Active Pharmacovigilance Surveillance

Active pharmacovigilance surveillance deliberately seeks out adverse events in defined populations rather than waiting for them to be reported voluntarily. By systematically following cohorts of treated patients or interrogating large healthcare databases, it aims to overcome the under-reporting and missing denominator that limit spontaneous reporting and to estimate how often reactions actually occur.

Definition

Active pharmacovigilance surveillance is a proactive approach to drug-safety data collection in which adverse events are systematically ascertained in a defined population of medicine users, allowing estimation of event frequency and comparison against a denominator.

Scope

The entry covers the rationale for actively collecting safety data, the main approaches — cohort event monitoring, prescription-event monitoring, and large-scale electronic-record or claims-based surveillance networks — and how active methods complement passive reporting. It is a reference overview of surveillance methodology, not clinical advice.

Core questions

• Why supplement spontaneous reporting with active methods?
• How does cohort or prescription-event monitoring ascertain events?
• How do database and sentinel networks enable large-scale surveillance?
• What can active surveillance estimate that passive reporting cannot?

Key concepts

• Cohort event monitoring
• Prescription-event monitoring
• Sentinel and distributed-data networks
• Common data model
• Denominator and incidence estimation
• Targeted (event-driven) surveillance
• Electronic health records and claims data

Mechanisms

Active surveillance defines a population of medicine users and then ascertains adverse events within it by design. In prescription-event monitoring and cohort event monitoring, patients dispensed a drug of interest are identified and followed, and events are solicited systematically (Kasliwal et al., 2008). In database and sentinel approaches, routinely collected electronic health records or insurance claims are queried — often through a shared common data model so that the same analysis can run across many data partners — to detect and quantify drug-outcome associations (Platt et al., 2009; Stang et al., 2010). Because the denominator of exposed patients is known, these methods can estimate incidence and relative risk, which spontaneous reporting cannot (Härmark & van Grootheest, 2008).

Clinical relevance

Active surveillance produces the population-level incidence and risk estimates that inform regulatory action and that clinicians encounter in safety communications. This entry describes how such evidence is collected and is not a basis for individual diagnostic or treatment decisions.

Epidemiology

Prospective studies illustrate the scale of drug-related harm that active ascertainment can quantify — for example, a large UK prospective analysis attributed roughly 1 in 16 hospital admissions to adverse drug reactions (Pirmohamed et al., 2004). Modern distributed networks extend such ascertainment to tens of millions of patients' records (Platt et al., 2009; Stang et al., 2010).

History

Active methods grew up alongside spontaneous reporting to address its blind spots. Prescription-event monitoring was developed in the United Kingdom from the 1980s to follow cohorts of patients on newly marketed drugs, and from the late 2000s large database-driven initiatives such as the US Sentinel Initiative and the Observational Medical Outcomes Partnership formalised active surveillance across networks of electronic healthcare data (Platt et al., 2009; Stang et al., 2010).

Debates

How should confounding be controlled in database surveillance?

Routinely collected data are not randomised, so apparent drug-outcome associations may reflect why a drug was prescribed rather than its effect; methods for confounding control and the reliability of automated signal screening across heterogeneous databases remain actively debated.

Key figures

• Richard Platt
• Saad Shakir
• Linda Härmark
• Patrick Ryan

Seminal works

• platt-2009
• stang-2010

Frequently asked questions

How does active surveillance differ from spontaneous reporting?

Spontaneous reporting waits for observers to volunteer reports, whereas active surveillance deliberately seeks events in a defined population. Because the exposed population is known, active methods can estimate how often a reaction occurs, which spontaneous reporting cannot.

Why use electronic health records or claims data for drug safety?

They capture large numbers of treated patients with their outcomes already recorded, allowing rapid, large-scale estimation of drug-outcome associations — though their non-randomised nature means confounding must be carefully addressed.