Keywords: Antipsychotics; Bayesian modeling; Diffusion path; Markov Chain Monte Carlo; Profiling
New treatment technologies are the primary driver of spending growth in the United States with physicians playing a pivotal role in their adoption. Studies of physician prescribing behavior indicate that the placement of a particular physician on the adoption curve for one drug does not necessarily predict where that physician falls for other drugs. We propose a joint model to summarize the diffusion paths across therapeutically similar antipsychotics, in which the diffusion path of each antipsychotic is modeled by a semiparametric Poisson model with physician-specific random effects. The joint model is constructed by concatenating the univariate models based on a correlated random effects assumption. We use Markov Chain Monte Carlo methods to obtain posterior inferences. We propose a class of performance indices to identify fast adopters of antipsychotics based on posterior tail probabilities of relevant model parameters and determine which set of physicians’ covariates are related to the adoption patterns. Methods are illustrated by using dispensing information for 16932 physicians between January 1,1997 and December 31,2007 from the Xponent database, maintained by Quintiles IMS.