In analysis of longitudinal data, the variance matrix of the parameter estimates is usually estimated by the 'sandwich' method, in which the variance for each subject is estimated by its residual products. We propose smooth bootstrap methods by perturbing the estimating functions to obtain 'bootstrapped' realizations of the parameter estimates for statistical inference. Our extensive simulation studies indicate that the variance estimators by our proposed methods can not only correct the bias of the sandwich estimator but also improve the confidence interval coverage. We applied the proposed method to a data set from a clinical trial of antibiotics for leprosy.
BT - Statistics in medicine C1 - http://www.ncbi.nlm.nih.gov/pubmed/17701950?dopt=Abstract DA - 2008 Mar 30 DO - 10.1002/sim.3027 IS - 7 J2 - Stat Med LA - eng N2 -In analysis of longitudinal data, the variance matrix of the parameter estimates is usually estimated by the 'sandwich' method, in which the variance for each subject is estimated by its residual products. We propose smooth bootstrap methods by perturbing the estimating functions to obtain 'bootstrapped' realizations of the parameter estimates for statistical inference. Our extensive simulation studies indicate that the variance estimators by our proposed methods can not only correct the bias of the sandwich estimator but also improve the confidence interval coverage. We applied the proposed method to a data set from a clinical trial of antibiotics for leprosy.
PY - 2008 SP - 937 EP - 53 T2 - Statistics in medicine TI - Smooth bootstrap methods for analysis of longitudinal data. VL - 27 SN - 0277-6715 ER -