Bayesian modeling and inference for meta-data with applications in efficacy evaluation of an allergic rhinitis drug.

Allergic rhinitis is an allergic inflammation of the nasal membranes. The symptoms include disorders in nose and eyes. Studies have been carried out on safety and efficacy evaluation of triamcinolone acetonide aqueous nasal spray. To combine the results from different studies, we propose random-coefficient regression models. The properties of the proposed models are examined. The models are compared via the deviance information criterion (DIC), and Bayesian computations are carried out via MCMC sampling. A set of meta-data from nine clinical trials is analyzed in detail via the proposed methodology.