VP26 Comparing Statistical Methods For Meta-Analysis Of Rare Event Data

Abstract

INTRODUCTION:

We aimed to identify the validity and robustness of effect estimates for serious rare adverse events in clinical study reports of antidepressant trials, across different meta-analysis methods for rare binary events data (1,2).

METHODS:

Four serious rare adverse events (all-cause mortality, suicidality, aggressive behaviour and akathisia) were meta-analyzed using different methods (3). The Yusuf-Peto odds ratio (OR), which ignores studies with no events in the treatment arms, was compared with the alternative approaches of generalized linear mixed models (GLMM), conditional logistic regression, a Bayesian approach using Markov Chain Monte Carlo (MCMC) and a beta-binomial regression model.

RESULTS:

Though the estimates for the four outcomes did not change substantially across the different analysis methods, the Yusuf-Peto method underestimated the treatment harm and overstimated its precision, especially when the estimated odds ratio (OR) deviated greatly from 1. For example the OR for suicidality for children and adolescents was 2.39 (95 percent Confidence Interval, CI 1.32 to 4.33, using the Yusuf-Peto method), but increased to 2.64 (95 percent CI 1.33 to 5.26) using conditional logistic regression, to 2.69 (95 percent CI 1.19 to 6.09) using beta-binomial, to 2.73 (95 percent CI 1.37 to 5.42) using the GLMM and finally to 2.87 (95 percent CI 1.42 to 5.98) using the MCMC approach.

CONCLUSIONS:

The method used for meta-analysis of rare events data influences the estimates obtained and the exclusion of double zero-event studies can give misleading results. To ensure reduction of bias and erroneous inferences, sensitivity analyses should be performed using different methods and we recommend that the Yusuf-Peto approach should no longer be used. Other methods, in particular the beta-binomial method that was shown to be superior, should be considered instead.