Please use this identifier to cite or link to this item:
|Title:||Meta-Analysis: The Effect of the Drug Fosamax on Bone Mineral Density in Multi-Dose and Multi-Year Osteoporosis Clinical Studies|
|Other Titles:||Meta-Analysis: Effect of the Drug Fosamax on Osteoporosis|
|Abstract:||Meta-analysis of clinical studies has reached a stage of general acceptance in both the statistical and medical fields. In the last two decades, meta-analysis has become an increasingly popular statistical procedure designed to integrate the findings of several independent studies that address a related set of research questions. The explosion of this relatively new scientific method for evaluating clinical medicine provides insight regarding the effect of a treatment so that inferences can be made that extend far beyond the findings of primary studies. Two models for study-to-study variation may be used in a meta-analysis: the fixed-effects model assumes that the population treatment effect is a single fixed value, and the random-effects model assumes that the population treatment effect is a randomly distributed variable with its own mean and variance. A fixed-effects approach exists for multi-dose and multi-time clinical studies whereby the correlations between treatment effect estimates resulting from the common control group between doses and years are taken into account. The goal of this research was to conduct a meta-analysis to evaluate the effect of the drug FOSAMAX on bone mineral density in postmenopausal women with established osteoporosis. In particular, the fixed-effects approach was used to determine whether a dose effect and/or time-effect existed across nine independent studies for the 4 anatomical sites: Lumbar Spine, Femoral Neck, Total Body and Trochanter. Statistical inferences about the dose-and/or time-effect were guided by fitting and testing regression models for the estimated treatment effects. Lastly, the summary treatment effects and 95% confidence intervals were calculated for each dose and year for all 4 sites via the fixed-effects model. The results of the meta-analysis indicate that a dose-and time-effect exists across the studies observed. By fitting and testing numerous regression models for the estimated treatment effects, it is concluded that the "full" model (a model with a parameter for each dose) is required to model the data adequately when tested with various models collapsed with respect to either dose or year of follow-up. By testing a dose-by-time interaction model with the full model for each of the 4 sites, the results indicate that terms in the interaction model are required to explain the data for the Lumbar Spine site, and thus, dose-effect depends on year.|
|Appears in Collections:||Digitized Open Access Dissertations and Theses|
Files in This Item:
|cihon_frank_1999Dec_masters.pdf||2.58 MB||Adobe PDF||View/Open|
Items in MacSphere are protected by copyright, with all rights reserved, unless otherwise indicated.