Medical Statistics Principles: Sensitivity, Specificity, and Predictive Values

Intrinsic Test Properties (Independent of Prevalence)

Sensitivity and specificity are strictly defined as fixed, intrinsic properties of the test itself. They do not fundamentally change based directly on how common or rare the disease is within the test population.

• Sensitivity (True Positive Rate): Mathematically defined as the strict proportion of patients who actually possess the disease who are correctly identified mathematically by a positive test result ($TP / (TP + FN)$). A highly sensitive test (like D-dimer for pulmonary embolism) produces very few false negatives. Therefore, a negative result confidently rules OUT the disease (the 'SnNOut' mnemonic).
• Specificity (True Negative Rate): Mathematically defined as the strict proportion of genuinely healthy patients without the disease who are correctly identified completely by a negative test result ($TN / (TN + FP)$). A highly specific test produces very few false positives. Therefore, a positive result confidently rules IN the disease (the 'SpPIn' mnemonic).

Clinical Predictive Values (Dependent on Prevalence)

Predictive values are intensely critical for the practicing physician because they directly answer the clinical question: 'Given this specific test result, what is the mathematical probability that my patient actually has the disease?'

• Positive Predictive Value (PPV): The strict proportion of patients with a mathematically positive test result who genuinely actually possess the disease ($TP / (TP + FP)$). PPV is profoundly, completely reliant on the underlying prevalence of the disease. If you screen for an extremely rare cancer in a totally healthy, low-risk population, almost all the positive tests will mathematically be false positives, yielding a dismal PPV.
• Negative Predictive Value (NPV): The strict proportion of patients with a negative test result who genuinely truly do not unfortunately have the disease ($TN / (TN + FN)$). Similar to PPV, it relies heavily on prevalence. If a disease is exceptionally common in extreme epidemics, even a negative test is less reassuring.