Statistical vs. clinical significant numbers.
Just because something is statistically significant does not mean that it is clinically significant. If the laboratory only used a narrow range of specimens, such as in a normal range study on both instruments, a small difference can make the comparison have a value of p<0.001 but the clinical significance is miniscule. (See McGlasson DL, et al., Rapid removal of platelets from plasma utilizing the Hepcheck heparin removal filter, Blood Coagulation and Fibrinolysis 8, 16-20, 1997).
A very well-explained definition of what that value means is the explanation about the probability indicating that the p<0.001, due to chance alone, is less than one in a thousand. Sometimes, laboratorians get caught up in numbers that do not have a clinically significant relevance in relation to their statistical outcome.
--David L. McGlasson, MS, CLS/NCA
59th Clinical Research Squadron
Lackland AFB, TX
MLO welcomes letters to the editor. We ask that you include a phone number for verification. While we prefer to publish the writer's name, we will publish a letter with "name withheld by request," but our editorial staff must have the writer's name confirmed for our files. MLO reserves the right to edit any letter for style and length.
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||Readers Respond|
|Author:||McGlasson, David L.|
|Publication:||Medical Laboratory Observer|
|Article Type:||Letter to the Editor|
|Date:||Dec 1, 2003|
|Previous Article:||Summer in the dead of winter.|
|Next Article:||Multiple views misleading.|
|Why readers respond. (Feedback).|
|Another candidate for the "top 10".|
|Testing toxic compounds in human subjects: ethical standards and good science.|
|Uniform Reporting Manual, 4th Edition.|