Positive predictive value

Method

Proportion of test-positive individuals in whom the target condition is confirmed on the reference standard, expressed as a percentage. Population, underlying disease prevalence, index test, threshold, and reference standard recorded on the edge and not optional.

The method is part of the endpoint identity. Studies measuring the same quantity by a different method are held as separate endpoints.

Notes

Confused with pooled specificity. Both rise with better rule-in performance and both are percentages, but specificity is conditional on disease status and is portable across populations, while a predictive value depends on prevalence and therefore differs for the SAME test in every population pooled. A predictive value carried into another setting is a false claim about the test, which makes this the subtler error of the two: not a sign inversion but a transportability failure that produces a plausible wrong number. Also distinct from negative predictive value, its counterpart on the other arm of the same 2x2 table.