Print this chapterPrint this chapter

PV Modules

MODULE 5: DATA CODING

Data coding is the process of standardizing all medical and drug-related information in a safety case using internationally accepted dictionaries. This ensures uniformity across all cases and enables accurate signal detection, regulatory reporting, and data analysis.

Data coding occurs after case registration and before narrative writing.

A. Why Coding is Needed

  • To ensure consistency across global databases
  • To allow system-based search and signal detection
  • To avoid ambiguity from free-text entries
  • To ensure regulatory compliance
  • To make global safety reports uniform across all regions

B. What All Items Are Coded?

  1. Adverse Events / Reactions
  2. Indications
  3. Medical History
  4. Past Drug History
  5. Concomitant Medications
  6. Laboratory Tests (sometimes)

C. Coding Dictionaries Used

1. MedDRA (Medical Dictionary for Regulatory Activities)

Used for:

  • Adverse events
  • Indications
  • Medical history
  • Procedures
  • Investigations

MedDRA Hierarchy Levels:

  1. SOC (System Organ Class) – Highest level
  2. HLGT (High-Level Group Term)
  3. HLT (High-Level Term)
  4. PT (Preferred Term) – Main coding term
  5. LLT (Lowest Level Term) – Selected term for coding

Example:
Free text: “Patient had swelling of face”
Code:

  • LLT: Facial swelling
  • PT: Oedema of face

2. WHO Drug Dictionary (WHODD)

Used for:

  • Suspect drugs
  • Concomitant medications
  • Past medications
  • ATC classification

Key Components:

  • Drug Name – brand or generic
  • Preferred Base Name (PBN) – core active ingredients
  • ATC Code – classification of drug’s pharmacologic group
  • DDI (Drug Dictionary Identifier)

Example:
Free text: “Tab Paracetamol 500 mg”
Coding in WHODD:

  • PBN: Paracetamol
  • ATC: N02BE01

D. How Coding is Performed (Step-by-Step)

1. Read the source document thoroughly

  • Extract all medical terms exactly as written.
  • Avoid assumptions or interpretations.

2. Match the term to the most appropriate LLT

  • Use the MedDRA browser or internal dictionary.
  • Select the LLT, but the case will map to the PT automatically.

3. Ensure medical accuracy

  • Do not code signs/symptoms separately if a diagnosis is provided.
    • Example: If diagnosis is “Pneumonia,” do not code “fever,” “cough,” etc.

4. Code based on the reporter’s verbatim

  • No medical interpretation unless allowed by SOP.

5. Avoid vague or generalized terms

Incorrect: “Bad reaction” → too vague
Correct: Identify exact symptoms from the narrative.

6. Code drugs using WHO Drug

  • Match strength, dose form (tablet, injection), combination drugs.
  • Add ATC codes.

7. Review coded terms

  • Ensure:
    • No duplications
    • No contradictions
    • All relevant terms are captured
    • Primary suspect drug is correctly marked

E. What Should NOT Be Coded

  • Social history (unless medically relevant)
  • Reporter’s comments unless containing medical information
  • Administrative information (dates, identifiers)
  • Duplicate medical terms

F. Examples of Correct Coding

Example 1

Free text: “Patient developed severe itching and red patches after taking amoxicillin.”
Coding:

  • LLT: Pruritus
  • LLT: Erythema
    WHODD:
  • Amoxicillin → ATC: J01CA04

Example 2

Free text: “Past history of diabetes mellitus type 2.”
Coding:

  • LLT: Type 2 diabetes mellitus

Example 3

Free text: “Patient was prescribed Ibuprofen 400 mg for pain.”
Coding:

  • Ibuprofen → ATC: M01AE01

G. Quality Checks in Coding

  • Is the LLT the closest match?
  • Is the PT medically meaningful?
  • Are all relevant terms coded?
  • Is the drug-role classification correct?
    • Suspect
    • Concomitant
    • Interacting
  • Is the preferred term aligned with narrative and medical review?