A screening test for a disease has a likelihood ratio positive (LR+) of 8 and a likelihood ratio negative (LR−) of 0.1. Which statement is most accurate regarding this test's diagnostic utility?

Explanation

## Likelihood Ratios (LR+ and LR−) Interpretation **Key Point:** Likelihood ratios quantify how much a test result changes the odds of disease. They are independent of prevalence and directly applicable to Bayesian reasoning at the bedside. ### Interpretation Guidelines | LR+ Value | Interpretation | Clinical Utility | | --- | --- | --- | | > 10 | Strong evidence for disease | Excellent for ruling **in** | | 5–10 | Moderate evidence for disease | Good for ruling **in** | | 2–5 | Weak evidence for disease | Fair for ruling **in** | | 1–2 | Minimal evidence for disease | Poor for ruling **in** | | 1 | No diagnostic value | No change in odds | | LR− Value | Interpretation | Clinical Utility | | --- | --- | --- | | < 0.1 | Strong evidence against disease | Excellent for ruling **out** | | 0.1–0.2 | Moderate evidence against disease | Good for ruling **out** | | 0.2–0.5 | Weak evidence against disease | Fair for ruling **out** | | 0.5–1 | Minimal evidence against disease | Poor for ruling **out** | | 1 | No diagnostic value | No change in odds | ### Analysis of This Test - **LR+ = 8**: Falls in the 5–10 range → **Good to moderate evidence for disease** → useful for ruling **in** - **LR− = 0.1**: Falls in the < 0.1 range → **Strong evidence against disease** → excellent for ruling **out** **High-Yield:** This test is **excellent for both ruling in AND ruling out disease**. An LR+ of 8 meaningfully increases the post-test probability of disease, and an LR− of 0.1 meaningfully decreases it. ### Mnemonic **Mnemonic:** **SnNOut, SpPIn** - **Sn**Out: High **Sensitivity** rules **OUT** disease (LR− < 0.1) - **Sp**PIn: High **Specificity** rules **IN** disease (LR+ > 10) This test has both good sensitivity (implied by LR− = 0.1) and good specificity (implied by LR+ = 8). ### Clinical Pearl **Clinical Pearl:** Likelihood ratios are **prevalence-independent** and can be applied to any population. A negative test with LR− = 0.1 will always substantially reduce the probability of disease, regardless of starting prevalence. ### Relationship to Sensitivity and Specificity $$LR^+ = \frac{Sensitivity}{1 - Specificity}$$ $$LR^- = \frac{1 - Sensitivity}{Specificity}$$ An LR+ of 8 and LR− of 0.1 indicate a test with both high sensitivity and high specificity — the ideal diagnostic test.