Pharmacovigilance

Why it exists after approval

A pivotal trial might enrol a few thousand carefully selected patients over months. Launch exposes hundreds of thousands of real patients — older, sicker, on other drugs, for longer. Adverse events with an incidence of 1 in 10,000 are effectively invisible pre-approval and unmissable afterwards. Pharmacovigilance is the system built to catch what trials structurally cannot.

The operating cycle

Collection. Reports arrive from healthcare professionals, patients, literature, studies, regulators and — increasingly — digital channels. Each becomes an ICSR.

Assessment. Each case is evaluated for seriousness, expectedness against the reference safety information, and causality. Seriousness and expectedness together set the reporting clock: serious unexpected cases carry the tightest deadlines.

Signal detection. Individual cases rarely prove anything. Signals emerge from patterns across the database, using disproportionality statistics alongside clinical review.

Risk management. A confirmed signal drives action — label change, communication, restricted distribution, or withdrawal.

What makes it hard

The work is dominated by volume and deadline risk. Case processing is labour-intensive, largely manual, and governed by clocks measured in calendar days from the moment anyone in the company became aware. Late submission is itself a compliance failure, independent of the safety finding.

The second difficulty is signal-to-noise. Reporting is voluntary, incomplete and confounded. A safety database records what was reported, not what happened, and every method has to work around that.