The use of appropriate robust statistical methods can provide important insights into the impact of pharmacological interventions. There is a long-standing, controversial, and unresolved debate regarding the association between antidepressant medication use and violent or aggressive behaviors. In particular, both suicide and homicide have been linked to antidepressant use. Various claims have been made that antidepressant use increases the risk of suicide and homicide, whereas counterclaims suggest that antidepressant use decreases these risks. Bouvy and Liem (2012) have entered this debate with an analysis of population-level, longitudinal data from the Netherlands. They report a significant negative linear association between lethal violence and prescriptions of antidepressants in the Netherlands. Their analysis, however, is flawed in numerous ways. So much so, in fact, that defensible conclusions cannot be made and use of this research to counter mounting evidence that antidepressants are linked with violent and aggressive behavior may have dangerous consequences.
- Causal relationships
- Linear regression