Evaluation of diagnostic sensitivity and specificity of tests for aquatic animal diseases

InstructorsIan Gardner (and others TBA)

This workshop will begin on September 2nd in the morning at 9:00 a.m. and continue through the afternoon to 4:30 p.m., with a break for lunch.  

Overview

The workshop will focus on aspects of diagnostic validation of tests based on the pathway described by the World Organisation for Animal Health (OIE) in its Aquatic Manual of Diagnostic Tests – see this website http://www.oie.int/standard-setting/aquatic-manual/access-online/. For the workshop, it will be assumed that bench/analytical validation (stage 1 of the OIE pathway) has been completed and the purpose(s) of the test (e.g. presumptive diagnosis, confirmatory diagnosis, or surveillance of healthy aquatic animals) has been defined.  Design, analysis and reporting considerations will be described.

Background includes for the morning session:

  • OIE context of “fitness-for-purpose” and validation pathway
  • Reporting guidelines
  • Design guidelines
  • Field & experimental studies
  • Avoiding biases
  • Statistical analysis
  • Binary (dichotomous) data (e.g. bacterial culture +/-)
  • Continuous data (e.g. CT values in real-time PCR)
  • Participant data and questions for discussion

Background includes for the afternoon session:

  • Medcalc software for data analysis
  • Example: Abalone herpesvirus
  • Overview of latent class analysis
  • Example: Nucleospora salmonis in rainbow trout
  • Discussion of design options and statistical analysis of data
  • Example: Francisella noatunensis in tilapia
  • Questions and answers and wrap-up

Learning objectives  

At the end of the workshop, participants will be able to:

  1. Describe the strengths and limitations of different sample types (reference samples of known status, field samples, and experimental challenge  samples) for use in diagnostic validation
  2. Be able  to develop a study design for diagnostic validation  of a test for a pathogen of interest including calculation of appropriate sample sizes
  3. Analyze binary test data assuming a perfect reference (gold) standard
  4. Understand the conceptual basis of data analysis in the absence of a perfect reference standard (latent class models)

This workshop will be restricted to 20 participants. Selection will be on a first come basis and will require all participants to provide their own laptop computer. This session will opens along with conference registration in January 2018.

The workshop will be provide 6 Hours of Continuing Education Credits for veterinarians, veterinary students and veterinary technicians.