Efficient quantile marginal regression for longitudinal data with dropouts
Authors:
Open Access
- 7 March 2016
- journal article
- research article
- Published by Oxford University Press (OUP) in Biostatistics
- Vol. 17 (3) , 561-575
Abstract
In many biomedical studies independent variables may affect the conditional distribution of the response differently in the middle as opposed to the upper or lower tail. Quantile regression evaluates diverse covariate effects on the conditional distribution of the response with quantile-specific regression coefficients. In this paper, we develop an empirical likelihood inference procedure for longitudinal data that accommodates both the within-subject correlations and informative dropouts under missing at random mechanisms. We borrow the matrix expansion idea of the quadratic inference function and incorporate the within-subject correlations under an informative working correlation structure. The proposed procedure does not assume the exact knowledge of the true correlation structure nor does it estimate the parameters of the correlation structure. Theoretical results show that the resulting estimator is asymptotically normal and more efficient than one attained under a working independence correlation structure. We expand the proposed approach to account for informative dropouts under missing at random mechanisms. The methodology is illustrated by empirical studies and a real-life example of HIV data analysis.Keywords
Funding Information
- National Science Foundation (DMS-1007535)
- NSA (H98230-15-1-0260)
This publication has 24 references indexed in Scilit:
- Efficient estimation for longitudinal data by combining large-dimensional moment conditionsElectronic Journal of Statistics, 2015
- Operator norm consistent estimation of large-dimensional sparse covariance matricesThe Annals of Statistics, 2008
- High dimensional covariance matrix estimation using a factor modelJournal of Econometrics, 2008
- Estimation of Large Covariance Matrices of Longitudinal Data With Basis Function ApproximationsJournal of Computational and Graphical Statistics, 2007
- Quantile regression for longitudinal data using the asymmetric Laplace distributionBiostatistics, 2006
- Covariance matrix selection and estimation via penalised normal likelihoodBiometrika, 2006
- Median regression for longitudinal dataStatistics in Medicine, 2003
- Bayesian Latent Variable Models for Median Regression on Multiple OutcomesBiometrics, 2003
- Quasi-Likelihood for Median Regression ModelsJournal of the American Statistical Association, 1996
- Zidovudine Compared With Didanosine in Patients With Advanced HIV Type I Infection and Little or No Previous Experience With ZidovudineArchives of internal medicine (1960), 1995