Hicksville, Newyork
+1-681-448-4202

ICSR Processing 101: How an Adverse Event Goes from Report to Regulatory Submission

Table of Contents

Ready To Automate Your Clinical Workflows?

Empower your research teams with Clinevo’s end-to-end unified eClinical platform for faster, data-driven decisions.
Summarize and analyze this article with

Perplexity

Grok

Google AI

Claude

Every minute, somewhere in the world, an adverse event report is being filed. EudraVigilance alone collected and managed 1,765,581 Individual Case Safety Reports (ICSRs) in 2025, taking its cumulative post-authorisation total to over 29 million reports since 2002.Across the Atlantic, the FDA’s Adverse Event Reporting System (FAERS) held more than 28 million reports as of December 2023, with around 95% of submissions originating from application holders in the pharmaceutical industry rather than direct public reporting.

Behind every one of those reports sits a tightly choreographed workflow. The job of pharmacovigilance case management is to convert a fragmented, often unstructured account of a patient experience into a fully coded, regulator-ready ICSR within tight, expedited reporting windows, with an audit trail that holds up under inspection.

This guide walks through how a spontaneous adverse event report becomes a structured ICSR, what each stage of the workflow is responsible for, and where AI-driven case intake and automation-led safety database tools change what is operationally possible.

What Makes an ICSR Valid: The Four Minimum Criteria

Before any case enters formal processing, it goes through a validity check. ICH E2B(R3), the global standard adopted by the FDA, EMA, PMDA, MHRA, and Health Canada, defines four minimum data elements an ICSR must carry to be accepted for regulatory submission.

A report missing any one of these is classified as non-valid. It is still logged for traceability, but it cannot be submitted until follow-up information closes the gap.

Why this matters operationally: Validity is the first quality gate. A weak intake step pushes invalid cases deeper into the workflow, where rework is far more expensive. PV teams that catch validity gaps at the point of intake protect every downstream stage from contaminated data.

Stage What Happens Compliance Anchor
1. Case Intake Reports arrive through call centres (MICC), email, web portals, partner forms, social media monitoring, and literature surveillance. Source documents (CIOMS, MedWatch, SAE forms, free-text emails, voice transcripts) are converted into structured data. ICH E2D Day 0 starts at receipt of the four minimum elements. This ‘Date of Awareness’ triggers the regulatory clock (7 or 15 calendar days) for expedited submission.
2. Validity Check & Triage Each report is checked against the four ICH E2B(R3) criteria. Valid cases are then triaged for seriousness using ICH E2D criteria (death, hospitalisation, life-threatening, disability, congenital anomaly, or other medically important events) and routed accordingly. ICH E2D, GVP Module VI.
3. Data Entry & MedDRA Coding Reactions, indications, and medical history are coded against MedDRA. Version 28.0 spans 27 System Organ Classes, 26,920 Preferred Terms, and 89,774 Lowest Level Terms, the controlled vocabulary used across global regulators. ICH M1 (MedDRA), ISO 27953-2.
4. Causality & Narrative Reviewers apply causality methods (WHO-UMC, Naranjo, or sponsor-defined criteria) to assess the relationship between the suspect product and the event. The case narrative captures the clinical sequence, concomitant medications, medical history, and outcome in structured but readable prose. WHO-UMC causality categories; sponsor SOPs.
5. Quality & Medical Review QC reviewers check completeness, coding accuracy, and narrative coherence. A qualified medical reviewer signs off on the clinical assessment before the case can be locked. 21 CFR Part 11, EU GMP Annex 11 audit trails.
6. E2B(R3) Submission Validated cases are converted to E2B(R3)-compliant XML and transmitted via secure gateways (typically AS2) to the relevant health authorities. ICH guidelines mandate 7 calendar days for fatal or life-threatening unexpected reactions and 15 days for other serious unexpected reactions. ICH E2B(R3), regional implementation guides.
7. Follow-up Management When initial information is incomplete, PV teams issue follow-up queries to the reporter, capturing all correspondence in the audit trail. Each response generates a versioned case record to track the evolution of data. This stage also includes Nullification and Duplicate Management to merge redundant reports. GVP Module VI Rev. 2.

50

New molecular entities cleared by the FDA in 2024. Each approval extends a post-marketing surveillance obligation that the sponsoring organization must maintain indefinitely — expanding workloads without automatic increases in team capacity.

A few operational realities the workflow has to absorb

These principles describe what the intake layer should do. The next question is what implementing them looks like in production.

Where the Modern PV Stack Earns Its Place

Modern ICSR processing succeeds or fails on two axes: how much friction the pipeline carries between intake and submission, and how defensible each step is under regulatory inspection. Patching individual stages with isolated tools tends to leave handoff gaps. Tools that work as a connected stack remove them.

Clinevo OnePV is built around both demands.

Compliance with 21 CFR Part 11, EU GMP Annex 11, GxP, and GDPR is built into the platform architecture. Every action, from intake through submission, is captured in timestamped, tamper-evident audit trails that are retrievable on demand for inspection.

Frequently Asked Questions

A spontaneous report (a phone call, email, web submission, or letter) is logged at intake with whatever information is available. The PV team then validates it against the four minimum criteria, triages for seriousness, codes the reaction terms in MedDRA, performs causality assessment, writes a narrative, and routes it through QC and medical review. Once locked, it is converted to an E2B(R3)-compliant XML message and submitted to the relevant regulator. Modern platforms automate most of the data capture and routing, with human reviewers focused on clinical judgement.

Four elements are required under ICH E2B(R3): an identifiable patient, an identifiable reporter, a suspect drug or biological product, and a suspected adverse event or reaction. Any one of several data points is sufficient to meet each criterion. A report missing one of the four is classified as non-valid and cannot be submitted until follow-up closes the gap.

AI-driven case intake tools can reliably extract candidate seriousness criteria from narrative text, hospitalisation language, fatal outcomes, life-threatening descriptors, congenital anomalies, and flag them for review with confidence scores. The final seriousness determination remains a regulated medical decision made by a trained reviewer. The role of AI is to surface signals consistently and accelerate triage, not to replace medical judgement.

The reviewer first identifies the verbatim reaction term as reported. They then locate the appropriate Lowest Level Term (LLT) in MedDRA, which auto-links to a Preferred Term (PT). The PT is what flows into reporting and signal detection. The PT then aggregates upward into High Level Terms, High Level Group Terms, and finally a System Organ Class. Coders use sponsor-specific coding conventions (for example, granularity rules for vague terms) to ensure consistency, and a second reviewer typically verifies the coding before case lock.

The case is logged with the available information and a follow-up request is issued to the reporter, capturing the request and any response in the audit trail. Each follow-up creates a new versioned case record. If the follow-up changes seriousness, causality, or expectedness in a way that meets expedited reporting criteria, the case clock starts again from the receipt of the new information. Multiple follow-up attempts may be required, and regulators expect documented evidence of what was asked, when, and what was received.

The receiving system returns an acknowledgement message confirming receipt and validation status. The case enters the regulator's database and becomes part of the ongoing safety surveillance dataset, available for signal detection, periodic safety reporting, and aggregate analysis. Marketing authorisation holders can download cases related to their products from EudraVigilance for their own signal management activities.