Lessons · Medicine foundations · positive predictive value
Positive predictive value: what a positive result is worth here
Positive predictive value is the share of positive results that turn out to be right: TP / (TP + FP). It reads across the row of positive results, not down a column.
Hone is a place to practise a career, one idea a day. This is one of its lessons, written out in full and free to read without an account.
What it is for
It is the question a person actually asks when a result comes back positive, and it is the one figure that changes when you move the same test from a group where the condition is common to a group where it is rare. Sensitivity does not move; this does.
How to think about it
Find the row of everyone the test called positive: the ones it got right plus the ones it got wrong. Divide the right ones by that row total. If you find yourself using false negatives, you have wandered into the wrong direction of the table, because a false negative was never called positive at all.
Worked example
PPV = TP / (TP + FP)The row of positive results, top and bottom.
True positives 90 and false positives 20, both givenThe two numbers that make up the positive row.
90 / (90 + 20) = 0.818, which is 81.8 %Eighty-two positives in every hundred are right, in this group.
Your turn
A table gives 40 true positives and 10 false positives. Write the line that gives the positive predictive value as a percentage.
40 / (40 + 10) × 100 = %
Solve one, graded on the server
The trap
Using false negatives in the denominator, which is sensitivity wearing a different name. Sensitivity is a fact about the test; positive predictive value is a fact about the test and the group it was used in, and the row you divide by is what keeps them apart.