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

Analysis of clustered matched-pair d...

Durkalski, Valerie Lynne.

Linked to FindBook

Google Book

Amazon

博客來

Analysis of clustered matched-pair data under an equivalence design.

Record Type:	Electronic resources : Monograph/item
Title/Author:	Analysis of clustered matched-pair data under an equivalence design./
Author:	Durkalski, Valerie Lynne.
Description:	116 p.
Notes:	Source: Dissertation Abstracts International, Volume: 62-10, Section: B, page: 4307.
Contained By:	Dissertation Abstracts International62-10B.
Subject:	Biology, Biostatistics. -
Online resource:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3029684
ISBN:	0493420495

Analysis of clustered matched-pair data under an equivalence design.
Durkalski, Valerie Lynne.

Analysis of clustered matched-pair data under an equivalence design. - 116 p.

Source: Dissertation Abstracts International, Volume: 62-10, Section: B, page: 4307.

Thesis (Ph.D.)--Medical University of South Carolina, 2001.

The clinical study design for the comparison of two diagnostic or screening procedures is well developed in the literature. The research goal is to design a study that will assess the accuracy (e.g., correct identification of diseased/non-diseased persons) of the new procedure and compare it with respect to a standard. When the response of interest is in a dichotomous form (e.g., disease, no disease), the McNemar test for matched-pair samples is used to determine if the proportion of correctly identified disease states is the same for both procedures. However, when a standard procedure is highly accurate, the development of a new procedure that is superior in terms of accuracy is not always of main interest. Rather, if the new procedure offers more appealing characteristics, such as outpatient-based services, non-invasiveness, low complication rates or low costs, the research goal may be to show that the difference in accuracy of the two procedures is within a clinically unimportant amount. A study designed to address this research objective is referred to as an equivalence study.

ISBN: 0493420495Subjects--Topical Terms:

1018416
Biology, Biostatistics.

Analysis of clustered matched-pair data under an equivalence design.
LDR:03220nmm 2200301 4500 001 1865159
005 20041216142218.5
008 130614s2001 eng d
020 $a 0493420495
035 $a (UnM)AAI3029684
035 $a AAI3029684
040 $a UnM $c UnM
100 1 $a Durkalski, Valerie Lynne. $3 1952619
245 1 0 $a Analysis of clustered matched-pair data under an equivalence design.
300 $a 116 p.
500 $a Source: Dissertation Abstracts International, Volume: 62-10, Section: B, page: 4307.
500 $a Chairman: Yuko Palesch.
502 $a Thesis (Ph.D.)--Medical University of South Carolina, 2001.
520 $a The clinical study design for the comparison of two diagnostic or screening procedures is well developed in the literature. The research goal is to design a study that will assess the accuracy (e.g., correct identification of diseased/non-diseased persons) of the new procedure and compare it with respect to a standard. When the response of interest is in a dichotomous form (e.g., disease, no disease), the McNemar test for matched-pair samples is used to determine if the proportion of correctly identified disease states is the same for both procedures. However, when a standard procedure is highly accurate, the development of a new procedure that is superior in terms of accuracy is not always of main interest. Rather, if the new procedure offers more appealing characteristics, such as outpatient-based services, non-invasiveness, low complication rates or low costs, the research goal may be to show that the difference in accuracy of the two procedures is within a clinically unimportant amount. A study designed to address this research objective is referred to as an equivalence study.
520 $a Extensions and adaptations to McNemar's test under equivalence designs for matched-pair data have received attention in the literature. However, if an individual contributes more than one response for each procedure, such as in a diagnostic procedure to identify colon polyps where one person can have more than one polyp, the assumption of independence between paired responses no longer holds. If the dependence among the paired responses within a subject is ignored, the variance of the parameter estimate is underestimated, leading to small p-values and false statistical conclusions. Equivalence designs for clustered matched-pair data are not available.
520 $a This research addresses the analysis of clustered matched-pair data under an equivalence design. The first paper introduces a test statistic that adjusts for clustering of matched-pair data and compares it to two currently available methods. The second paper applies this method to a clinical trial designed to compare two diagnostic procedures. The third paper adapts the proposed test statistic to an equivalence hypothesis and applies it to a clinical trial comparing the accuracy of two procedures for detecting abnormal parathyroid glands.
590 $a School code: 0122.
650 4 $a Biology, Biostatistics. $3 1018416
650 4 $a Statistics. $3 517247
690 $a 0308
690 $a 0463
710 2 0 $a Medical University of South Carolina. $3 700119
773 0 $t Dissertation Abstracts International $g 62-10B.
790 1 0 $a Palesch, Yuko, $e advisor
790 $a 0122
791 $a Ph.D.
792 $a 2001
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3029684