
Bayesian influence diagnostics and outlier detection for metaanalysis of diagnostic test accuracy
Metaanalyses of diagnostic test accuracy (DTA) studies have been gainin...
read it

Bayesian modelbased outlier detection in network metaanalysis
In a network metaanalysis, some of the collected studies may deviate ma...
read it

Methods for the inclusion of real world evidence in network metaanalysis
Background: Network MetaAnalysis (NMA) is a key component of submission...
read it

Permutation inference methods for multivariate metaanalysis
Multivariate metaanalysis is gaining prominence in evidence synthesis r...
read it

MetaLearning for Relative DensityRatio Estimation
The ratio of two probability densities, called a densityratio, is a vit...
read it

Degree irregularity and rank probability bias in network metaanalysis
Network metaanalysis (NMA) is a statistical technique for the compariso...
read it

Network metaanalysis of rare events using penalized likelihood regression
Network metaanalysis (NMA) of rare events has attracted little attentio...
read it
Outlier detection and influence diagnostics in network metaanalysis
Network metaanalysis has been gaining prominence as an evidence synthesis method that enables the comprehensive synthesis and simultaneous comparison of multiple treatments. In many network metaanalyses, some of the constituent studies may have markedly different characteristics from the others, and may be influential enough to change the overall results. The inclusion of these "outlying" studies might lead to biases, yielding misleading results. In this article, we propose effective methods for detecting outlying and influential studies in a frequentist framework. In particular, we propose suitable influence measures for network metaanalysis models that involve missing outcomes and adjust the degree of freedoms appropriately. We propose three influential measures by a leaveonetrialout crossvalidation scheme: (1) comparisonspecific studentized residual, (2) relative change measure for covariance matrix of the comparative effectiveness parameters, (3) relative change measure for heterogeneity covariance matrix. We also propose (4) a modelbased approach using a likelihood ratio statistic by a meanshifted outlier detection model. We illustrate the effectiveness of the proposed methods via applications to a network metaanalysis of antihypertensive drugs. Using the four proposed methods, we could detect three potential influential trials involving an obvious outlier that was retracted because of data falsifications. We also demonstrate that the overall results of comparative efficacy estimates and the ranking of drugs were altered by omitting these three influential studies.
READ FULL TEXT
Comments
There are no comments yet.