PakMediNet Discussion Forum : Biostatistics : importance os statistics
I wonder about the importance of statistics in paramedial subjects. I don't think it is really imporatnt and i will mention an example. suppose you are comparing 2 solutions to detect which one is better for the antibody identification. you found that soln. A detected more antibodies than soln. B?you used statistical test and say it is sign test, the p value was insignificant? does that mean i should use soln. B despite A is better?
thank you
Posted by: palestinePosts: 7 :: 22-09-2006 :: | Reply to this Message
Dear Palestine,
a.s.a.
The statistical test you used for a test is inaccurate. To define a "better" diagnostic test, we need estimates of its accuracy. Estimates of diagnostic accuracy include, Sensitivity and Specificity, Receiver Operating Characteristic (ROC) curves, Diagnostic Odds Ratio, Likelihood Ratios, and Positive and Negative Predictive Values.
Many tests are adopted in clinical practice when their diagnostic accuracy are evaluated in dubious clinical trials. This results in wastage of limited health care resources. Therefore critical appraisal of diagnostic tests is very important more so for developing countries with scarce resources for healthcare.
[Edited by asiddiqui on 23-09-2006 at 09:46 PM GMT]
Posted by: asiddiquiPosts: 26 :: 23-09-2006 :: | Reply to this Message
thank you very much for your reply...just i would like to tell you why i thought and i am till convinsed that statistics is not always necessary for paramedical students..a paper evaluated 2 LISS solutins to detect which solution can detect more antibodies than the other..he found that solution A is better than B for the detection of Rhesus antibodies,hoever from statistical point of viewwhen he used sign test he found that there is nodifference between the 2 solutions in the ability of their detection of Rhesus antibodiesdoes that make any sense for you?
thank you
Posted by: palestinePosts: 7 :: 02-10-2006 :: | Reply to this Message