Editing AI for Drug Adverse Effect Detection (section)

== <span style="color: #FFFFFF;">Analyzing</span> ==
{| class="wikitable"
|+ Pharmacovigilance Signal Detection Methods
! Method !! Data Source !! Signal Speed !! Causal Evidence Strength
|-
| Disproportionality (ROR, IC) || Spontaneous reports || Months || Weak (association only)
|-
| SCCS (self-controlled) || EHR || Months || Moderate (controls confounding)
|-
| CPRD prescription sequence || EHR || Months-years || Moderate
|-
| NLP from social media || Twitter, Reddit || Weeks || Weak (unverified)
|-
| NLP from clinical notes || Hospital EHR || Months || Moderate
|-
| ML DDI prediction || Molecular data || Pre-market || Moderate (in silico)
|}

'''Failure modes''': Notoriety bias — high-media-coverage drugs generate more reports regardless of true ADR rate. Indication confounding — patients taking drug X for condition Y; outcome Y attributed to drug X. Weber effect — spontaneous reporting spikes in years 1-2 post-launch, then declines, creating spurious temporal trends. Social media noise — most drug-related posts don't contain genuine ADE reports. NLP coding errors — mapping free text to MedDRA can introduce systematic errors.
</div>

<div style="background-color: #483D8B; color: #FFFFFF; padding: 20px; border-radius: 8px; margin-bottom: 15px;">