ICSR Drug Coding Automation (WHODrug + NLP)
Automates a large share of case-level drug coding to reduce manual pharmacovigilance workload Evidence basis: WHODrug Koda evaluation on 4.8 million VigiBase entries reported major automation gains with high coding accuracy; additional NLP work supports scalable coding from unstructured narratives with expert oversight for exceptions
The Problem
“ICSR Drug Coding Automation (WHODrug + NLP)”
Organizations face these key challenges:
Automates a large share of case-level drug coding to reduce manual pharmacovigilance workload
Impact When Solved
The Shift
Human Does
- •Review each ICSR and identify reported drug names from case details
- •Manually code drugs to WHODrug terms and resolve ambiguous matches
- •Coordinate case follow-up and track coding status in spreadsheets
- •Perform retrospective quality checks and correct coding inconsistencies
Automation
- •No meaningful automation in routine drug coding
- •Limited rule-based lookup support for reference checking
- •Minimal prioritization of cases needing urgent coding attention
Human Does
- •Review low-confidence or ambiguous coding suggestions and make final coding decisions
- •Approve coded cases that require expert validation before submission
- •Handle exceptions from incomplete narratives, unusual products, or conflicting source data
AI Handles
- •Extract reported drug information from structured fields and unstructured case narratives
- •Match case drugs to WHODrug terms and propose standardized codes at scale
- •Prioritize straightforward cases for rapid processing and flag uncertain cases for review
- •Monitor coding consistency and surface potential discrepancies for human follow-up
Operating Intelligence
How ICSR Drug Coding Automation (WHODrug + NLP) runs once it is live
AI runs the first three steps autonomously.
Humans own every decision.
The system gets smarter each cycle.
Who is in control at each step
Each column marks the operating owner for that step. AI-led actions sit above the divider, human decisions and feedback loops sit below it.
Step 1
Assemble Context
Step 2
Analyze
Step 3
Recommend
Step 4
Human Decision
Step 5
Execute
Step 6
Feedback
AI lead
Autonomous execution
Human lead
Approval, override, feedback
AI handles assembly, analysis, and execution. The human gate sits at the decision point. Every cycle refines future recommendations.
The Loop
6 steps
Assemble Context
Combine the relevant records, signals, and constraints.
Analyze
Evaluate options, risk, and likely outcomes.
Recommend
Present a ranked recommendation with supporting rationale.
Human Decision
A human accepts, edits, or rejects the recommendation.
Authority gates · 1
The system must not finalize coding for low-confidence, ambiguous, or conflicting cases without review by a pharmacovigilance coding specialist or safety reviewer. [S1]
Why this step is human
The decision carries real-world consequences that require professional judgment and accountability.
Execute
Carry out the approved action in the operating workflow.
Feedback
Outcome data improves future recommendations.
1 operating angles mapped
Operational Depth
Technologies
Technologies commonly used in ICSR Drug Coding Automation (WHODrug + NLP) implementations: