Which LR- range corresponds to moderate evidence to rule OUT disease?

Which LR- range corresponds to moderate evidence to rule OUT disease?

• A. <.1: strong (rules OUT)
• B. .1-.2: moderate evidence
• C. .2-.5: small evidence
• D. .5-1: little clinical evidence

Answer: (B)

Negative likelihood ratio shows how much a negative test lowers the chance of disease. The smaller the LR-, the stronger the test’s ability to rule out. In practice, these commonly used bands help interpret how convincing a negative result is: LR- less than 0.1 is considered strong evidence to rule out, 0.1 to 0.2 is moderate evidence, 0.2 to 0.5 is small evidence, and 0.5 to 1 provides little clinical value. Therefore, the range 0.1 to 0.2 corresponds to moderate evidence to rule OUT disease. To see why, think in terms of post-test odds. Post-test odds = pre-test odds × LR-. For example, a person with a pre-test probability of 20% has pre-test odds of 0.25. If the test is negative with an LR- of 0.15, the post-test odds become 0.25 × 0.15 = 0.0375, which translates to a post-test probability of about 3.6%. If the pre-test probability were 30%, the post-test probability would drop to around 6%. These substantial reductions illustrate why an LR- in the 0.1–0.2 range is viewed as moderate evidence to rule out disease.

Negative likelihood ratio shows how much a negative test lowers the chance of disease. The smaller the LR-, the stronger the test’s ability to rule out. In practice, these commonly used bands help interpret how convincing a negative result is: LR- less than 0.1 is considered strong evidence to rule out, 0.1 to 0.2 is moderate evidence, 0.2 to 0.5 is small evidence, and 0.5 to 1 provides little clinical value. Therefore, the range 0.1 to 0.2 corresponds to moderate evidence to rule OUT disease.

To see why, think in terms of post-test odds. Post-test odds = pre-test odds × LR-. For example, a person with a pre-test probability of 20% has pre-test odds of 0.25. If the test is negative with an LR- of 0.15, the post-test odds become 0.25 × 0.15 = 0.0375, which translates to a post-test probability of about 3.6%. If the pre-test probability were 30%, the post-test probability would drop to around 6%. These substantial reductions illustrate why an LR- in the 0.1–0.2 range is viewed as moderate evidence to rule out disease.