How is Bayes' theorem applied in clinical decision-making?

Study for the Medical Council of Canada Qualifying Examination (MCCQE) I. Engage with flashcards and multiple-choice questions with detailed feedback. Prepare confidently!

Multiple Choice

How is Bayes' theorem applied in clinical decision-making?

Bayes' theorem in clinical decision-making updates the probability of disease after a test by starting with a pretest probability and adjusting it with the test’s performance. The post-test probability is derived from the pretest probability and the test’s sensitivity and specificity through likelihood ratios: for a positive result, LR+ = sensitivity/(1−specificity); for a negative result, LR− = (1−sensitivity)/specificity. This converts the prior probability into updated odds and then back to a probability. For example, with a pretest probability of 20%, sensitivity 90%, and specificity 85%, a positive result gives LR+ of 6, so the pretest odds (0.2/0.8 = 0.25) become post-test odds of 1.5, corresponding to about a 60% probability. A negative result gives LR− around 0.118, making the post-test probability drop to about 3%. This illustrates why the best description is updating after a test by combining pretest probability with the test’s sensitivity and specificity. The other options either miss the post-test update, rely on predictive values instead of the underlying test characteristics, or imply independence from test results.

Bayes' theorem in clinical decision-making updates the probability of disease after a test by starting with a pretest probability and adjusting it with the test’s performance. The post-test probability is derived from the pretest probability and the test’s sensitivity and specificity through likelihood ratios: for a positive result, LR+ = sensitivity/(1−specificity); for a negative result, LR− = (1−sensitivity)/specificity. This converts the prior probability into updated odds and then back to a probability.

For example, with a pretest probability of 20%, sensitivity 90%, and specificity 85%, a positive result gives LR+ of 6, so the pretest odds (0.2/0.8 = 0.25) become post-test odds of 1.5, corresponding to about a 60% probability. A negative result gives LR− around 0.118, making the post-test probability drop to about 3%.

This illustrates why the best description is updating after a test by combining pretest probability with the test’s sensitivity and specificity. The other options either miss the post-test update, rely on predictive values instead of the underlying test characteristics, or imply independence from test results.

More practice questions