Workstream 2: Safety Monitoring Statistical Methodology
- Who Are We: An ASA task force looking at quantitative methods applied to benefit risk analysis in multiregional clinical trial settings
- Purpose: Aligned with ICH E17 & E9, to provide guidance on benefit risk analysis in multiregional clinical trial settings.
- Background: Wrote book chapter on benefit risk
- Our Role: Review literature and case studies on benefit risk in multi-regional clinical trials to prepare and put together content for the 2022 Deming short course. In parallel, will try to draft a white paper (timing to be determined later)
- Who Are We: An ASA task force looking at quantitative methods applied to safety data in the drug's lifecycle
- Background: Wrote several book chapters on quantitative methods applied to safety
- Our Role: Develop training materials and identify/develop tools for implementing these methods outlined in the chapters
- Connection: We are looking for more participation and input on medical assessment, discernment, and interpretation of these methods in practice
- Who Are We: An ASA task force looking at ML, AI, and DL applied to drug safety data and use cases
- Background: With both clinical trials and healthcare in general evolving with regards to digital and big data regulatory agencies have been actively evaluating the impact of this ongoing development (Han et al (2017), Ly et al (2018), Gupta et al (2018), https://www.fda.gov/about-fda/website-policies/fda-social-media-policy, https://www.ema.europa.eu/en/news/artificial-intelligence-medicine-regulation, http://www.icmra.info/drupal/sites/default/files/2021-08/horizon_scanning_report_artificial_intelligence.pdf). It is quite evident that the use of ML, NLP, and DL in drug safety and pharmacovigilance will continue to experience a rapid uptake that will dictate a need to develop new regulations.
- Our Role: Review the literature and use case of machine learning, artificial intelligence, and deep learning applied to drug safety data. Draft a white paper (timing to be determined later) on the literature review findings
- Connection: We are looking for more participation and input on on both methodology and use of these methods and medical assessment, discernment, and safety interpretation of these methods in practice
- Who Are We: An ASA task force looking at approaches and use of visual analytics of drug safety and benefit risk data in particular those approaches that incorporate one or more quantitative elements
- Background: A continuation of the work from the book (Chapter 15) going beyond clinical trial data
- Our Role: Continue to identify tools, resources, and software for visual analytics of healthcare safety data and creation of training materials and use cases. Develop training materials and identify/develop and share tools for implementing the method and tools via some modality online or manuscript
- Connection: We are looking for more participation and input on medical understanding, assessment, discernment, and interpretation of these visual analytics and benefit risk methods in practice.
- Who Are We: An ASA task force looking at digital health in the context of drug safety data and methodology for analyzing data obtained using digital media for clinical trial and post-marketing data
- Background: Digital technologies continue to transform healthcare. Realizing the full potential of digital health can help accelerate the shift towards patient-centered and outcomes-focused access, sustainable healthcare. Regulatory agencies have also taken note of this development (see for example: https://www.fda.gov/medical-devices/digital-health-center-excellence,)
- Our Role: Review the literature and use case of digital health and methods for analyzing data from digital tools
- Connection: We are looking for more participation and input on medical understanding, assessment, discernment, and interpretation of these visual analytics and benefit risk methods in practice.
- Who Are We: An ASA task force looking at quantitative methods for identifying new safety signals and characterizing and comparing safety data between groups receiving anticancer treatments in oncology studies.
- Background: We are concerned with being able to integrate safety data and use it to describe the safety profile of a single treatment, compare safety endpoints between treatments or groups exposed to the same treatment, evaluate benefit risk, perform safety monitoring, and identify safety signals. The data may include AE data, overall survival (OS) endpoint, patient-reported outcome (PRO) data, laboratory data, data from wearable devices, etc.
- Our Role: Explore methods to accurately summarize and analyze oncology data in clinical study reports, submission documents, labeling, information requests, and publications.
- Connection: We seek collaboration with other ASA SWG taskforces as well as researchers from academia, industry, and regulatory agencies to improve the appropriateness, accuracy, and interpretation of safety statistics for oncology therapies.
- Who Are We: An ASA task force examining how Bayesian Statistics can be applied to quantitative problems in drug safety.
- Background: Safety is less concerned with confirmation and more aligned with learning. This lends itself to Bayesian methods that may incorporate external information and quantify risk.
- Our Role: Explore ways Bayesian Statistics can be applied to understanding and interpreting safety data to quantify risk. We seek to leverage the functionality found in Bayesian analysis to incorporate historical information, estimate many parameters in a multi-level model, and update information as it accrues to understand and safety date for decision making.
- Connection: Bayesian methods can be applied to safety and benefit risk, to causality and real world data, and to post approval marketing.
- Who Are We: We are an ASA task force dedicated to exploring and advancing the application of causal inference methods in drug safety and benefit-risk analysis. Our mission is to enhance the scientific rigor of safety analyses by promoting methodologies that support valid causal conclusions.
- Background: Safety analysis often lacks the rigor of efficacy analysis, failing to address causal questions in drug safety evaluation and benefit-risk assessment. Integrating causal inference methods into safety and benefit-risk assessments is essential for enhancing ongoing safety monitoring, improving accurate risk assessment, strengthening robust regulatory submissions, and providing clearer insights into the true effects of drug interventions.
- Our Role: We aim to review and synthesize existing applications of causal inference in drug safety and benefit-risk analysis; educate and disseminate by developing accessible materials such as blog posts, website content and presentations to increase awareness and promote best practices; innovate and advance methodologies tailored to the specific challenges of safety analysis, improving the reliability and interpretability of findings.
- Connection: We seek collaboration with other ASA taskforces, as well as researchers from academia, industry and regulatory agencies. By aligning efforts, we aim to synergize efforts for improving drug safety and benefit-risk assessment.
- Who Are We: An ASA task force looking at the estimand framework for safety in low mortality trials, and statistical considerations for integrated safety analyses across multiple clinical trials.
- Background: Wrote 2 articles on the topic; Safety estimand in late-phase studies with lower mortality (Clin Trials 2024), and Estimand framework and statistical considerations for integrated analyses (TIRS 2024) using signal detection of common AEs as case.
- Our Role: We identify principles for integrating safety data across multiple studies, to avoid unnecessary bias and to ensure proper comparative analysis.
- Connection: We strive to stress-test our new strategies in real drug projects, and reach out across industry, to make our new strategies well known and applicable and get feedback for further improvement. We seek to align with other TFs in WS2 (e.g., TF Oncology Drug Safety).