Conditional estimation for generalized linear models when covariates are subject-specific parameters in a mixed model for longitudinal measurements.

Erning Li; Daowen Zhang; Marie Davidian

doi:10.1111/j.0006-341x.2004.00170.x

Home/Archive/PubMed

PubMed · 15032767

Conditional estimation for generalized linear models when covariates are subject-specific parameters in a mixed model for longitudinal measurements.

Erning Li·Daowen Zhang·Marie Davidian

View original source ↗Export metadata ↓

Abstract

The relationship between a primary endpoint and features of longitudinal profiles of a continuous response is often of interest, and a relevant framework is that of a generalized linear model with covariates that are subject-specific random effects in a linear mixed model for the longitudinal measurements. Naive implementation by imputing subject-specific effects from individual regression fits yields biased inference, and several methods for reducing this bias have been proposed. These require a parametric (normality) assumption on the random effects, which may be unrealistic. Adapting a strategy of Stefanski and Carroll (1987, Biometrika74, 703-716), we propose estimators for the generalized linear model parameters that require no assumptions on the random effects and yield consistent inference regardless of the true distribution. The methods are illustrated via simulation and by application to a study of bone mineral density in women transitioning to menopause.

Explore related subjects

Biometry Bone Density Bone Diseases, Metabolic Female Humans Linear Models Logistic Models Longitudinal Studies Menopause Middle Aged Models, Biological Progesterone Journal Article Research Support, U.S. Gov't, P.H.S.

Keep this discovery

Explore connections, maps & timelines

BibTeX RIS

Erning Li, Daowen Zhang, Marie Davidian. 2004. Conditional estimation for generalized linear models when covariates are subject-specific parameters in a mixed model for longitudinal measurements.. https://doi.org/10.1111/j.0006-341x.2004.00170.x

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related citations

PubMed2006

Multivariate adaptive regression splines: a powerful method for detecting disease-risk relationship differences among subgroups.

In a wide variety of medical research scenarios one is interested in the question whether regression curves differ for subgroups in the sample. Examples are gender differences in the effect of drug treatment or the study of genotype-environment interactions. To address this question exploratory techniques are often required because detailed knowledge concerning the shape of the regression curves and how that shape differs across subgroups is lacking. In this article we explored the power of two such exploratory techniques: multivariate adaptive regression splines (MARS) and least squares curve fitting using polynomials. For this purpose simulations were performed using linear, logistic, and complex non-linear curves. The power obtained from MARS was on average 1.4 times higher than with polynomials. It was shown that power was higher even if the regression curve was linear, that gains increased with the complexity of the curve, and that for highly non-linear curves model-free methods such as MARS might be the only alternative.

Biometry↗

PubMed2006

Properties of R(2) statistics for logistic regression.

Various R(2) statistics have been proposed for logistic regression to quantify the extent to which the binary response can be predicted by a given logistic regression model and covariates. We study the asymptotic properties of three popular variance-based R(2) statistics. We find that two variance-based R(2) statistics, the sum of squares and the squared Pearson correlation, have identical asymptotic distribution whereas the third one, Gini's concentration measure, has a different asymptotic behaviour and may overstate the predictivity of the model and covariates when the model is mis-specified. Our result not only provides a theoretical basis for the findings in previous empirical and numerical work, but also leads to asymptotic confidence intervals. Statistical variability can then be taken into account when assessing the predictive value of a logistic regression model.

Biometry↗

PubMed2006

National Biometry Audit II.

PURPOSE: To determine the change in compliance with the Royal College of Ophthalmologists biometry guidelines since the last National Audit 2 years ago and in particular to quantify the adoption of modern methods of axial length measurement and customization of A constants. METHOD: A structured telephone questionnaire of individuals who perform biometry in all eye departments in the United Kingdom. RESULTS: A biometrist was interviewed in 94 of the 178 United Kingdom Ophthalmology departments. Compared with 2 years ago, nurses alone perform biometry more frequently (67 vs 51%) and junior doctors less frequently (9 vs 15%). More biometrists now attend external training courses (45 vs 37%). The Royal College of Ophthalmologists recommended intraocular lens calculation formulae (SRK-T, Hoffer Q, and Holladay) are used more commonly (30 and 15%) and audit of prediction error is being performed more frequently (78 vs 71%). The routine use of a partial coherence laser interferometry has increased from 35 to 61% in United Kingdom Ophthalmology departments. Currently, only one United Kingdom department is routinely using immersion ultrasound biometry. 'A' constants are customized in 47% of departments. CONCLUSION: Over the last 2 years, there has been improved implementation of the Royal College of Ophthalmologists guidelines on biometry. It is essential that the Royal College of Ophthalmologists guidelines are updated to include current best practice of routine use of partial coherence laser interferometry or immersion biometry and customization of A constants. A benchmark standard of 85-90% of patients achieving a final postoperative refraction within 1 dioptre of the predicted should be established.

Biometry↗