PiVily Quarterly Report
Governance became visible in the tools
Common AI principles, a unified FDA adverse-event platform and more actionable real-world data quality guidance made governance concrete.
The quarter in one reading
Q1 2026 connected high-level principles to the systems and evidence environments in which safety decisions are made. The quarter rewarded organisations that could make accountability visible at the point of use.
This retrospective edition was reconstructed from official materials and completed on 11 August 2026. It describes the quarter; it does not imply contemporaneous publication.
Material signals
What moved—and why it mattered.
EMA and FDA established a shared direction for good AI practice.
In January 2026, EMA and FDA published ten common principles for AI use in evidence generation and monitoring across the medicines lifecycle, including safety monitoring.
EMA / FDA: Common principles for AI in medicine development EMA / HMA: Artificial intelligence in medicines regulation
Principles such as clear context of use, multidisciplinary expertise, data governance, performance assessment and lifecycle management move the discussion beyond a generic 'human in the loop.'
Can the organisation name the qualified decision-maker, intervention point, evidence and override authority for each material use?
FDA began consolidating adverse-event access through AEMS.
FDA launched the Adverse Event Monitoring System in March 2026, beginning a transition from multiple legacy systems toward a unified platform, more timely publication, APIs and broader analytics.
Better access expands surveillance and public scrutiny. It does not convert spontaneous reports into incidence estimates or causal evidence.
Are scientific and communications teams prepared to explain what more accessible safety data can—and cannot—support?
Real-world data quality guidance became more actionable.
EMA published the application of its medicines-regulation Data Quality Framework to real-world data in March 2026, including recommendations on relevance, reliability, systems and processes, and data-quality metrics.
Data quality is not a generic property of a database. It must be assessed in relation to a research question and the processes that produced the data.
Does the evidence record connect the research question to data relevance, reliability and the controls underpinning the source?
EU guidance sharpened expectations for pregnancy and breastfeeding.
EMA's final GVP considerations for pregnant and breastfeeding women and children exposed in utero or via breastmilk became legally effective in February 2026.
Population-specific PV requires connected evidence plans, follow-up, analysis, communication and risk management—not an isolated data-collection activity.
Are population-specific responsibilities integrated across case management, aggregate evaluation, studies, risk management and communication?
Next-quarter watch
Questions to carry forward.
- How common AI principles translate into use-case controls
- The operational and interpretive consequences of the AEMS transition
- Whether RWD quality assessments become consistent across functions and partners
Method & source register
Trace the reading back to the record.
PiVily selects developments for system-level relevance to human pharmacovigilance. Facts are attributed to official publishers; interpretation is editorial. This is not an exhaustive jurisdictional change log or organisation-specific advice.