Median bias reduction in random-effects meta-analysis and meta-regression

@article{Kyriakou2019MedianBR,
  title={Median bias reduction in random-effects meta-analysis and meta-regression},
  author={Sophia Kyriakou and Ioannis Kosmidis and Nicola Sartori},
  journal={Statistical Methods in Medical Research},
  year={2019},
  volume={28},
  pages={1622 - 1636}
}
The reduction of the mean or median bias of the maximum likelihood estimator in regular parametric models can be achieved through the additive adjustment of the score equations. In this paper, we derive the adjusted score equations for median bias reduction in random-effects meta-analysis and meta-regression models and derive efficient estimation algorithms. The median bias-reducing adjusted score functions are found to be the derivatives of a penalised likelihood. The penalised likelihood is… 
Efficient implementation of median bias reduction
In numerous regular statistical models, median bias reduction (Kenne Pagui et al., 2017) has proven to be a noteworthy improvement over maximum likelihood, alternative to mean bias reduction. The
Efficient implementation of median bias reduction with applications to general regression models.
In numerous regular statistical models, median bias reduction (Kenne Pagui et al., 2017) has proven to be a noteworthy improvement over maximum likelihood, alternative to mean bias reduction. The
On the Properties of Simulation-based Estimators in High Dimensions
TLDR
An algorithm is considered, namely the Iterative Bootstrap (IB), to efficiently compute simulation-based estimators by showing its convergence properties, and the properties of simulation- based estimators are proved, more specifically the unbiasedness, consistency and asymptotic normality when the number of parameters is allowed to increase with the sample size.
Phase transition unbiased estimation in high~dimensional settings
TLDR
A new property is introduced to assess the bias, called phase transition unbiasedness, which is weaker than unbiasedness but stronger than asymptotic results, and is computed using an algorithm which is shown to converge exponentially fast.
Contribution to the discussion of “When should meta‐analysis avoid making hidden normality assumptions?”
The article by Jackson and White (2018) explores how the normal distribution is used in the different stages of meta-analysis, which is frequently not made transparent in meta-analytic applications.

References

SHOWING 1-10 OF 21 REFERENCES
Improving the accuracy of likelihood-based inference in meta-analysis and meta-regression
SummaryRandom-effects models are frequently used to synthesize information from different studies in meta-analysis. While likelihood-based inference is attractive both in terms of limiting properties
On random-effects meta-analysis.
TLDR
The asymptotic properties of the meta-analysis estimator and the joint maximum likelihood estimator when the number of studies is either fixed or increases at a slower rate than the study sizes are established and a surprising result is discovered: the former estimator is always at least as efficient as the latter.
Integrated Likelihood Inference in Small Sample Meta‐analysis for Continuous Outcomes
This paper proposes the use of the integrated likelihood for inference on the mean effect in small‐sample meta‐analysis for continuous outcomes. The method eliminates the nuisance parameters given by
Improved tests for a random effects meta-regression with a single covariate.
TLDR
This work examines the commonly used tests on the parameters in the random effects meta-regression with one covariate and proposes some new test statistics based on an improved estimator of the variance of the parameter estimates based on some theoretical considerations.
Median bias reduction of maximum likelihood estimates
For regular parametric problems, we show how median centering of the maximum likelihood estimate can be achieved by a simple modification of the score equation. For a scalar parameter of interest,
A likelihood approach to meta-analysis with random effects.
TLDR
It is concluded that likelihood based methods are preferred to the standard method in undertaking random effects meta-analysis when the value of sigma B2 has an important effect on the overall estimated treatment effect.
Random-effects meta-analysis: the number of studies matters
TLDR
The overall recommendation is to avoid the DerSimonian and Laird method when the number of meta-analysis studies is modest and prefer a more comprehensive procedure that compares alternative inferential approaches.
A re-evaluation of the ‘quantile approximation method’ for random effects meta-analysis
TLDR
It is shown that altering the study parameters, and in particular introducing changes to the distribution of the within-study variances, can have a dramatic impact on the resulting quantiles.
A comparison of statistical methods for meta-analysis.
TLDR
It is shown that the commonly used DerSimonian and Laird method does not adequately reflect the error associated with parameter estimation, especially when the number of studies is small, and three methods currently used for estimation within the framework of a random effects model are considered.
Bias reduction in exponential family nonlinear models
In Firth (1993, Biometrika) it was shown how the leading term in the asymptotic bias of the maximum likelihood estimator is removed by adjusting the score vector, and that in canonical-link
...
...