|
Abstract:
|
Meta-analysis has been widely used to synthesize information from related studies to achieve reliable findings. However, in studies of rare events, the event counts are often low or even zero, and so standard meta-analysis methods may cause substantial bias in estimation. Recently, for the overall treatment effect based on a random-effects (RE) model, Bhaumik et al. showed that a simple average estimator with the continuity correction factor 0.5 (SA_0.5) is the least biased for large samples, and it has superior performance when compared with other commonly used estimators. However, the RE models used in previous work are restrictive. Under a general framework that explicitly allows treatment groups with unequal variability but assumes no direction, we prove that SA_0.5 is still the least biased for large samples. Meanwhile, we show that SA_0.5 fails to minimize the mean squared error. Under a new RE model that accommodates groups with unequal variability, we thoroughly compare the performance of various methods for both large and small samples via simulation, and draw conclusions about when to use which method. A data example of rosiglitazone is used to provide comparison.
|
