|
Abstract:
|
Confounders often complicate the analysis of treatment effects in biomedical studies. Without proper handling, confounders may bring bias to the estimated effect of a non-randomized treatment. In a Parkinson's Disease study, it is of interest to evaluate the effect of Monoamine oxidase B (MAOB) inhibitor use on long-term clinical decline. However, MAOB inhibitor is not a randomized treatment in this clinical trial. In this project, we implemented several statistical methods for exploring and handling potential confounders, including multivariate regression adjustment, propensity score matching, propensity score weighting and propensity score adjustment. The results using different methods are consistent in this data analysis.
|
