Besides the likelihood ratio, which statistics are mentioned as useful in evaluating test usefulness?

Besides the likelihood ratio, which statistics are mentioned as useful in evaluating test usefulness?

• A. Positive predictive value and negative predictive value
• B. Sensitivity and specificity
• C. Hazard ratio
• D. Accuracy

Answer: (A)

When evaluating a test in everyday practice, you want to know how a result translates into the chance a patient actually has the disease. Besides the likelihood ratio, the positive predictive value and negative predictive value provide those post-test probabilities: PPV is the probability of disease given a positive result, and NPV is the probability of no disease given a negative result. These measures are directly useful for clinical decisions, as they reflect real-world risk for the patient being seen and they incorporate disease prevalence. Sensitivity and specificity describe how well the test detects disease and its absence, but they don’t by themselves give the actual probability a patient has or doesn’t have the disease after testing in a given setting. Hazard ratio relates to time-to-event outcomes, and accuracy can be misleading when disease prevalence is extreme.

When evaluating a test in everyday practice, you want to know how a result translates into the chance a patient actually has the disease. Besides the likelihood ratio, the positive predictive value and negative predictive value provide those post-test probabilities: PPV is the probability of disease given a positive result, and NPV is the probability of no disease given a negative result. These measures are directly useful for clinical decisions, as they reflect real-world risk for the patient being seen and they incorporate disease prevalence. Sensitivity and specificity describe how well the test detects disease and its absence, but they don’t by themselves give the actual probability a patient has or doesn’t have the disease after testing in a given setting. Hazard ratio relates to time-to-event outcomes, and accuracy can be misleading when disease prevalence is extreme.