AI in Pharmacovigilance: Five Decisions Before Go-Live
Decide where rules, AI or people belong in a PV process and record why
Know what GVP, the EU AI Act and draft Annex 22 add to an AI use case in PV
Validate an AI tool against its intended use with a reference set and real numbers
Design reviews that catch AI errors and keep the tool under control after go-live
Pharma
Online
Diese Themen erwarten Sie
Programmpunkte
09:00 Opening: one use case per participant
Where each organisation stands with AI in PV
Choosing the use case each participant carries through the day
09:20 Decision 1 – Does this step need AI at all?
Rule-based automation, machine learning and generative AI compared
Mapping the process and its handovers before automating anything
Readiness check: decision owner, criteria, reference data, data use
10:30 – 10:45 Break
10:45 Decision 2 – Which rules apply?
Why MAH responsibility and QPPV oversight stay with people
GVP, the EMA reflection paper on AI and the EU AI Act, including AI literacy
Draft Annex 22 and the revised Annex 11: what can be applied now
Data protection when safety data reaches cloud services
11:45 Decision 3 – How do we know it works?
From a vague wish to a testable intended use
The reference set: recall, precision and performance in subgroups
Release, restrict or reject: a worked decision record
12:30 – 13:15 Lunch break
13:15 Decision 4 – Who checks the AI, and how?
Automation bias in experienced reviewers
Error types reviewers must recognise, from invented data to false certainty
Review depth by impact and an audit trail of what the reviewer changed
Seeded errors as a test of the review step
14:15 Workshop – Score your use case
Participants rate their own use case on benefit and PV risk
Decision: go, go with conditions, or not now, with one reason and one next step
15:00 – 15:15 Break
15:15 Decision 5 – How do we run it?
Monitoring, change control for prompts and model versions, fallback
Vendor claims, contract clauses and the questions an inspector will ask
Short demonstration on synthetic data: an AI-drafted narrative and its review
16:15 Takeaways: the next step for your use case
16:30 End of the seminar
Details
About this Workshop
More and more PV teams test AI for case intake, literature screening or narrative writing. Responsibility for every decision stays with the marketing authorisation holder and the QPPV, whatever the tool does. What has to change is how that responsibility is designed into the process and documented. This seminar is for PV professionals who decide whether, where and how AI enters regulated PV work. The day is built around five decisions every AI project in PV has to make before go-live: whether a step needs AI at all, which rules apply, how performance is proven, how human review is designed so that it catches errors, and how the tool is run and changed afterwards. Each participant brings one use case and takes it through all five decisions. A workshop scores these use cases against benefit and PV risk, and a short demonstration on synthetic data shows an AI-drafted narrative and the review it needs. The seminar is led by a practitioner who holds the EU QPPV role and works with automation and AI in PV. It is held online in English. Participants take home a readiness check, a scoring scheme for use cases and a template for the release decision.
Choose the right tool
Tell rule-based automation from machine learning and generative AI, and choose the simplest approach that is adequate for each single step of a PV process.
Know the rules
Explain what GVP, the EU AI Act, the AI literacy obligation and draft Annex 22 require for an AI use case in PV, and where data protection sets limits.
Prove performance
Write a testable intended use, build a reference set and decide on release, restriction or rejection using recall, precision and results for relevant subgroups.
Keep control
Design review steps, monitoring and change control that let you show an inspector the AI tool stays under control after go-live and through every change.
EU QPPVs, deputies and QPPV office staff
Heads of Pharmacovigilance, PV operations leads and global safety leads
PV specialists in case processing, literature, signal management and safety systems
QA, compliance and inspection readiness professionals
PV technology, automation and CSV/IT quality leads supporting PV systems
Vendor management and outsourcing oversight professionals in PV
Unsere Referent:innen
Reinhold Schilling, M.D.R.A.
Reinhold Schilling has headed global pharmacovigilance at Wörwag Pharma GmbH & Co. KG in Böblingen since 2021, where he is EU QPPV and Stufenplanbeauftragter. A chemist by training with a Master of Drug Regulatory Affairs, he has worked in drug safety since 2011, including earlier roles as Stufenplanbeauftragter and Informationsbeauftragter for Germany and Austria. He trains PV teams, audits PV systems and teaches pharmacovigilance at Reutlingen University and Goethe Business School. He is a member of the BPI pharmacovigilance working group.
EU QPPV in practice
Head of Global Pharmacovigilance, EU QPPV and Stufenplanbeauftragter at Wörwag Pharma GmbH & Co. KG.
Trainer and auditor
Trains pharmacovigilance staff and audits PV systems, processes and service providers.
Teaching
Lecturer in pharmacovigilance at Reutlingen University and in the Pharma MBA at Goethe Business School; member of the BPI PV working group.
Academic background
Diplom-Chemiker and Master of Drug Regulatory Affairs (M.D.R.A.), University of Bonn.
Seminar-Nr.
1226PPV601
Schnellübersicht
Alle wichtigen Infos auf einen Blick
Dauer
7,5 Stunden (9:00–16:30 Uhr)
Format
Online
Preis
390 EUR (zzgl. gesetzl. USt)
Nächster Termin
10.12.2026
Referent:in
Reinhold Schilling, M.D.R.A.
Sie haben weitere Fragen?
Kontaktieren Sie uns!
Sie haben Fragen zu diesem Seminar oder möchten eine individuelle Beratung? Wir freuen uns auf Ihre Nachricht.
