Interpretation of diagnostic tests

Question 1

How can lab tests be used to diagnose?

Answer 1

Reference intervals - what we expect in healthy animals
Interpretative thresholds - diagnostic cut-offs
Positive/negative tests

Question 2

Define sensitivity and specificity

Answer 2

Sensitivity = proportion of animals with the disease that yield a positive test result (good screening test)
Specificity = proportion of animals that don’t have the disease that yield a negative test result (good confirmatory test)

Question 3

What groups of animals are required for testing sensitivity and specificity?

Answer 3

Diseased - used to derive sensitivity
Healthy - used to derive specificity

Question 4

What are the consequences of bad sensitivity?

Answer 4

Higher false negatives:
- diagnosis missed
- animal may re-present
- outbreak may worsen
- costs

Question 5

What are the consequences of bad specificity?

Answer 5

Higher false positives:
- diagnosis where disease is absent
- unnecessary life long therapy
- unnecessary euthanasia
- costs

Question 6

What is prevalence/pre-test probability?

Answer 6

the proportion of animals in the tested population that have the condition
Widespread healthy population screen for an infectious disease - low pre-test probability
Only testing animals for a disease that have several relevant clinical circumstances - high pre-test probability

Question 7

What are positive predicted values and negative predicted values?

Answer 7

PPV = the proportion of positive results that are actually affected cases
NPV = proportion of negative results that are actually non-affected cases

Question 8

Describe positive predicted vaule

Answer 8

The proportion of positive results that are actually affected cases
Tell you if a positive results is reliable
Specificity has the greatest influence on PPV
High specificity tests have good PPV even at low prevalence
Confidence to confirm

Question 9

Describe negative predicted values

Answer 9

The proportion of negative results that are actually non-affected cases
Sensitivity has the greatest influence on NPV
High sensitivity tests have good NPV even when prevalence is high
Confidence to rule out

Question 10

Why are endocrine tests not very good and how can they be improved?

Answer 10

Endocrine systems respond to the environment physiologically in way that looks similar to pathology of the endocrine system
e.g., Increased cortisol when stressed and increased cortisol in hyperadrenocorticism

Improved by using dynamic endocrine function tests:
- stimulate hormone production - test for hypofunction
- suppress hormone production - test for hyperfunction