Timeline for Validate a medical test
Current License: CC BY-SA 3.0
6 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Mar 18, 2013 at 18:21 | comment | added | cbeleites | The widely used measures are: recovery rate (that's the bias which is still there after calibration - happens e.g. when calibration samples are prepared as solution, but real samples [and test samples] have some other matrix - usually marked in a plot of $c_{pred}$ over $c_{ref}$), RMSE as summary measure, dynamic range or calibration range. Usually also LOD and LOQ - these two have possible definitions using the relative error, so they are conveniently found in a plot of $sd (c_{ref})$ over $c_{ref}$ or $\frac{sd (c_{ref})}{c_{ref}}$ over $c_{ref}$ [$c$ as usual symbol for concentration]. | |
| Mar 18, 2013 at 8:55 | comment | added | cbeleites | Calibration samples are usually either prepared in a way that ensures a known concentration or the analyte concentration of the calibration samples is measured by a reference method. In any case you have samples where you have the known concentration and the instrument response. Which way you took experimentatlly usually doesn't matter for the further data analysis. | |
| Mar 18, 2013 at 8:54 | history | edited | Glen_b | CC BY-SA 3.0 |
added 346 characters in body
|
| Mar 18, 2013 at 8:49 | history | edited | Glen_b | CC BY-SA 3.0 |
added 346 characters in body
|
| Mar 18, 2013 at 8:44 | history | edited | Glen_b | CC BY-SA 3.0 |
added 346 characters in body
|
| Mar 18, 2013 at 8:32 | history | answered | Glen_b | CC BY-SA 3.0 |