Assessing performance of diagnostic tests Flashcards
True prevalence
Proportion of truly infected individuals
Apparent prevalence
Proportion of test positive individuals
Accuracy
Proportion of infected and non-infected animals correctly classified by test
(TP + TN)/N
Sensitivity
The ability of a test to detect infected animals or proportion of infected animals that test positive
TP/I+
I+ = TP + FN
Specificity
Ability of a test to detect non-infected animals or proportion of non-infected animals that test negative
TN/I-
I- = TN + FP
Predictive values reflect
The way test results are used in the clinic/hospital/population
Predictive values are used
as a method for test selection
Predictive values are affected by
- SE
- SP
- I+ (disease prevalence)
Negative predictive value
Proportion of non-infected animals among those that test negative
TN/T-
T- = TN + FN
Positive predictive value
Proportion of infected animals among those that test positive
TP/T+
T+ = TP + FP
Use of positive predictive value:
- If the test result is positive, what’s the probability that this patient is infected?
- If we screen a population, what’s the proportion of animals who have the infection that will be correctly identified.
Selection of diagnostic tests to avoid intro of dz:
- High SE and NPV
- Important to reduce number of FN
Selection of diagnostic tests to confirm diagnosis:
- High SP and PPV
- Important to reduce number of FP
Selection of diagnostic tests to avoid unnecessary elimination of animals
- High SP and PPV
- Important to reduce number of FP
Testing in parallel
- The results of two or more tests must be negative
- Increases SE and NPV
