Bayesian Analysis

Question 1

When to use Bayesian

Answer 1

• carrier risk given # normal children
• carrier risk given negative test result (test not 100% sensitive)
• likelihood to have a disease after testing positive (test not 100% specific)
• adult-onset conditions
• non-penetrance

Question 2

Bayesian Set-Up

Answer 2

Carrier vs Noncarrier

Prior
Probability

Conditional
Probability (sensitivity and specificity)

Joint
Probability (P1 * C)

Posterior
Probability (normalized J)

Question 3

Sensitivity

Answer 3

is the test able to detect what it’s supposed to?

sensitivity = true positives/(true positives + false negatives)

Question 4

Specificity

Answer 4

is the test able to tease apart who is affected and who is not?

specificity = true negatives/(true negatives + false positives)

Question 5

Positive predictive value

Answer 5

if someone tests positive, are they truly positive?

PPV = true positives/(true positives + false positives)

Question 6

Negative predictive value

Answer 6

if someone tests negative, are they truly negative?

NPV = true negatives/(true negatives + false negatives)