Screening/Diagnostic Tests

Question 1

Sensitivity

Answer 1

True positives/people with disease

Question 2

Specificity

Answer 2

True negative/people without disease

Question 3

False positive

Answer 3

1-specificity

Question 4

Positive predictive value

Answer 4

true positives/number who tested positive

Question 5

Negative predictive value

Answer 5

true negatives/number who tested negative

Question 6

Prior probability

Answer 6

probability that a randomly selected person would have the disease, essentially the prevalence

Question 7

Post test probability

Answer 7

probability that someone doesn’t have the disease after a negative test or opposite for positive test

Question 8

Likelihood ratio of positive test

Answer 8

LR+ = sensitivity / (1-specificity)
essentially change from pre-test odds to get post-test odds
post-test odds= pre-test x (LR+)
Should be more than 1, the higher the better

Question 9

LR-

Answer 9

(1-sensitivity)/ specificity

LR- should be less than 1, lower the better as looking to exclude people with true negatives

Question 10

Diagnostic Odds Ratio (DOR)

Answer 10

(True positive/false negative) / (false positive/true negative)
Useful for comparing multiple tests

Question 11

Accuracy for diagnostic tests

Answer 11

(True positives + true negatives) / total tested

Question 12

Importance of cut-off points for tests

Answer 12

High specificity but low sensitivity minimises false positives
High sensitivity but low specificity avoids false negatives

Question 13

Finding balance between sensitivity and specificity for tests

Answer 13

Receiver operating characteristic (ROC) curves plot sensitivity vs (1-specificity)
Point closest to top left corner, or maximising area under curve optimises both

Question 14

Ideal characteristics of screening test

Answer 14

Not too uncomfortable
Repeatable
Sensitive
Specific
Quick + easy to interpret

Question 15

Bias in diagnostics

Answer 15

Verification - e.g. gold standard tests only used to check those with a + result on new test
Review - When results of test is known before gold standard or vice versa
Inconsistencies in borderline tests

Question 16

Bias in screening

Answer 16

Selection - volunteers for screening may be health conscious
Lead-time - diagnosis earlier so longer between diagnosis and outcome, even though outcome has not been pushed back
Spectrum - results influenced by incorrect expectation that test will perform as well in different populations