Meta-analysis with application to estimating combined estimators of effect sizes in biomedical research.
Year: 2018
Type: dissertation
Study design: Systematic Review
Abstract: Meta-analysis is a statistical analysis that combines results from different independent \nstudies. In meta-analysis a number of statistical methods are currently used for \ncombining effect sizes of different studies. The simplest of these methods is based \non a fixed-effects model, which assumes that all studies in the meta-analysis share \na common true effect size and that the effect sizes in our meta-analysis differ only \nbecause of sampling error. Another statistical method that is used in meta-analysis, \nis the random-effects model, which assumes sampling variation due to fixed-effects \nmodel assumptions and random variation because the effect sizes themselves are \nsampled from a population of effect sizes. These models are compared to determine \nwhich model is appropriate and under what circumstances is the model appropriate. \nWe illustrate these models by applying each model to a collection of 3 studies examining \nthe effectiveness of new drug versus placebo to treat patients with duodenal \nulcers and meta-analysis of 9 studies of the use of diuretics during pregnancy to prevent \nthe development of pre-eclampsia. Results indicated that the choice between \nthe two model depends on the question of which model fits the distribution of effect \nsizes better and takes account of the relevant source(s) of error. We further study \nthe meta-analysis of longitudinal studies where effect sizes are reported at multiple \ntime points. Univariate meta-analysis is a statistical approach which may be used to \nstudy effect sizes reported at multiple time point. The problem with this approach \nis that it ignores correlation between the effect sizes, which might increase the standard \nerror of the point estimates. We used the linear mixed-effects model, which \nborrows ideas from multivariate meta-analysis. One of the advantages of the linear \nmixed-effects model is that it accounts for correlation between effect sizes both \nwithin and between studies. The independence model where separate univariate \nmeta-analysis is done at each of the time points was compared against models where \ncorrelation was accounted for different alternatives; including random study effects, \ncorrelated random time effects and/or correlated within-study errors, or unstructured \ncovariance structures. We implemented these methods through an example \nof meta-analysis of 16 randomized clinical trials of radiotherapy and chemotherapy \nversus radiotherapy alone for the post-operative treatment of patients with malignant gliomas, where in each trial, survival is evaluated at 6, 12, 18 and 24 months post randomization. The results revealed that models that accounted for correlations had better fit. \nKeywords: meta-analysis, fixed-effects model, random-effects model, heterogeneity, \npublication bias, linear mixed-effects model.
Author: Mhlengi Corrigan. Mgaga
Language: English
Subfield: Statistics, Probability and Uncertainty
Field: Decision Sciences
Domain: Social Sciences
Open Access status: green
OpenAlex ID: W3017978623