A study on the effects of different levels of data on meta-analysis estimates /
This research sought to assess the implication of utilising studies with different levels of data, namely, the aggregate data (AD), individual patient data (IPD) and mixed of AD:IPD data (MD), on the overall meta-analysis estimates. Another primary focus of this research was to assess the performanc...
Saved in:
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English |
| Published: |
Kuantan, Pahang :
Kulliyyah of Science, International Islamic University Malaysia,
2017
|
| Subjects: | |
| Online Access: | Click here to view 1st 24 pages of the thesis. Members can view fulltext at the specified PCs in the library. |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | This research sought to assess the implication of utilising studies with different levels of data, namely, the aggregate data (AD), individual patient data (IPD) and mixed of AD:IPD data (MD), on the overall meta-analysis estimates. Another primary focus of this research was to assess the performance of the available methods that are used in combining the AD and IPD studies in a meta-analysis, particularly, the two-stage method and one-stage method, and to examine how the overall meta-analysis estimates are affected by combining the two levels of data. Simulation approach utilising R-software was used to generate the AD, IPD, and the combined AD:IPD level data. The assessments of the estimates were based on the mean bias, standard error, and root-mean-square error (RMSE). The findings revealed that the level of studies used in meta-analysis has some effects on the overall meta-analysis estimates. The studies that used individual data (IPD) have better estimates than those that used the AD studies. However, if IPD was used, there is a greater tendency of producing a false-positive conclusion. Additionally, the simulation results showed that the MD produced more accurate and more precise estimates than the AD. The results further showed that the method used to integrate the AD and IPD studies had some influence on the efficacy of the estimates. In general, the two-stage method gave comparable or better overall estimates in terms of accuracy and precision compared to the one-stage method for combining the AD and IPD studies. Whenever possible, IPD meta-analysis is recommended. However, as IPD is more difficult to access and more costly, combining the IPD and AD would be the better option than utilising AD alone. The two stage method is easier to implement and relatively more accurate than the one stage method. Therefore, the two-stage method is sufficient for combining the AD and IPD if the main interest of the study is to estimate the overall treatment effects. |
|---|---|
| Physical Description: | xii, 93 leaves : illustrations ; 30cm. |
| Bibliography: | Includes bibliographical references (leaves 85-87). |