Diagnostic Test Evaluation

Reliability – Precision
 Reliability is the ability of a test to measure something consistently – repeatedly
1. testing situations (test-retest reliability)
2. within a test (split-half reliability)
3. across judges (inter-rater reliability).

Validity – Accuracy

 Validity is the degree to which a test measures that which was intended (compared with gold
standard test)

1|Page
Sensitivity

 Sensitivity is the probability of correctly identifying a case of disease

 Sensitivity = TP/(TP + FN)
 Snout = If the test is (-)  can be sure that patient doesn’t have disease

Calculate TP and FN with formula

True positive = (sensitivity) X (Number of patients confirmed with disease)

False negative = (1-sensitivity) X (Number of Patients confirmed with disease)

Specificity
 Specificity is the probability of correctly identifying disease-free persons.
 Specificity = TN/(TN + FP)
 Spin = If the test is +  can be sure that patient has disease

Calculate TN and FP with formula

2|Page
3|Page
PPV

 The probability that a person with a positive test is a true positive  (has the disease)
 Also known as pretest probability of patient with a positive test is really positive with disease
 PPV is directly proportional to prevalence of disease in a population (common diseases have ↑
PPV)

 PPV = TP/(TP + FP)

NPV

 The probability that a person with a negative test is a true negative (does not have the disease)
 Also known as pretest probability of patient with a negative test is really negative with disease
 NPV is inversely proportional to prevalence of disease in a population (uncommon diseases have
↑ NPV)

 NPV = TN/(TN + FN)

4|Page
Graphic Analysis of Sensitivity and Specificity

5|Page
6|Page
7|Page
8|Page