ENCePP Working Group on AI in Pharmacoepidemiology & Pharmacovigilance · 2026

Perception & Use of AI in
Pharmacovigilance & Pharmacoepidemiology

ENCePP has asked its Artificial Intelligence Working Group to conduct this survey to understand how its members and the broader PE/PV community use AI tools in pharmacoepidemiology and pharmacovigilance.

⏱ Estimated time: 12–15 min
🔒 Fully anonymous — GDPR-compliant
🏛 Data controller: University of Copenhagen

Built on Colleague Feedback

Every question in this survey was shaped by comments and suggestions from ENCePP WG members. This initiative directly incorporates input from the February 2026 WG meeting, the ARCS PV survey, and the infrastructure subgroup contributions led by WG colleagues. Your prior engagement is woven into every section below.

PV Survey (ARCS) — use cases, conversational AI, benefit-risk WG Meeting Feb 2026 — inventory, barriers, PE/PV split Infrastructure Subgroup (JD, FR, MG) — computing, data, governance Barriers Subgroup (AL, CP, DL) — education, legal, privacy Inventory Subgroup (LC, IT) — AI tool landscape mapping EMA AI Observatory (2024) — scope alignment

📖 Definitions used in this survey

Artificial Intelligence (AI)
A machine-based system that, for a given set of objectives, makes predictions, recommendations, or decisions influencing real or virtual environments. AI systems are designed to operate with varying levels of autonomy. (Adapted from OECD AI Principles, 2019; CIOMS Working Group XIV, 2025)
AI Tool
Software applications or programs that utilise artificial intelligence algorithms to perform tasks such as decision-making, problem-solving, and data analysis — including off-the-shelf systems (e.g., Microsoft Copilot, ChatGPT, Claude AI, Gemini), domain-specific platforms, and custom-built models used to support PE or PV activities. (ScienceDirect: AI Tools; adapted for PE/PV context)
Note: Questions about "AI tools" encompass both AI and machine learning (ML) approaches unless otherwise specified. ML is a subset of AI; where answer options mention ML explicitly, these are illustrative examples of AI methods.
👥
Who should fill in this survey?
Everyone involved in pharmacoepidemiology or pharmacovigilance is welcome — regardless of role, seniority, sector, or level of AI experience. You do not need to be an AI expert to participate. This survey is for all people working in or around PE and PV, including — but not limited to:
Drug safety and PV professionals (associates, managers, directors, QPPVs)  ·  Pharmacoepidemiologists and RWE researchers  ·  Clinical pharmacologists and physicians  ·  Regulatory scientists and HTA analysts  ·  Data scientists, AI engineers, and bioinformaticians  ·  Biostatisticians and epidemiologists in industry  ·  Hospital pharmacists and clinical researchers  ·  CRO, consulting, and outsourcing professionals  ·  Academic researchers, postdocs, and PhD students  ·  Medical writers and regulatory affairs specialists  ·  IT professionals, system administrators, and project managers  ·  Administrative staff and coordinators supporting PE/PV teams  ·  Patient advocates and representatives involved in safety activities  ·  …and anyone else whose work touches pharmacoepidemiology or pharmacovigilance in any capacity.
If you are unsure whether the survey applies to you — it does. Your perspective matters.
🔄
Do you work in both pharmacovigilance and pharmacoepidemiology? The survey has two separate tracks — one for PV and one for PE. If your work spans both fields, we kindly ask you to complete the survey twice: once selecting the PV track and once selecting the PE track. Your two responses will each capture a different and complementary perspective, and both are equally valuable to us.
📢
Multiple people from the same organisation are encouraged to participate. There is no limit on the number of responses from a single institution — different colleagues will have different perspectives on AI adoption, barriers, and priorities, and all of them are welcome. We kindly ask you to share this survey widely within your organisation and networks: with colleagues, collaborators, and anyone working at the intersection of AI and drug safety or pharmacoepidemiology. The more diverse and representative our responses, the stronger the evidence base we can build together.
⚠️
Please complete this survey only once per track. If you have already submitted a response for a given track (PV or PE), please do not complete that same track again. Duplicate responses from the same individual within the same track cannot be identified and will bias the results.

Data Protection & Privacy

This survey is conducted in compliance with GDPR (EU 2016/679). The University of Copenhagen is the data controller. No personally identifying information is collected — responses are entirely anonymous and cannot be linked to individual participants. Data will be stored securely on SurveyXact (EU data protection standards, EEA storage) for five years and deleted thereafter. Results may be published in aggregate form in academic publications, ENCePP newsletters, and related venues. By continuing, you confirm you have read this information and consent to participate.

