Evaluation of factors affecting performance of ...
|Title||Evaluation of factors affecting performance of diagnostic tests for infectious salmon anaemia virus|
|Author(s)||Charles G. B. Caraguel|
|Institution||University of Prince Edward Island|
|Supervisor(s)||Larry Hammell, Ian Dohoo|
|Committee member(s)||Javier Sanchez, Ian Gardner, Peter Wright, Crawford Revie, Larry Hammell|
|Degree earned||Doctor of Philosophy|
|Place published||Charlottetown, P.E.I.|
|Abstract||To secure the safety of international trade of animal and their derived products, it is required that animals should be proven infection-free using a validated diagnostic test certified by the World Organization for Animal Health (OIE). Test validation involves a multi-step evaluation process to assess test accuracy and its fitness for a specific purpose. Contrary to common understanding, diagnostic accuracy is not constant for each test and may vary within and between populations according to the distribution of biological factors that influence the pathophysiology of the disease. Diagnostic test accuracy combines the precision (repeatability & reproducibility) and the trueness of the test (diagnostic sensitivity, DSe, & specificity, DSp). The objective of this research program was to extend diagnostic test evaluation methods by estimating accuracy specifically for influential covariate factors. This research was applied to reverse-transcriptase polymerase chain reaction (RT-PCR) for infectious salmon anaemia virus (ISAV). Early detection of ISAV is the base of an efficient control of this devastating disease for salmon aquaculture industries. RT-PCR accuracy was shown to differ among stages of infection, revealing substantial variation of test precision across prevalences of infection stages in the tested populations. Submission factors such as homogenization and testing laboratory also significantly impacted the precison and were accounted for to predict test result agreement across prevalences. Latent Class Modeling (LCM) was used to evaluate test trueness in absence of true status information. LCM assumes that DSe & DSp are constant across populations, which was revealed as an invalid assumption. Extending to 3-class LCM revealed different DSe between low- and high-infected salmon. Finally, the selection of the proper cutpoint for the real-time version of RT-PCR was dependent on the distribution of infection stages in the target population. Both analytical and epidemiological approaches to select the cutpoint were reviewed to improve the fit for intended purpose of the test. Potential applications for specific estimates of test accuracy and new perspectives on diagnostic test evaluation and use were discussed.|
|Use/Reproduction||In presenting this thesis in partial fulfillment of the requirements for a postgraduate
degree from the University of Prince Edward Island, I agree that the Libraries of this
University may make it freely available for inspection. I further agree that permission
for extensive copying of this thesis for scholarly purposes may be granted by the
professor or professors who supervised my thesis work, or, in their absence, by the
Chair of the Department or the Dean of the Faculty in which my thesis work was done.
It is understood any copying or publication or use of this thesis or parts thereof for
financial gain shall not be allowed without my written permission. It is also understood
that due recognition shall be given to me and to the University of Prince Edward Island
in any scholarly use which may be made of any material in my thesis.
Address: AVC-Centre for Aquatic Health Sciences & Department of Health Management
Atlantic Veterinary College
University of Prince Edward Island
Charlottetown, Prince Edward Island, Canada
Using APA 6th Edition citation style.
Times viewed: 881