You need to know in advance anything about the disease / status of your patient / sample.
This is the gold standard. Either the sample is from a positive or negative patient / sample. So the gold test could be e.i. a clinical diagnosis, a pathological diagnosis.
These definitions leads to the set of "positive" and "negative" samples which you need to get you cut off (which should be the point where you reach the best differentiation between both populations). The is quite a helpful tool: The Youden-Index as the sum of specificity and sensitivity -1. If you make a XY-graph over the cut-off values (X) and the Youden-index on Y, you will get an optimum curve. (see also: Problems of Cut-Off Level Determination in Enzyme Immunoassays: The Case of TBE-ELISA], Klin. Lab., 39, 877 (1993)

## All Answers (2)

Fatkhanuddin Aziz· Gadjah Mada UniversityStephan T. Kiessig· Ruhr-Plasma-ZentrumThis is the gold standard. Either the sample is from a positive or negative patient / sample. So the gold test could be e.i. a clinical diagnosis, a pathological diagnosis.

These definitions leads to the set of "positive" and "negative" samples which you need to get you cut off (which should be the point where you reach the best differentiation between both populations). The is quite a helpful tool: The Youden-Index as the sum of specificity and sensitivity -1. If you make a XY-graph over the cut-off values (X) and the Youden-index on Y, you will get an optimum curve. (see also: Problems of Cut-Off Level Determination in Enzyme Immunoassays: The Case of TBE-ELISA], Klin. Lab., 39, 877 (1993)

Can you help by adding an answer?