How do likelihood ratios affect clinical decision-making?

How do likelihood ratios affect clinical decision-making?

• A. They quantify how much a test result shifts the probability of disease, moving from pretest odds to post-test odds.
• B. They determine population prevalence.
• C. They are used to estimate sample sizes.
• D. They are equal to the p-value.

Answer: (A)

Likelihood ratios quantify how much a test result should change your estimate of whether a patient has the disease. They combine sensitivity and specificity into a single multiplier that you apply to your pretest odds to get post-test odds, then convert that back to probability. Use them by starting with your pretest probability (your initial clinical judgement), convert to odds, multiply by the appropriate LR (positive LR for a positive result, negative LR for a negative result), and convert back to probability. This updated probability then informs decisions about treatment, further testing, or ruling out the disease. For example, with a pretest probability of 30% (odds ≈ 0.43) and a positive test with LR+ of 5, post-test odds ≈ 2.14, a post-test probability of about 68%. If the test is negative with LR− of 0.2, post-test odds ≈ 0.086, a post-test probability of about 8%. This approach helps decide whether to proceed with therapy or additional testing, based on how informative the test is, and it isn’t about population prevalence, sample size, or p-values.

Likelihood ratios quantify how much a test result should change your estimate of whether a patient has the disease. They combine sensitivity and specificity into a single multiplier that you apply to your pretest odds to get post-test odds, then convert that back to probability.

Use them by starting with your pretest probability (your initial clinical judgement), convert to odds, multiply by the appropriate LR (positive LR for a positive result, negative LR for a negative result), and convert back to probability. This updated probability then informs decisions about treatment, further testing, or ruling out the disease. For example, with a pretest probability of 30% (odds ≈ 0.43) and a positive test with LR+ of 5, post-test odds ≈ 2.14, a post-test probability of about 68%. If the test is negative with LR− of 0.2, post-test odds ≈ 0.086, a post-test probability of about 8%.

This approach helps decide whether to proceed with therapy or additional testing, based on how informative the test is, and it isn’t about population prevalence, sample size, or p-values.