|
Activity Number:
|
229
- Random Effect/Mixed Models
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, July 31, 2017 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Biometrics Section
|
|
Abstract #322609
|
|
|
Title:
|
Alternative Methods to Estimate Pooled Effect Size and Its Uncertainty in Meta-Analysis
|
|
Author(s):
|
Alok Dwivedi* and Mona Pathak and Sada Nand Dwivedi and Rakesh Shukla
|
|
Companies:
|
Texas Tech University Health Sciences Center El Paso and All India Institute of Medical Sciences and All India Institute of Medical Sciences and University of Cincinnati
|
|
Keywords:
|
Meta-analysis ;
Pooled effects ;
Uncertainty coefficient ;
Weight ;
Random effects ;
Conclusive studies
|
|
Abstract:
|
Meta-analysis is the standard statistical procedure for computing pooled effect size from multiple studies. In the meta-analysis, the weight is typically assigned for each study depending on precision of the estimate. The current methods of assigning weights ignore conclusiveness of the study and thus produce unreliable pooled estimate. We define that a study is conclusive if the 95% confidence interval does not overlap with the minimum significance limits otherwise inconclusive. We may obtain conclusive pooled effect size based on all inconclusive studies or number of unjustified significant studies using standard meta-analysis approach. The weight should reflect the value of evidence (conclusiveness and statistically precise estimate) in a particular study. To take into account this, we proposed an alternative weight scheme and uncertainty coefficients to obtain fixed and random effects pooled estimates. The proposed methods were compared using standard methods of meta-analysis. Meta-analysis based on our proposed methods provides more reliable pooled effect size, suggesting to apply weight and uncertainty coefficient according to the value of evidence within and across studies.
|
Authors who are presenting talks have a * after their name.
