Please use this identifier to cite or link to this item:
http://hdl.handle.net/10773/27714
Title: | Meta-analysis of a very low proportion through adjusted wald confidence intervals |
Author: | Afreixo, V. Cruz, S. Freitas, A. Hernandez, M. A. |
Keywords: | Meta-analysis Proportion Adjusted wald confidence intervals Linear combination of binomial proportions |
Issue Date: | 3-Jul-2019 |
Publisher: | Crimson Publishers |
Abstract: | In this paper we will discuss the meta-analysis of one low proportion. It is well known, that there are several methods to perform the meta-analysis of one proportion, based on a linear combination of proportions or transformed proportions. However, in the context of a linear combination of binomial proportions has been proposed some approximate estimators with some improvements on low proportion estimation. In this paper we will show, with a simple adaptation, the possible contribution of several approximate adjusted Wald confidence intervals (CIs) for the meta-analysis of proportions. In the context of low proportions, a simulation study scenario is carried out to compare these CIs amongst themselves and with other available methods with respect to bias and coverage probabilities, using the fixed effect or the random-effects model. Pointing our interest in rare events (analogous for the abundant events) and taking into account the prevalence estimation of the Methicillin-resistant Staphylococcus aureus with mecc gene, we discuss the choice of the meta-analysis methods on this low proportion. The default meta-analysis methods of meta-analysis software programs are not always the best choice, in particular to the meta-analysis of one low proportion, where the methods including the adjusted Wald can outperform. |
Peer review: | yes |
URI: | http://hdl.handle.net/10773/27714 |
DOI: | 10.31031/OABB.2019.02.000545 |
ISSN: | 2578-0247 |
Appears in Collections: | CIDMA - Artigos DMat - Artigos PSG - Artigos |
Files in This Item:
File | Description | Size | Format | |
---|---|---|---|---|
Afreixo2019a.pdf | 542.7 kB | Adobe PDF | View/Open |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.