Sensitivity and Specificity

Question 1

Sensitivity

Answer 1

SPIN - Ability of a diagnostic test to identify who has the disease. A highly sensitive test if negative rules out the disease. TP/TP+FN (cases/Diseased) - True positive rate

Question 2

Specificity

Answer 2

SPOUT - Ability of the diagnostic test to identify who does not have the disease. A highly specific test if Positive rules in the disease. True Negative/ Controls(non-diseased) - True negative rate

Question 3

Positive predictor value and Negative predictor value

Answer 3

PPV - What is the probability of having the disease if I test positive? (True positive/True positive + False positive)

NPV - What is the probaility of not having the disease if i test negative? (True Negative/ False Negative + True Negative)

Question 4

Relation between sensitivity & specificity vs Positive Predictor Value and Negative Predictor Value

Answer 4

Higher Sn → higher NPV (clinician is more confident that a negative test result rules out the disease)

Higher Sp → higher PPV (clinician can be more confident that a positive test rules in the diagnosis)

Question 5

Does sensitivity and specificity depend on the prevalance?

Answer 5

No a test result will give the same answers does not depend on the prevalance

Question 6

Does Positive predictor Value and Negative predictor value depend on prevalance?

Answer 6

Yes, when prevalance increases, PPV also increases but NPV decreases

Question 7

Likelihood Ratio positive

Answer 7

Probability of testing positive in the diseased individuals/ probability of testing positive in the non diseased individuals (LR +: Sensitivity/ 1 - specificity)

LR + should be more than 10 (why because we want the probability of testing positive in the diseased individuals to be higher)

Question 8

Liklihood Ratio Negative

Answer 8

Probability of testing negative in the diseased individuals/ probaility of testing negative in the non -diseased individuals (LR -: 1- Sensitivity/ Specificity)

LR value to be less than 0.1 (why because we want the probability of testing negative in the non-diseased individuals to be higher)

Question 9

Receiver Operating Characteristic Curve (ROC curve)

Answer 9

ROC analysis has become a popular method for evaluating the accuracy of medical diagnostic systems. The most desirable property of ROC analysis is that the accuracy indices derived from this technique are not distorted by fluctuations caused by the use of arbitrarily chosen decision criteria or cut-offs. The derived summary measure of accuracy, such as the area under the curve (AUC) determines the inherent ability of the test to discriminate between the diseased and healthy populations. For a test to be accurate the ROC curve needs to be closer to one. Sensitivity (true positive) in Y axis and 1-specifity of the false positives in the x axis.

Question 10

Screening test

Answer 10

Inexpenisve test administered in largerly asymptomatic or population at high risk. high sensitivity, simple test, indicates suspicion of disease

Question 11

Diagnostic tests

Answer 11

Confirm diagnosis, expensive, invasive, given to symptomatic individuals with positive screening test (first you do TB skin test as a screen if positive do a diagnostic TB test to confirm) ; high specificity