Meta-analysis of migraine headache treatments: Combining information from heterogeneous designs

Migraine headache is a common condition in the United States for which a wi de range of drug and nondrug treatments are available. There is wide disagr eement about which treatments are most effective; meta-analysis of existing clinical trials can help to bring existing evidence to bear on this questi on. Conducting a meta-analysis is a challenging statistical problem because of the absence of a uniform accepted definition of headache syndromes, the diversity of treatments, and the heterogeneous and incomplete nature of pu blished information. The results of studies are summarized in various ways; most studies report continuous treatment effect?, for each treatment, but some report only differences in effectiveness for pairs of treatments, and others report only 2 x 2 contingency tables for dichotomized responses. In this article we present a hierarchical Bayesian grouped random-effects mode l for synthesizing evidence from several clinical trials comparing the effe ctiveness of commonly recommended prophylactic treatments for migraine head aches. We incorporate explicitly the relationships among the different clas ses of treatments and introduce latent auxiliary variables to create a comm on scale for combining information from studies that report results in very different forms. This model permits us to synthesize this heterogeneous in formation and to make inferences about treatment effects and the relative r anks of treatment without understating uncertainty. Estimation, ranking, mo del validation, and sensitivity analysis are all implemented through simula tion-based methods.