Meta-Analysis Resources

Tools for Those Who Summarize the Evidence Base

Resources and networking for those who conduct or interpret meta-analyses related to any phenomenon that is gauged in multiple studies.

The importance of effect size calculations in any meta-analytic study is of utmost importance. However, despite this importance effect size accuracy is an area of meta-analysis that receives great scrutiny. The apples and oranges criticism of meta-analysis makes it crucial for analysts to accurately measure effect sizes so ensure good quality science. Recently, Hon-Cheong So and Pak C. Sham out of the University of Hong Kong looked at effect size measures in genetic association studies (GWAS) and age-conditional risk prediction. (Effect Size Measures in Genetic Association Studies and Age-Conditi...) So, clarified the relationship between odds ratio (OR), relative risk (RR) and incidence rate ratios (IRR) in context with GWAS studies and demonstrated that sampling prevalent cases and controls, OR approximates IRR. So did this by constructing a framework to compute breast cancer and Alzheimer’s disease risk given the participant’s current age and follow up period with consideration of competing risks of mortality (i.e. Cardiovascular Disease). So found that simply multiplying the OR by the average lifetime risk estimates yielded a final estimate > 100%, while So’s more all encompassing method accounted for competing risks that produced an estimate of 63%. So concluded that companies offering direct consumer genetic testing should enact more rigorous algorithms that consider competing risks.  With So’s recent findings among GWAS studies and age-conditional risk prediction using competing variables, meta-analysts who conduct health related analyses, especially in the genetic sciences should strive to incorporate competing variables in order to gain a more focused and accurate effect size statistic. See below for full pdf.

Views: 281

Attachments:

Replies to This Discussion

Hey Mike thanks for this. This is actually the kind of thing I was wondering about; how can we compare effect sizes in genetics studies when the wide variety of approaches to genetics studies (single SNP candidate studies, high-throughput GWAS, approaches intermediate to these two extremes) often makes this such 'apples and oranges' as you put it.

 

We should remember Thursday to ask about having some class time devoted to genetics meta-analysis at some point this semester.

 

Yes, thanks. It will still help to have a diverse sample of studies to see how, specifically, ES can be calculated.

Garrett Ash said:

Hey Mike thanks for this. This is actually the kind of thing I was wondering about; how can we compare effect sizes in genetics studies when the wide variety of approaches to genetics studies (single SNP candidate studies, high-throughput GWAS, approaches intermediate to these two extremes) often makes this such 'apples and oranges' as you put it.

 

We should remember Thursday to ask about having some class time devoted to genetics meta-analysis at some point this semester.

 

RSS

© 2022   Created by Blair T. Johnson.   Powered by

Badges  |  Report an Issue  |  Terms of Service