Are Causal Statements Reported in Pharmacovigilance Disproportionality Analyses Using Individual Case Safety Reports...

Area of Science:

• Pharmacovigilance
• Epidemiology
• Data Analysis

Background:

• Meta-epidemiological surveys reveal significant misinterpretation of disproportionality analyses.
• This study investigates the link between causal language strength in original disproportionality studies and their citing literature.

Purpose of the Study:

• To explore the relationship between causal statements in the title/abstract conclusions of pharmacovigilance disproportionality analyses and the causal language in citing studies.
• To quantify the strength of causal language using a four-level scale.

Main Methods:

• Selected 30 high-impact disproportionality studies and analyzed 1434 citing studies.
• Assessed causal statement strength in source articles and citing studies using a four-level scale (appropriate, ambiguous, conditionally causal, unconditionally causal).
• Employed multinomial regression models to evaluate associations between causal statement strengths.

Main Results:

• 27% of source studies used unconditionally causal titles, 30% used them in abstract conclusions.
• Only 20% of source studies used appropriate statements in both title and abstract conclusions.
• 45.8% of citing studies used unconditionally causal statements, and 26.4% used appropriate language, positively associated with source article conclusions (LogLRT p < 0.00001).

Conclusions:

• Nearly half of studies citing disproportionality analyses employ causal claims, especially when source articles use stronger causal language.
• A significant need exists for increased caution in the writing, interpretation, and citation of disproportionality analyses.
• Findings highlight potential for misinterpretation and overstatement of causality in pharmacovigilance research.