Ainsworth C. Assessing the robustness of direct meta-analysis in the presence of heterogeneity. J Comp Eff Res. 2018 Oct;7(10):1009-25. doi: 10.2217/cer-2018-0024.

AIMS: To assess robustness of findings in recent systematic reviews of atypical antipsychotics in schizophrenia patients in the presence of heterogeneity.

METHODS: Relative efficacy was measured in seven direct comparisons of mean difference (MD) in average Positive and Negative Syndrome Scale (PANSS) total score and comparisons between conventional meta-analysis results and four alternate meta-analytic strategies using the difference in MD (∆MD).

RESULTS: MDs in PANSS total score were smaller in the conventional meta-analyses than those for three of the four analytic strategies. Such differences were small (all ∆MD <4). No considerable differences in effect size were observed in the limit meta-analysis comparison (|∆MD| <1).

CONCLUSIONS: The analyses validated the systematic review results and demonstrate the value of confirmatory sensitivity analysis.

Share on: