Uptodate Article: “Evaluating diagnostic tests”
Calculating likelihood ratios (LR) using sensitivity and specificity is a common practice in medical diagnosis and other fields. Here’s the step-by-step guide:

Positive Likelihood Ratio (LR+)
LR+ = Sensitivity / (1 – Specificity)

Negative Likelihood Ratio (LR-)
LR- = (1 – Sensitivity) / Specificity

Where:

1. Sensitivity: True Positive Rate (proportion of true positives)
2. Specificity: True Negative Rate (proportion of true negatives)

Example
Suppose a diagnostic test has:

1. Sensitivity: 0.9 (90% chance of positive test result if disease is present)
2. Specificity: 0.95 (95% chance of negative test result if disease is absent)

Calculating LR+
LR+ = 0.9 / (1 – 0.95)
= 0.9 / 0.05
= 18

Calculating LR-
LR- = (1 – 0.9) / 0.95
= 0.1 / 0.95
= 0.105

Interpreting Likelihood Ratios

1. LR+ | 1: Supports presence of disease
2. LR- < 1: Supports absence of disease
3. LR+ | 10 or LR- < 0.1: Strong evidence
4. LR+ < 0.1 or LR- | 10: Weak evidence

Real-World Applications
Likelihood ratios are used in:

1. Medical diagnosis
2. Financial risk assessment
3. Machine learning
4. Forensic science

Important Considerations

1. Prior probabilities: Incorporate prior probabilities to calculate posterior probabilities using Bayes’ theorem.
2. Model assumptions: Ensure the statistical model accurately reflects the underlying data.