The meeting will start Tuesday June 20 at 12.00 and end Thursday June 22 at 16.00.
|
Tuesday June 20 |
|
|
12:00 – 13:00 |
Talk: Opening/Introducing the Guidelines! Koen Mintiens - Introducing this workshop, the previous one. The objective of the coming 2½ days. - Introducing the guidelines and their environment: the context of OiE, their test validation for gold standard guidelines, their purpose of testing, outlining the elements of the guidelines, and setting up unresolved questions that this workshop will seek to answer.
|
|
13:00 – 13:45 |
Late lunch break |
|
13:45 – 14:30 |
Talk: Estimation of sensitivity and specificity of a diagnostic test in the absence of a gold standard Nils Toft - We will introduce the concepts involved in evaluation of diagnostic tests in the absence of a gold standard in order to 1) familiarize everyone with the latent class analysis in the context of the ‘classic’ Hui-Walter model; 2) outline the challenges involved in latent class analysis from an analytical as well as biological point of view; 3) introduce the problems with the assumption of conditional independence of diagnostic tests given disease status. The emphasis is put on examples and illustrations rather than equations.
|
|
14:30 – 15:15 |
Discussion Session 1: Addressing conditional independence and particularly the lack thereof. Chairman: Koen Mintiens.
Introductory talk: Estimation of Se and Sp of diagnostic tests when the assumption of conditional independence cannot be justified. Bas Engel
While the conditional independence assumption is central to the original Hui-Walter model, it is possible to formulate and analyze models where this assumption is replaced by an assumption of some form of conditional dependence between tests given disease status. This talk introduces approaches to address the lack of conditional independence…(keywords bias, identifiability, model selection, data requirements, Bayesian vs. frequentist etc.)
Discussion:
|
|
15:15 – 15:45 |
Coffee break |
|
15.45 – 16.30 |
Discussion Session 2: Addressing conditional independence and particularly the lack thereof. Chairman: Koen Mintiens.
Introductory talk: Biology vs. statistics when modeling conditional dependence in latent class analysis. Nils Toft
While the assumption of conditional independence is unique, there is more than one way to model conditional dependence between two tests. We briefly discuss a couple of these with respect to the interpretation, application and potential sample issues. Some examples of test evaluations are used to illustrate how the choice of tests influences the definition of disease in the latent class model and thus the Se and Sp of the tests.
Discussion
|
|
16.30 -17.15 |
Discussion session 3: Latent class analysis, model assumptions, data requirements, sampling and the purpose of testing paradigm. Chairman: Nils Toft
Introductory talk: Other issues that arise when performing no-gold standard diagnostic test evaluation studies. Koen Mintiens
Conditional independency is an important assumption when applying latent class analysis. Still other issues arise when planning and performing a no-gold standard diagnostic test evaluation study. These issues will briefly be presented and discussed.
|
|
17.30 - ??.?? |
Social programme (incl. dinner) |
|
Wednesday June 21 |
|
|
9:00-9.45 |
Talk: Challenges in analytical validation and quality assurance of microbiological assays. Koen Mintiens - An introduction to challenges and the tools and techniques used in the lab-validation element of diagnostic test evaluation.
|
|
9.45 – 10:30 |
Discussion Session 4: Examples of quality control Chairman: Kris De Clercq
- Introductory talk: Measurement of uncertainty of ELISA tests. Jean François Toussaint - - Discussion: accuracy and precision, validation of test methods, microbiological issues |
|
10:30 – 11:00 |
Coffee break |
|
11:00 – 11:45 |
Discussion Session 5: Examples of quality control Chairman: Kris De Clercq
- Introductory talk: Development of secondary standards for a foot-and-mouth disease ELISA and internal quality control using control charts. Nesya Goris
Discussion: Methods for quality control, selection of reference sera |
|
11:45 – 12:30 |
Discussion Session 6: Examples of quality control Chairman: Kris De Clercq
- Introductory talk: Performing proficiency testing for microbiological laboratories Katia Knapen
Discussion: Methods for quality control, selection of reference sera |
|
12:30 – 13:30 |
Lunch |
|
13:30 – 17:00
(incl. coffee break)
|
Case studies: Foot-and-mouth virus, Trichinellosis, Paratuberculosis. Chairman: Koen Mintiens
- Using a virus, a parasite and a bacteria as examples we are bound to discover something interesting about diagnostic test evaluation with and without a gold standard.
-Talk: Bayesian evaluation of diagnostic test for FMD for vaccinated and non-vaccinated cattle. Bas Engel
-Talk: Comparison of two 3ABC ELISA for FMD in an endemic, multiple serotype cattle population using gold standard as well a no-gold standard analysis. Nils Toft
-Talk: Diagnostic Test Evaluation For Trichinella Infections In Pigs Dorte Döpfer To evaluate the usefulness of serological tests applied to monitor Trichinella free herds, Bayesian methodology was used to estimate the diagnostic test parameters: sensitivity, specificity and prevalence in the absence of a Gold Standard test. In the absence of Dutch serum samples for positive pigs, it is discussed how the diagnostic test parameters derived from panels of sera originating from regions with endemic Trichinellosis in Argentina can be used to estimate the test parameters.
-Talk: Evaluation and use of diagnostic tests for paratuberculosis. Nils Toft.
-Discussion: When can/should latent class analysis be used? When is it inappropriate? Are there any important lessons from the examples? |
|
18:00 - |
Social event |
|
Thursday June 22 |
|
|
9:00 – 12:00 |
Lessons from the Case studies Chairman: Koen Mintiens
Groupwork on the guidelines! The participants are divided into groups and are given a set of the draft guidelines. The groups (of 6-8 people) are then asked to fill out a template for a given no-gold standard test evaluation for a given scenario, i.e. a given test, a disease, the known facts (such as available prescribed tests) and the purpose of the new test. |
|
12:00 -13:30 |
Lunch |
|
13:30 – 15:30 |
Closing discussion Chairman: Koen Mintiens - Summary of the problems and solutions discussed in the morning - Identified problems with the guidelines - The road ahead/ Training courses/ Manuals - Future meetings - Evaluation |
|
15:30 – 16:00 |
Coffee and departure! |
|
|
|