Positive Predictive Value (PPV)
Definition: The proportion of individuals detected by a system or test that actually have the disease of interest. Positive predictive value (PPV) addresses the question, "of those who were detected with the laboratory test or surveillance system, what proportion truly have disease?" (Detected cases may be cases that actually have the disease or they may be false positives.)
Low PPV
- Many of the detected cases are not true cases
- Less specific case definition
- May waste resources in investigating non-cases or false outbreaks
High PPV
- Most detected cases are true cases
- More specific case definition
Calculating PPV
Positive Predictive Value can be calculated using a 2x2 table and the following formula: PPV = A / (A+B).
| Disease Yes | Disease No | Total |
|
| Detected Yes | True + (A) | False + (B) | A+B |
| Detected No | False – (C) | True – (D) | C+D |
| A+C | B+D | A+B+C+D |
Example: In the diagrams below, red people truly have the disease; the outlined boxes those people whom the test or surveillance system detected, that is, identified, as having the disease. The diagrams below show a population of 50 people.
Diagram 1
Since all five people in the outlined box (those detected by the test or surveillance system) are red (have the disease), the PPV is 100%.
| Disease Yes | Disease No | Total | |
| Detected Yes | 5 | 0 | 5 |
| Detected No | 0 | 45 | 45 |
| 5 | 45 | 50 |
PPV calculation = A / (A+B) = 5 / (5+0) = 100%.
Diagram 2
In the diagram below, 10 people were detected with the disease (10 people in outlined box); of those 10 people, half truly had the disease (five people are red). So the PPV is = 50%.
Note that five people were detected as having the disease, when they truly didn't have the disease. These cases are called false positives.
| Disease Yes | Disease No | Total | |
| Detected Yes | 5 | 5 | 10 |
| Detected No | 0 | 40 | 40 |
| 5 | 45 | 50 |
PPV calculation = A / (A+B) = 5 / (5+5) = 5 / 10 = 50%.
Diagram 3
In the diagram below, three people were detected with the disease (three people in outlined box); of those three people, all truly had the disease (all are red). So the PPV is 100%.
Note that two people were not detected as having the disease.
| Disease Yes | Disease No | Total | |
| Detected Yes | 3 | 0 | 3 |
| Detected No | 2 | 45 | 47 |
| 5 | 45 | 50 |
PPV calculation = A / (A+B) = 3 / (3+0) = 3 / 3 = 100%.
