東華大學圖書館 |

Language: English

Help

回圖書館首頁

手機版館藏查詢

Back

Switch To: Labeled | MARC Mode | ISBD

Estimating vaccine efficacy in the p...

Chu, Haitao.

Linked to FindBook

Google Book

Amazon

博客來

Estimating vaccine efficacy in the presence of correlated data and missing data.

Record Type:	Electronic resources : Monograph/item
Title/Author:	Estimating vaccine efficacy in the presence of correlated data and missing data./
Author:	Chu, Haitao.
Description:	104 p.
Notes:	Source: Dissertation Abstracts International, Volume: 64-02, Section: B, page: 0493.
Contained By:	Dissertation Abstracts International64-02B.
Subject:	Biology, Biostatistics. -
Online resource:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3080308

Estimating vaccine efficacy in the presence of correlated data and missing data.
Chu, Haitao.

Estimating vaccine efficacy in the presence of correlated data and missing data. - 104 p.

Source: Dissertation Abstracts International, Volume: 64-02, Section: B, page: 0493.

Thesis (Ph.D.)--Emory University, 2003.

Estimating vaccine efficacy in the presence of correlated data and missing data presents unique challenges. When there are multiple infectives in a transmission unit, the pairwise transmission between an infective and a susceptible can not be observed. We can only observe the overall transmission from all infectives to a susceptible. Previously, analyses either excluded transmission units with multiple infectives (Préziosi and Halloran, 2003) or ignored co-infectives (Fine, Clarkson and Miller, 1988). However, excluding transmission units with multiple infectives is statistically less efficient and ignoring co-infectives can lead to biased estimation. Through Bayesian hierarchical models, we showed how to estimate heterogeneous transmission and thus vaccine efficacy including the transmission units with multiple infectives in the analyses. In vaccine studies, specific diagnosis of a suspected case by culture or serology of the infectious agent is expensive and difficult. Vaccine efficacy estimates could be severely attenuated if only based on a nonspecific auxiliary outcome. Implementing validation sets in the study can correct the bias while maintaining statistical efficiency (Halloran and Longini, 2001). We compared the performance of a Bayesian method with two commonly used missing data methods, the mean score method (Pepe, Reilly and Fleming, 1994) and multiple imputation (Little and Rubin, 2002) in analysis of a field study of influenza vaccine and with simulations. We demonstrated the superiority of the Bayesian method when the sample size in the validation set is small and the vaccine is highly efficacious. We also propose the vaccine efficacy acceptability curve, defined as the posterior probability that the measure of vaccine efficacy <italic>VE</italic> ≥ <italic> k</italic> for each acceptable value <italic>k</italic>, to represent the uncertainty for the estimate of the vaccine efficacy graphically.Subjects--Topical Terms:

1018416
Biology, Biostatistics.

Estimating vaccine efficacy in the presence of correlated data and missing data.
LDR:02844nmm 2200277 4500 001 1858943
005 20041020095006.5
008 130614s2003 eng d
035 $a (UnM)AAI3080308
035 $a AAI3080308
040 $a UnM $c UnM
100 1 $a Chu, Haitao. $3 1946611
245 1 0 $a Estimating vaccine efficacy in the presence of correlated data and missing data.
300 $a 104 p.
500 $a Source: Dissertation Abstracts International, Volume: 64-02, Section: B, page: 0493.
500 $a Adviser: M. Elizabeth Halloran.
502 $a Thesis (Ph.D.)--Emory University, 2003.
520 $a Estimating vaccine efficacy in the presence of correlated data and missing data presents unique challenges. When there are multiple infectives in a transmission unit, the pairwise transmission between an infective and a susceptible can not be observed. We can only observe the overall transmission from all infectives to a susceptible. Previously, analyses either excluded transmission units with multiple infectives (Préziosi and Halloran, 2003) or ignored co-infectives (Fine, Clarkson and Miller, 1988). However, excluding transmission units with multiple infectives is statistically less efficient and ignoring co-infectives can lead to biased estimation. Through Bayesian hierarchical models, we showed how to estimate heterogeneous transmission and thus vaccine efficacy including the transmission units with multiple infectives in the analyses. In vaccine studies, specific diagnosis of a suspected case by culture or serology of the infectious agent is expensive and difficult. Vaccine efficacy estimates could be severely attenuated if only based on a nonspecific auxiliary outcome. Implementing validation sets in the study can correct the bias while maintaining statistical efficiency (Halloran and Longini, 2001). We compared the performance of a Bayesian method with two commonly used missing data methods, the mean score method (Pepe, Reilly and Fleming, 1994) and multiple imputation (Little and Rubin, 2002) in analysis of a field study of influenza vaccine and with simulations. We demonstrated the superiority of the Bayesian method when the sample size in the validation set is small and the vaccine is highly efficacious. We also propose the vaccine efficacy acceptability curve, defined as the posterior probability that the measure of vaccine efficacy <italic>VE</italic> ≥ <italic> k</italic> for each acceptable value <italic>k</italic>, to represent the uncertainty for the estimate of the vaccine efficacy graphically.
590 $a School code: 0665.
650 4 $a Biology, Biostatistics. $3 1018416
650 4 $a Health Sciences, Immunology. $3 1017716
650 4 $a Statistics. $3 517247
690 $a 0308
690 $a 0982
690 $a 0463
710 2 0 $a Emory University. $3 1017429
773 0 $t Dissertation Abstracts International $g 64-02B.
790 1 0 $a Halloran, M. Elizabeth, $e advisor
790 $a 0665
791 $a Ph.D.
792 $a 2003
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3080308