What is a Good Positive Predictive Value?
What is a good positive predictive value?
When you are screened for a disease or infection, some people will test ‘positive’ (indicating they have the illness) and some will test ‘negative’ (indicating they don’t have the illness). If there were no false positives then it would be very helpful to know how many of those whose test result was ‘positive’ actually had the disease.
If, however, there were a lot of false positives then the screening programme would soon become less useful and, in fact, might be making matters worse. This is because, for a screening programme to be effective, the prevalence of the disease in the population being screened must be high enough to make the screen worthwhile and the tests used must have a good positive predictive value.
What are the advantages of sensitivity and specificity?
Using the 2×2 table method, sensitivity and specificity can be calculated from any set of two assays that have exactly the same test sensitivity and specificity. For example, a 2×2 table that has a 50% prevalence of glioma (the disease we want to test for) will have exactly the same sensitivity and specificity as the same 2×2 table with a 1% prevalence of glioma.
This means that, if the majority of the people who tested negative are actually negative for glioma, then we have a 99% chance that the person whose result was a positive has the disease! This is called a positive predictive value or PPV.
What is the effect of high sensitivity and high specificity?
The positive predictive value increases as the sensitivity of a test improves and fewer false positives are found. This is because more people with the disease will have a positive test result and there will be fewer false negatives. This is why the positive predictive value for a test goes up when a disease becomes more common in a population and lowers when the disease is not as prevalent.
What is the effect of higher disease prevalence?
A PPV is directly proportional to the disease prevalence. Therefore, if the disease prevalence in a study is higher, then the PPV is going to be higher and the NPV will be lower.
For example, if you are testing for a disease that has low disease prevalence then the NPV will be higher because more people with the disease are going to have a negative test result and there will be fewer false positives. On the other hand if the disease prevalence is very high then the PPV will be lower because there are more people with the disease and fewer false negatives.
How can I increase the positive predictive value of my test?
There are a few things that you can do to improve the positive predictive value of your test. The best way to do this is to try and avoid ordering a test that has a low sensitivity or low specificity.
Another way to increase the positive predictive value of your test is to order the test on groups of people that have a high likelihood of having the disease. For example, order a prostate specific antigen test on men over age 65 or on older men with a palpable nodule. You can also try to order the test on people with a symptom such as pain or a lump.