Predictive Value Calculator (PPV/NPV)

How it is calculated

From the four cells of a test result table the tool computes the predictive values: PPV = TP ÷ (TP + FP) × 100 and NPV = TN ÷ (TN + FN) × 100. TP are true positives, FP false positives, TN true negatives and FN false negatives, all entered as counts of people or samples. Counts cannot be negative, there must be at least one positive result (TP + FP) and one negative result (TN + FN), or nothing is shown. Results are percentages with two decimals. PPV answers how often a positive result is right, NPV how often a negative result is right. Both depend on how common the condition is in the group tested.

Common mistakes

• Mixing up PPV with sensitivity. Sensitivity divides TP by all people who have the condition (TP + FN), while PPV divides by all positive results (TP + FP).
• Carrying PPV from one population to another. The same test has a lower PPV when the condition is rarer, so figures from a clinic do not transfer to general screening.
• Entering percentages or rates in the four boxes. They must be counts from the same table.
• Putting the false negatives and false positives in the wrong cells, which swaps the meaning of the two results.
• Entering an empty group, such as no positives at all. PPV is then undefined and the tool shows nothing.