For a patient with intermediate probability, which testing characteristic provides the biggest diagnostic leverage?

For a patient with intermediate probability, which testing characteristic provides the biggest diagnostic leverage?

• A. Very high sensitivity only
• B. High specificity only
• C. High LR (diagnostic test), high LR+ or low LR- (big shift for or against disease)
• D. High accuracy across all settings

Answer: (C)

When you’re dealing with an intermediate pretest probability, the key idea is how much a test result changes the likelihood of disease. The most powerful way to do that is through likelihood ratios, which directly translate what you know before testing into what you should think after testing. A test with a high LR+ makes a positive result strongly supportive of disease, while a test with a very low LR- makes a negative result strongly supportive of no disease. These large shifts in probability, driven by the likelihood ratios, provide the biggest diagnostic leverage for a single test in a patient whose pretest probability sits in the middle. Sensitivity or specificity alone don’t tell you how much the post-test probability will change for an individual, and a test with good overall accuracy doesn’t guarantee a meaningful update for a given case. For example, with a 50% pretest probability, a test with LR+ around 10 can push the post-test probability up to about 90%, while an LR- around 0.1 can reduce it to around 10%. That substantial change is what makes likelihood ratios the most levered diagnostic characteristic in this setting.

When you’re dealing with an intermediate pretest probability, the key idea is how much a test result changes the likelihood of disease. The most powerful way to do that is through likelihood ratios, which directly translate what you know before testing into what you should think after testing.

A test with a high LR+ makes a positive result strongly supportive of disease, while a test with a very low LR- makes a negative result strongly supportive of no disease. These large shifts in probability, driven by the likelihood ratios, provide the biggest diagnostic leverage for a single test in a patient whose pretest probability sits in the middle. Sensitivity or specificity alone don’t tell you how much the post-test probability will change for an individual, and a test with good overall accuracy doesn’t guarantee a meaningful update for a given case. For example, with a 50% pretest probability, a test with LR+ around 10 can push the post-test probability up to about 90%, while an LR- around 0.1 can reduce it to around 10%. That substantial change is what makes likelihood ratios the most levered diagnostic characteristic in this setting.