Course Outline

Evaluation of the accuracy of diagnostic tests for animal diseases: tips, tricks and traps.

Instructor: Ian Gardner, Canada Excellence Research Chair (Aquatic Epidemiology), UPEI

 

8:00 – 8:30.      Introduction: OIE context of fitness for purpose; analytical and other criteria; what’s new in test evaluation (lecture)

 

8:30 – 10:00.   Test evaluation with a perfect reference standard (field data) – estimation of 95% confidence intervals, receiver-operating characteristic analysis for comparing continuous tests (lecture)

 

10:00 -10.30    Analysis of experimental challenge studies for diagnostic sensitivity – some tips (lecture and discussion)

 

10:30-10:45     Coffee break

 

10:45-12:15     Discussion of participant data and hands-on analysis (1.5 hr of discussion/computer lab using Medcalc)

 

12:15-13:00     Lunch

 

13:00- 14:00    Test evaluation without a perfect reference standard (field data) – when is it most necessary and why?;  basic principles underlying method, practical example using web-based software (TAGS), example reanalyzed from a Bayesian perspective using prior information (lecture)

 

14:00 – 16:00  Discussion of participant data sets and hands-on analysis ctd. (2 hr of discussion/computer lab using Medcalc/TAGS) including a 15 min coffee break at about 15:00

 

16:00 – 16:30  Design and reporting considerations for diagnostic test accuracy studies (30 min of lecture and discussion)

 

16:30 -17:00   Questions and answers; wrap-up (30 min of discussion)

 

Note: Participants should bring their own laptop to the course although sharing a laptop with another colleague is also acceptable. We will use the Medcalc software (www.medcalc.org – downloadable free for 15-day trial use) when there is a perfect reference standard and the TAGS software for evaluation of tests without a perfect reference standard (there will be a stand-alone app for this based on the R software). Background powerpoint presentations and a useful reference list covering the main topics will be provided approximately 1 week prior to the course so that much of the workshop can be used for hands-on analysis and discussion of test validation scenarios brought to the workshop by participants.