Signal Management Knowledge Assessment

Delete the signal record because no action was taken Retain the assessment and define monitoring measures and triggers for reopening where appropriate Assume the association can never be reconsidered Transfer responsibility to the original reporter

The result proves that the product caused the event The result is irrelevant unless at least one randomised trial confirms it The result is an alert requiring clinical context, validation and potentially further assessment The result automatically requires a label change

Assume routinely coded diagnoses are valid without further testing Document data provenance, phenotype definitions, analytic methods and limitations such as confounding Use the largest available dataset regardless of its data quality Interpret an association as causal if it is statistically significant

A committee charter defining membership, quorum, conflicts of interest and decision rights Allowing any employee to change signal status without approval Using only informal discussions with no minutes Restricting assessment to one discipline

Automatically withdrawing the product whenever a signal is validated Selecting among monitoring, additional PV, risk minimisation, studies or regulatory referral according to evidence and benefit-risk considerations Taking no action whenever causality is not proven beyond doubt Issuing a DHPC for every disproportionality alert

To establish definitive causality To make a governance decision that a validated signal should undergo structured assessment To close every signal that has not produced a regulatory action To determine the exact incidence of the event

To demonstrate how risk-based decisions were made and how changing evidence affected priority To replace the need for a written assessment To ensure that a signal can never be reprioritised To calculate the number of adverse reactions in the general population

They directly estimate absolute incidence in the treated population They are unaffected by reporting behaviour or coding quality They can be influenced by confounding by indication, stimulated reporting and MedDRA coding They are only valid when all cases have rechallenge information

Discard the alert because replication across all databases is mandatory Treat the discrepancy as evidence of causality in the positive database Investigate differences in populations, coding, reporting practices, study design and data quality Average the two results and close the signal

A new case series or epidemiological finding materially changes the evidence The signal register has reached its monthly storage limit The original signal owner changes job title The product has an established brand name