Background: Cervarix is a prophylactic vaccine used in preventing cervical cancer associated with human papillomavirus types 16/18. A previous study that investigated the risk of miscarriage associated with this bivalent vaccine by pooled analysis of data from two clinical trials, showed a numerically higher but statistically insignificant miscarriage rate in the HPV arm for pregnancies that began within three months following the vaccination. We explored this issue using an alternative statistical approach.
Objectives: To develop a hierarchical Bayes model to identify the potential time-dependent risk window of miscarriage rate associated with the bivalent HPV vaccine.
Methods: This study comprised the development of a hierarchical Bayes Model with its model inference and the application of this model to a real-world question. A multivariate logistic model was proposed that involved an indicator variable to accommodate a risk window with lower and higher cut-off points. Gibbs Sampling algorithms were used for the inference on the parameters of interest. Over ninety sets of simulation studies were conducted to evaluate the performance of the proposed method and estimate the power in detecting the risk effect. The Bayesian approach was then compared to the existing traditional approaches (e.g. the permutation test). The proposed model was applied to the subpopulation of pregnant women from the Costa Rica Vaccine Trial.
Results: In simulation studies, the Bayesian model demonstrated a better performance over the traditional approaches. It showed higher power than the traditional hypothesis testing in detecting the risk effect; it was more informative than the permutation test because it provided both the point estimates and the corresponding credible intervals for the cut-points and the ratio of odds ratios. In the analysis of the CVT data, we observed an effect of 1.13 (95% credible interval: 0.49 to 2.75), implying no significant evidence to support the hypothesis that HPV is associated with a higher miscarriage rate.
Conclusions: The hierarchical Bayes model can be applied to investigate the time-dependent risk of adverse events in clinical trials. Using the new Bayesian method, no significant risk of miscarriage in the Costa Rica Vaccine Trial was established, which is consistent with previous report. / Thesis (Master, Community Health & Epidemiology) -- Queen's University, 2013-01-30 23:20:39.859
| Identifer | oai:union.ndltd.org:LACETR/oai:collectionscanada.gc.ca:OKQ.1974/7790 |
| Date | 31 January 2013 |
| Creators | FANG, TIAN |
| Contributors | Queen's University (Kingston, Ont.). Theses (Queen's University (Kingston, Ont.)) |
| Source Sets | Library and Archives Canada ETDs Repository / Centre d'archives des thèses électroniques de Bibliothèque et Archives Canada |
| Language | English, English |
| Detected Language | English |
| Type | Thesis |
| Rights | This publication is made available by the authority of the copyright owner solely for the purpose of private study and research and may not be copied or reproduced except as permitted by the copyright laws without written authority from the copyright owner. |
| Relation | Canadian theses |
Page generated in 0.0017 seconds