Definition · Plain-language
Signal detection
Signal detection is the pharmacovigilance process of sifting adverse-event data to identify new or changing safety signals — possible causal associations between a medicine and an event that warrant further evaluation.
The step most authors miss
Doing CRediT right? Don’t stop at the statement.
A CRediT statement credits you inside one paper. The recognition CRediT was built for happens when those roles are tied to you, persistently. Sign in with your ORCID — free — and claim your CRediT contributions on casrai.org, the home of the standard. They become a verified, portable part of your identity, not a line that disappears into one PDF.
Free: claim your contributions, then export a journal-ready CRediT statement, schema.org structured data, JATS XML, CSV or BibTeX — and preview your public profile. A membership publishes that profile publicly and verifies the journals you serve.
What a safety signal is
In pharmacovigilance, a signal is information that suggests a new potentially causal association, or a new aspect of a known association, between a medicine and an adverse event — information judged sufficient to justify further investigation. Signals can point to a previously unrecognised reaction, a change in frequency or severity, or a new at-risk group. Crucially, a signal is provisional: it raises a question rather than answering it, and detecting one is the start of an evaluation process, not its conclusion.
Quantitative methods
Large adverse-event databases are screened using statistical data-mining techniques known collectively as disproportionality analysis. These methods compare how often a particular drug-event pair is reported against what would be expected if there were no association, using measures such as the proportional reporting ratio (PRR) and the reporting odds ratio (ROR). A disproportionately high reporting frequency flags the pair for attention. Because the underlying data are spontaneous reports, these measures highlight statistical patterns in reporting; they do not by themselves quantify real-world risk.
From signal to assessment
Quantitative flags are combined with qualitative clinical review of the underlying case reports, the biological plausibility of the association and any supporting evidence from other sources. Validated signals are prioritised and formally assessed, which may draw on epidemiological studies or other data. The outcome can range from no action, where the signal is not substantiated, to regulatory measures such as updated product information or risk-minimisation activities. Throughout, the distinction between a detected signal and an established causal risk is maintained.
Key facts
At a glance
- Definition: Identifying new or changing safety signals from pharmacovigilance data.
- Signal: Information suggesting a possible new causal association, warranting review.
- Quantitative tools: Disproportionality analysis (e.g. PRR, ROR).
- Inputs: Spontaneous reports and pharmacovigilance databases.
- Nature: A signal is a hypothesis, not a confirmed risk.
- Next step: Validation and formal signal assessment.
Common misconceptions
What people often get wrong
Often heard: A detected signal means the medicine has been shown to cause the event.
Actually: A signal is a hypothesis flagged for further evaluation. It indicates a possible association worth investigating, not a confirmed causal relationship.
Often heard: Disproportionality measures like PRR and ROR give the real-world risk of a side effect.
Actually: These measures describe patterns in reporting data, not true incidence or risk. They help prioritise what to investigate, not quantify how often an event occurs.
Often heard: Signal detection is a purely automated statistical exercise.
Actually: Quantitative screening is combined with clinical and qualitative review of case reports and plausibility. Human assessment is integral to deciding whether a statistical flag is a genuine signal.
Going deeper
