|
Abstract:
|
Model-based standardization uses a statistical outcome model or exposure model to estimate a population-average association that is unconfounded by selected covariates. We developed an approach based on an outcome model, in which the mean of the outcome is modeled conditional on the exposure and the confounders. In our approach, there is a confounder that clusters the observations into a large number of categories. We treated the parameters for the clusters as random effects. We used a between-within generalized linear mixed model (GLMM) to account for the association of the random effects with the exposure variables and the cluster sizes. We reviewed alternative approaches in the literature, and we compared our approach to recently proposed exposure-modeling approaches. We illustrated our method with simulation studies and we applied it to a data example with application in medical research. We used 2014 Truven Health MarketScan Research data to compare proportions of Acute Respiratory Tract Infection (ARTI) diagnoses with antibiotic prescription for emergency department versus outpatient visit, adjusting for confounding by cluster and measured diagnosis level confounders.
|
