Statistics in medicine
-
Statistics in medicine · Jul 2012
Comparative StudyMultiple imputation for left-censored biomarker data based on Gibbs sampling method.
Biomarkers, increasingly used in biomedical studies for the diagnosis and prognosis of acute and chronic diseases, provide insight into the effectiveness of treatments and potential pathways that can be used to guide future treatment targets. The measurement of these markers is often limited by the sensitivity of the given assay, resulting in data that are censored either at the lower or at the upper limit of detection. For the Genetic and Inflammatory Markers of Sepsis (GenIMS) study, many different biomarkers were measured to examine the effect of different pathways on the development of sepsis. ⋯ We assume a multivariate normal distribution to account for the correlations between biomarkers and use the Gibbs sampler for the estimation of the distributional parameters and the imputation of the censored markers. We evaluate and compare the proposed methods with some simple imputation methods through simulation. We use a data set of inflammatory and coagulation markers from the GenIMS study for illustration.
-
Statistics in medicine · Jul 2012
Model misspecification and robustness in causal inference: comparing matching with doubly robust estimation.
In this paper, we compare the robustness properties of a matching estimator with a doubly robust estimator. We describe the robustness properties of matching and subclassification estimators by showing how misspecification of the propensity score model can result in the consistent estimation of an average causal effect. The propensity scores are covariate scores, which are a class of functions that removes bias due to all observed covariates. ⋯ The implication is that there are multiple possibilities for the matching estimator in contrast to the doubly robust estimator in which the researcher has two chances to make reliable inference. In simulations, we compare the finite sample properties of the matching estimator with a simple inverse probability weighting estimator and a doubly robust estimator. For the misspecifications in our study, the mean square error of the matching estimator is smaller than the mean square error of both the simple inverse probability weighting estimator and the doubly robust estimators.