Quick question before you start:
Does your organisation use at least one off-the-shelf AI system (e.g., Microsoft Copilot, ChatGPT, Claude AI, Gemini) for work purposes?
Takes approximately 12–15 minutes  ·  All fields marked * are required
Screening Question

What is your primary area of expertise?

Your answer determines which tailored module you will receive.

Q0 Gate *
Which best describes your primary professional focus?
Please select your primary area of expertise to continue.
Section 1 of 4

Your Profile

A few background questions to help us understand respondents. All data is reported in aggregate only.

Q1 Profile *
What is your current role and organisation?

Two sub-questions below capture your professional profile. Responses are reported in aggregate only — in the unlikely event that only a small number of respondents select a category, results for that category will not be reported separately.

1a. Your current role *

Options will reflect your selected track.

Please select your current role.
1b. Type of organisation *
Please specify your organisation type
Please select your organisation type.
Q3a Profile
In which European country is your organisation primarily based?
💬
Any comments on this section? Optional — share anything about your background or context not captured above
Section 2 — Shared (all respondents)

AI Adoption & Use Case Inventory

These questions apply to all respondents and build the ENCePP community inventory of AI adoption across PE and PV settings.

Q21Use Case*
Does your organisation currently use or actively develop AI tools for any pharmacoepidemiology or pharmacovigilance activities?
Q23bUse Case
Which AI methods or approaches does your organisation currently use or develop for PE/PV? (Select all that apply)
Select all that apply
Please specify the method
Q24Wishes
Which of the following AI use cases would your organisation most like to implement or expand, but has not yet been able to? (Select up to 3)
Select up to 3
Please specify the use case you wish to expand
Section 3 of 4

Infrastructure, Governance & Skills

These questions were developed by the ENCePP Infrastructure Subgroup (JD, FR, MG) and Barriers Subgroup (AL, CP, DL), specifically requested during the February 2026 WG meeting.

I · Overall Readiness
I-2Infrastructure
Overall, how would you rate your organisation's infrastructure readiness for adopting, maintaining, and scaling AI in pharmacoepidemiology or pharmacovigilance?
What would most improve your readiness? (optional)
💬
Comments on overall AI readiness? Optional — your broader view on what readiness means for your organisation
II · Computing & Hardware
I-3Computing
What type of computing infrastructure does your organisation currently have available for AI workloads in pharmacoepidemiology or pharmacovigilance?
I-4Computing
How would you rate the adequacy of your organisation's computing resources (processing power, GPU availability, storage) for running AI/ML applications?
💬
Comments on computing infrastructure? Optional — any specifics about your compute setup or challenges
III · Data Infrastructure
I-5Data
What is the primary type of data environment your organisation uses for PE/PV research involving AI?
I-6Data
To what extent does your organisation have access to datasets suitable for training or applying AI/ML models in PE/PV?
I-7Barrier
What is the primary barrier your organisation faces regarding health data use for AI-based PE/PV research?
💬
Comments on data infrastructure or access? Optional — data access challenges, GDPR constraints, interoperability issues
I-DSData
Which data sources does your organisation use — or has it used — for AI-supported PE/PV activities? (Select all that apply)

This question maps available data infrastructure — select all sources your organisation has access to for AI-driven work, even if not yet fully used.

Select all that apply
Please specify the data source
IV · Software, Tools & Support
I-8Tools
Which best describes the availability of AI/ML software tools and platforms (e.g., Python/R libraries, TensorFlow, commercial AI platforms) at your organisation?
I-9Tools
Does your organisation have dedicated technical support (e.g., data engineers, AI/ML engineers, bioinformaticians, cybersecurity experts) to assist researchers with AI/ML applications?
I-10Tools
Does your organisation provide access to validated or pre-approved AI tools specifically designed for PE/PV tasks (e.g., signal detection, case narrative assessment, automated coding)?
💬
Comments on AI tools or technical support? Optional — tools you recommend, gaps in tooling, or support challenges
V · Governance & Compliance
I-11Governance
Does your organisation have a formal policy or governance framework for the use of AI in research (including pharmacoepidemiology and pharmacovigilance)?
I-12Governance — EU AI Act
How well prepared is your organisation to comply with EU AI Act requirements as they pertain to AI applications in pharmacoepidemiology or pharmacovigilance?
I-13Governance
Does your organisation have a process in place for the validation and quality assurance of AI/ML models used in PE/PV research?
I-14Governance
Are you personally aware of current national or international regulations on AI in pharmacovigilance or pharmacoepidemiology?
💬
Comments on AI governance or regulatory compliance? Optional — your experience with the EU AI Act, EMA guidance, or internal AI policies
VI · Training & Workforce
I-15Skills
What best describes the current level of AI/ML expertise among researchers in your organisation who work in PE/PV?
I-16Skills
Have you personally received training on how to use AI tools in your work?
I-17Skills
Does your organisation offer training programmes or capacity-building activities specifically related to AI/ML for PE/PV researchers?
I-18Skills
Please rate your personal understanding of the following AI-related terms in the context of PE/PV:

