Related Experiment Videos
Do antidepressants cause suicidality in children? A Bayesian meta-analysis.
Eloise E Kaizar1, Joel B Greenhouse, Howard Seltman
1Department of Statistics, Carnegie Mellon University, Pittsburgh, PA 15217, USA.
Clinical Trials (London, England)
|June 16, 2006
Summary
The evidence linking antidepressant use to suicidality in children is weak. A Bayesian hierarchical model revealed increased risk for major depressive disorder and SSRIs, but overall causality remains uncertain.
Area of Science:
- Child and Adolescent Psychiatry
- Pharmacovigilance
- Biostatistics
Background:
- The U.S. Food and Drug Administration (FDA) mandated a black-box warning for antidepressants in children due to concerns about suicidal behavior or ideation (suicidality).
- A meta-analysis of randomized placebo-controlled trials found that 1.7% of 4487 children exhibited suicidality, though no completed suicides occurred.
Purpose of the Study:
- To investigate the impact of relaxing the FDA's assumption of equivalence across different drug formulations and psychiatric diagnoses on the risk estimate of suicidality.
- To re-evaluate the FDA's meta-analysis by incorporating potential underestimation of risk variance.
Main Methods:
- Employed a Bayesian hierarchical model to extend the FDA's analysis, allowing for study-level variability.
- Conducted extensive sensitivity analyses to assess the robustness of findings to different model assumptions.
Main Results:
- An association between antidepressant use and increased suicidality risk was observed in trials for major depressive disorder (OR 2.3) and with SSRIs (OR 2.2).
- No significant association was found in other trial subsets.
- The hierarchical model results were robust, but the FDA's original meta-analysis robustness to assumptions is less clear. Trial generalizability is limited by patient exclusion and short trial durations.
Conclusions:
- The evidence supporting a causal link between antidepressant use and childhood suicidality is weak due to model specification and interpretation issues.
- Bayesian hierarchical models offer advantages in meta-analysis by incorporating variability and enabling sensitivity analyses for regulatory decision-making.