on these topics, clear cohesive recommendations are not readily available for statisticians and clinicians working on an ISS. The team is therefore developing a toolkit for ISS development. For each topic, a summary will be available summarizing literature and guidance, providing best practice recommendations along with examples or recommendations for the ISS SAP text and shells. Code to produce the recommended analyses and displays may also be provided. This talk will provide a status update of this work to date.