MODULE 13: SIGNAL DETECTION

A signal is: “Information that suggests a new causal association or a new aspect of a known association between a medicinal product and an adverse event.”

A signal does not confirm causality—it indicates that further evaluation is required.

Sources of Signals: Signals may come from:

A. Spontaneous Reporting Systems

  • Individual Case Safety Reports (ICSRs)
  • Company safety database

B. Scientific Literature

  • Case reports
  • Drug safety studies

C. Clinical & Epidemiological Studies

  • Clinical trial data
  • Post-marketing studies

D. Periodic Reports

  • PSUR / PBRER
  • DSUR
  • PADER

E. Other Sources

  • Social media (some agencies consider)
  • Patient support programs
  • Drug utilization data

Steps in Signal Detection

Step 1: Data Collection

Sources include:

  • Safety databases (global & local)
  • Literature screenings
  • Periodic reports (PSUR, PBRER)
  • Regulatory agency reports (e.g., FDA FAERS)
  • Clinical trial data (DSUR)

Objective: Gather all AE data and identify unusual patterns.

Step 2: Statistical Methods

Used to detect disproportionality—i.e., whether a drug-event pair is reported more often than expected.

Common tools:

  • PRR (Proportional Reporting Ratio)
  • ROR (Reporting Odds Ratio)
  • IC (Information Component – Bayesian)

Example:
If “Drug X – Liver Injury” shows a PRR > 2 → potential signal.

Step 3: Medical Review

Medical experts evaluate:

  • Clinical relevance
  • Biological plausibility
  • Severity & medical seriousness
  • Case quality (dechallenge/rechallenge)
  • Time-to-onset pattern
  • Consistency across cases

Outcome: Decide if the signal seems medically meaningful.

Step 4: Signal Validation

Determine whether the signal is:

  • Valid
  • Not valid
  • Needs more data

Criteria used:

  • Case quality
  • Plausibility
  • Consistency across sources
  • Exposure data

Step 5: Signal Prioritization

High priority signals:

  • Serious AEs (death, life-threatening)
  • High frequency
  • Preventable issues
  • High public health impact

Low priority signals:

  • Non-serious & rare events
  • Weak biological plausibility

Step 6: Signal Assessment

A deep evaluation involving:

  • Causality assessment
  • Benefit–risk impact
  • Case series analysis
  • Literature review
  • Comparison with similar drugs

Step 7: Regulatory Actions (If Confirmed)

Actions may include:

  • Safety label update
  • SPC / PI revisions
  • Boxed warning addition
  • Risk Minimisation Measures (RMMs)
  • Updating RMP
  • Health authority communication (DHPC, alert)

Examples of Signals

Drug

Observed AE

Signal Type

Action

Fluoroquinolones

Tendon rupture

New safety signal

Label update

Metformin

Vitamin B12 deficiency

Known, new aspect

RMP update

DPP-4 inhibitors

Pancreatitis

Potential signal

Further assessment

 

6. Important Terms

New Signal

Previously unknown; requires investigation.

Emerging Safety Issue

Urgent signal requiring immediate communication.

Potential Signal

Needs more data; not yet validated.

7. Case Example of Signal Detection

Drug: "Cardiotab"
Reported AE: Arrhythmia
Data:

  • 15 cases reported in 3 months
  • PRR = 3.5 (significant)
  • Consistent time-to-onset (1–3 days)
  • 3 positive rechallenge cases
  • Plausible mechanism (QT prolongation)

Outcome:
→ Validated signal
→ Label updated
→ ECG monitoring added as RMM