General Statistics

Question 1

Sensitivity

Answer 1

[True-positive] ÷ [true positive + false negative]

A test with high sensitivity has a high NPV: meaning that a negative test result most likely rules out the disease (both have low false negatives)

The proportion of people with the disease who have a positive test result

Question 2

Specificity

Answer 2

[True-negative] ÷ [true negative + false positive]

The proportion of people without the disease who have a negative test result

Question 3

Good screening test?

Answer 3

High SENSITIVITY, so that few cases are missed

Question 4

Positive predictive value

Answer 4

[True-positive] ÷ [all positive results]

Probability that someone with a positive result actually has the disease. An increase in the disease prevalence increases the PPv

Question 5

Negative predictive value

Answer 5

[True-negative] ÷ [all negative results]

Probability that someone with a negative result does not have the disease

A test with high sensitivity has a high NPV: meaning that a negative test result most likely rules out the disease (both have low false negatives)

Question 6

Likelihood Ratio of a positive test

Answer 6

sensitivity ÷ (1 - specificity)

Odds of getting a positive result in a diseased patient

Used to determine post-test probability of a test result

Question 7

Likelihood Ratio of a negative test

Answer 7

Odds of getting a negative result in an unaffected patient

Used to determine post-test probability of a test result

Question 8

Post-test probability

Answer 8

post-test odds ÷ (1 + post-test odds)

Question 9

Post-test odds

Answer 9

Pretest odds (disease prevalence) x likelihood ratio

Question 10

Coefficient of variation (CV)

Answer 10

Measure of precision. Accomplished by repeating a test approx 20 times using the same sample, and determining the range of variation

CV (%) = 100 x SD/mean

CV can vary over the assay’s analytical range, and should thus be determined at low, mid and high values.

Within run/ intra-assay: same sample, same day
Between run/inter-assay: same sample, different day

Question 11

Standard deviation (SD)

Answer 11

Measure of dispersion from the mean in a Gaussian normal distribution

Question 12

Area under the ROC curve?

Answer 12

Receiver Operating Characteristic (ROC) curve: graphic representation of sensitivities and specificities that are possible when the varying the test result cutoff value

Area under the curve? measure of the test’s overall ability to discriminate between disease and non-disease (accuracy)

Area of 1: perfect test
0.5: no ability to discriminate
>0.8: good discriminatory power