Skip to Main Content
Frequently Asked Questions
Submit an ETD
Global Search Box
Need Help?
Keyword Search
Participating Institutions
Advanced Search
School Logo
Files
File List
NicoletDeedraRae2006 mt.pdf (642.79 KB)
ETD Abstract Container
Abstract Header
Enhancement of SAS and R for meta-analysis of observational studies
Author Info
Nicolet, Deedra Rae
Permalink:
http://rave.ohiolink.edu/etdc/view?acc_num=osu1407409901
Abstract Details
Year and Degree
2006, Master of Science, Ohio State University, Public Health.
Abstract
Meta-analysis is used to get an overall effect estimate from a collection of previously conducted studies on the same subject. When this method first became popular, it was in the setting of clinical trials. Meta-analysis identifies homogeneity of effect estimates between studies. This type of analysis also takes into consideration confounding factors and decreases their effect on the estimate. Meta-analysis consists of searching the literature for relevant studies, deciding which studies to include in the analysis, extracting relevant information and analyzing the data. The goal of meta-analysis is to estimate the effect estimate from a collection of relevant studies. In order to find this estimate, we consider two possible models for the data. The data could be fit using a fixed-effects or random-effects model. The fixed-effects model has two popular methods of finding the estimate that we will consider here, the Mantel-Haenszel method and the confidence interval method. This model is based on the assumption that there is only a within study variance component. The random-effects model is the second model that we address and in this case, we examine the DerSimonian-Laird method for estimating the effect estimate. There is an additional variance component that is used in the random-effects model. This variance component addresses the variance between studies. Meta-analysis has been extensively used with clinical trials and observational studies. In available statistical software, the programs available for meta-analysis can only be used with clinical trials. The goal of this work was to modify the programs in R and SAS, so they would be suitable for use with observational studies. Using data on aspirin use and its effect on colon cancer, we demonstrate the use of the modified R and SAS code. The results presented here demonstrated how we extended the programs for meta¬-analysis of clinical trials to observational studies. As these are very common in medical studies, this development will allow additional analysis to be done when they are the object of a given study. The hope for this work is to provide researchers with programs suitable for meta-analysis with observational studies and clinical trials.
Committee
Randall Harris (Advisor)
Deborah Burr Doss (Advisor)
Pages
69 p.
Subject Headings
Public Health
Recommended Citations
Refworks
EndNote
RIS
Mendeley
Citations
Nicolet, D. R. (2006).
Enhancement of SAS and R for meta-analysis of observational studies
[Master's thesis, Ohio State University]. OhioLINK Electronic Theses and Dissertations Center. http://rave.ohiolink.edu/etdc/view?acc_num=osu1407409901
APA Style (7th edition)
Nicolet, Deedra.
Enhancement of SAS and R for meta-analysis of observational studies.
2006. Ohio State University, Master's thesis.
OhioLINK Electronic Theses and Dissertations Center
, http://rave.ohiolink.edu/etdc/view?acc_num=osu1407409901.
MLA Style (8th edition)
Nicolet, Deedra. "Enhancement of SAS and R for meta-analysis of observational studies." Master's thesis, Ohio State University, 2006. http://rave.ohiolink.edu/etdc/view?acc_num=osu1407409901
Chicago Manual of Style (17th edition)
Abstract Footer
Document number:
osu1407409901
Download Count:
208
Copyright Info
© 2006, all rights reserved.
This open access ETD is published by The Ohio State University and OhioLINK.