Covariate adjustment and estimation of mean response in randomised trials

Research output: Contribution to journalArticle

2 Citations (Scopus)
14 Downloads (Pure)

Abstract

Analyses of randomised trials are often based on regression models which adjust for baseline covariates, in addition to randomised group. Based on such models, one can obtain estimates of the marginal mean outcome for the population under assignment to each treatment, by averaging the model-based predictions across the empirical distribution of the baseline covariates in the trial. We identify under what conditions such estimates are consistent, and in particular show that for canonical generalised linear models, the resulting estimates are always consistent. We show that a recently proposed variance estimator underestimates the variance of the estimator around the true marginal population mean when the baseline covariates are not fixed in repeated sampling and provide a simple adjustment to remedy this. We also describe an alternative semiparametric estimator, which is consistent even when the outcome regression model used is misspecified. The different estimators are compared through simulations and application to a recently conducted trial in asthma.

Original languageEnglish
Pages (from-to)648-666
Number of pages19
JournalPharmaceutical Statistics
Volume17
Issue number5
Early online date11 Jul 2018
DOIs
Publication statusPublished - 1 Sep 2018

Fingerprint

Randomized Trial
Covariates
Baseline
Adjustment
Estimator
Regression Model
Estimate
Population
Asthma
Linear Models
Variance Estimator
Empirical Distribution
Generalized Linear Model
Averaging
Assignment
Model-based
Prediction
Alternatives
Simulation
Model

ASJC Scopus subject areas

  • Statistics and Probability
  • Pharmacology
  • Pharmacology (medical)

Cite this

Covariate adjustment and estimation of mean response in randomised trials. / Bartlett, Jonathan W.

In: Pharmaceutical Statistics, Vol. 17, No. 5, 01.09.2018, p. 648-666.

Research output: Contribution to journalArticle

@article{8d7905245e6a4dbaba19eecde5908b6c,
title = "Covariate adjustment and estimation of mean response in randomised trials",
abstract = "Analyses of randomised trials are often based on regression models which adjust for baseline covariates, in addition to randomised group. Based on such models, one can obtain estimates of the marginal mean outcome for the population under assignment to each treatment, by averaging the model-based predictions across the empirical distribution of the baseline covariates in the trial. We identify under what conditions such estimates are consistent, and in particular show that for canonical generalised linear models, the resulting estimates are always consistent. We show that a recently proposed variance estimator underestimates the variance of the estimator around the true marginal population mean when the baseline covariates are not fixed in repeated sampling and provide a simple adjustment to remedy this. We also describe an alternative semiparametric estimator, which is consistent even when the outcome regression model used is misspecified. The different estimators are compared through simulations and application to a recently conducted trial in asthma.",
author = "Bartlett, {Jonathan W}",
note = "{\circledC} 2018 John Wiley & Sons, Ltd.",
year = "2018",
month = "9",
day = "1",
doi = "10.1002/pst.1880",
language = "English",
volume = "17",
pages = "648--666",
journal = "Pharmaceutical Statistics",
issn = "1539-1604",
publisher = "John Wiley and Sons Inc.",
number = "5",

}

TY - JOUR

T1 - Covariate adjustment and estimation of mean response in randomised trials

AU - Bartlett, Jonathan W

N1 - © 2018 John Wiley & Sons, Ltd.

PY - 2018/9/1

Y1 - 2018/9/1

N2 - Analyses of randomised trials are often based on regression models which adjust for baseline covariates, in addition to randomised group. Based on such models, one can obtain estimates of the marginal mean outcome for the population under assignment to each treatment, by averaging the model-based predictions across the empirical distribution of the baseline covariates in the trial. We identify under what conditions such estimates are consistent, and in particular show that for canonical generalised linear models, the resulting estimates are always consistent. We show that a recently proposed variance estimator underestimates the variance of the estimator around the true marginal population mean when the baseline covariates are not fixed in repeated sampling and provide a simple adjustment to remedy this. We also describe an alternative semiparametric estimator, which is consistent even when the outcome regression model used is misspecified. The different estimators are compared through simulations and application to a recently conducted trial in asthma.

AB - Analyses of randomised trials are often based on regression models which adjust for baseline covariates, in addition to randomised group. Based on such models, one can obtain estimates of the marginal mean outcome for the population under assignment to each treatment, by averaging the model-based predictions across the empirical distribution of the baseline covariates in the trial. We identify under what conditions such estimates are consistent, and in particular show that for canonical generalised linear models, the resulting estimates are always consistent. We show that a recently proposed variance estimator underestimates the variance of the estimator around the true marginal population mean when the baseline covariates are not fixed in repeated sampling and provide a simple adjustment to remedy this. We also describe an alternative semiparametric estimator, which is consistent even when the outcome regression model used is misspecified. The different estimators are compared through simulations and application to a recently conducted trial in asthma.

UR - http://www.scopus.com/inward/record.url?scp=85060020423&partnerID=8YFLogxK

U2 - 10.1002/pst.1880

DO - 10.1002/pst.1880

M3 - Article

C2 - 29998483

VL - 17

SP - 648

EP - 666

JO - Pharmaceutical Statistics

JF - Pharmaceutical Statistics

SN - 1539-1604

IS - 5

ER -