Analysis · Health
Why a Drug Safety Signal Cannot Tell You Your Risk
By AWEI · AI-compiled · Published · Analysis prepared · 1 source · doi.org
A medicine can appear unusually often alongside a health problem in a reporting database without showing how often that problem occurs among people taking it. The missing information is crucial: reports of suspected reactions do not provide a complete count of everyone exposed, or every event that occurred.
Michele Fusaroli and colleagues explain this distinction in their 2024 READUS-PV explanation and elaboration paper, published in Drug Safety. This is methodological guidance, not a treatment experiment. It explains 14 reporting recommendations, connecting each to methodological concerns and practical details that researchers should disclose. The checklist’s development is described in a companion paper.
The underlying method, disproportionality analysis, searches for drug–event combinations reported more often than expected relative to a chosen comparison group within a database. That makes the comparison group consequential: changing which reports form the background can change the result. The authors therefore call for researchers to specify and justify their comparator, statistical measure and threshold for flagging an association.
Other seemingly routine choices matter too. Researchers must explain how they removed duplicate reports, handled missing information and grouped drug names or medical events. Reporting can also vary with warnings, geography and prescribing patterns. Underlying illnesses and other medicines offer alternative explanations for an apparent association. A statistical signal alone cannot establish causation or rank medicines by risk.
The paper recommends examining individual cases for details such as timing, other treatments and alternative causes, alongside evidence from other sources. These steps can help assess whether a signal merits further investigation. Equally, finding no disproportionate reporting does not establish safety, and lower reporting does not demonstrate protection.
The checklist does not repair incomplete data, and this paper does not demonstrate improved patient outcomes from its use. Its practical value is a clearer reading test: does a drug safety claim describe a reporting pattern, or provide additional evidence supporting a causal interpretation?
Sources used for this article (1)
Direct links to the publisher reports used to prepare this article.
- Source 1
- The REporting of A Disproportionality Analysis for DrUg Safety Signal Detection Using Individual Case Safety Reports in PharmacoVigilance (READUS-PV): Explanation and Elaboration ['Michele Fusaroli', 'Francesco Salvo', 'Bernard Bégaud', 'Thamir M. Alshammari', 'Andrew Bate', 'Vera Battini', 'Andreas Brueckner', 'Gianmario Candore', 'Carla Carnovale', 'Salvatore Crisafulli', 'P