1 = No understanding  ·  5 = Full expert understanding

Term 12345
Machine learning (ML)
Natural language processing (NLP)
Large language models (LLMs)
Predictive analytics
Signal detection algorithms
VII · Overall Barriers — Synthesis

Now that you have worked through all the specific questions above, please give us your overall assessment:

I-1Barriers — Overall
Having considered all the dimensions above — what is the single most critical barrier your organisation faces in adopting or expanding AI in PE/PV?
Please describe the barrier
💬
Comments on AI training or workforce skills? Optional — gaps, successes, or what kind of training would help most
Section 4 of 4 — Shared

Perceptions, Attitudes & Wishes

This final section is shared by all respondents and captures your personal views, community priorities, and open feedback for the ENCePP WG.

Perception Statements
P-1Perception
To what extent do you agree or disagree with the following statements? (Strongly Disagree → Strongly Agree)

SD = Strongly Disagree  ·  D = Disagree  ·  N = Neutral  ·  A = Agree  ·  SA = Strongly Agree

Statement SDDNASA
Generative AI has already changed how PE/PV is conducted
Ethical concerns about AI in PE/PV need more attention
AI threatens the job security of PE/PV professionals
Regulatory guidance on AI in PE/PV is sufficiently clear
My organisation is ready to scale AI applications in PE/PV
AI tools currently used in my field are trustworthy and validated
P-2Perception
How comfortable would you be with implementing conversational AI tools for patients or HCPs to interact with in your organisation's PE/PV context (e.g., chatbots for reporter follow-up or patient communication) — as opposed to using AI yourself for internal tasks?
Enablers & Priorities
P-3Wishes
What would most enable better AI adoption in PE/PV in your context? Select up to 3

This question asks for your personal or organisational priority — distinct from the objective infrastructure assessment in Section 3.

Select up to 3 options
Open Questions
P-4Open
What advice would you give to pharma leaders or regulators regarding AI in PE/PV?

Free text — share as much or as little as you like

P-5Open — Wishes
Would you be interested if ENCePP or the WG provided AI training sessions or resources specifically for PE/PV professionals? If yes, what topics would you most want covered?

Please mention specific topics, formats, or time commitments you would find valuable

P-6Open — WG Feedback
Do you have any other comments, suggestions, or questions for the ENCePP Working Group on AI in PE & PV?

Your suggestions directly feed into the WG's work programme and survey revisions

💬
Any further thoughts on AI perceptions or priorities? Optional — views not captured in the structured questions above
Section 5 of 5 — Shared

AI Tool Inventory & Use Case Showcase (Optional) 0 entries

Document every AI tool you have used in PE or PV — one structured entry per tool. This builds the ENCePP inventory of AI applications across the community.

This section is entirely optional. Privacy note: Providing tool names, URLs, or vendor details may make your response identifiable. Please share only what you are comfortable disclosing. You may use generic descriptions (e.g., "a commercial NLP tool") if you prefer. If you have used one or more specific AI tools in PE or PV, you are warmly invited to document them below — each entry asks you to describe a tool you have actually used — its purpose, the data it ran on, how it performed, limitations, and references. Fields marked * are required for each entry you add. You may add as many entries as you like — one per tool, method, or project. The more detail you provide, the more useful the inventory becomes for the community.

Thank you for your participation!

Your input helps us understand how AI is — and should be — used in pharmacoepidemiology and pharmacovigilance across Europe. Your voice is heard, and your response will directly inform the ENCePP WG's evidence base and future work programme.

🏛 University of Copenhagen 🔒 Anonymous & GDPR-compliant 📊 Results published in aggregate 🤝 ENCePP WG on AI in PE & PV