A Bayesian model for combining standardized mean differences and odds ratios in the same meta-analysis
Conference
64th ISI World Statistics Congress - Ottawa, Canada
Format: IPS Abstract
Keywords: meta-analysis
Session: IPS 94 - Advances in Research Synthesis in Healthcare Research
Tuesday 18 July 10 a.m. - noon (Canada/Eastern)
Abstract
Meta-analysts frequently face studies that report the same outcome differently, such as a continuous or binary variable. To combine these two types of results, a simple conversion method has been widely used to handle standardized mean differences (SMDs) and odds ratios (ORs). This conventional method uses a linear function connecting the SMD and log OR and assumes logistic distributions for continuous measures. However, the normality assumption is more commonly used, and the conventional method may be inaccurate when effect sizes are large or cutoff values are extreme. This talk presents a Bayesian hierarchical model to synthesize SMDs and ORs without using the conventional conversion. Our model assumes exact likelihoods for continuous and binary outcome measures that account for full uncertainties in the synthesized results. We performed simulation studies to compare the performance of the conventional and Bayesian methods. The Bayesian method generally produced less biased results with smaller mean squared errors and higher coverage probabilities than the conventional method in most cases. We used case studies to illustrate the proposed Bayesian method in real-world settings.