Sponsored Links
Results 1 to 7 of 7
  1. #1
    qaiser is offline Newbie 510 points
    Join Date
    Nov 2006
    Posts
    16
    Downloads
    12
    Uploads
    0
    Thanks
    0
    Thanked 0 Times in 0 Posts

    Sensitivity vs. Specificity of a test

    Advertisements



    From clinical point of view, it is very important to understand the concept of sensitivity and specificity of a test.
    Sensitivity: Suppose 10 subjects were screened for a test. In reality all 10 did have the disease but test came positive only for 8 people. This test failed to catch 2 people who had the disease. Test showed false negative tests for those two persons. This test is only 80 % sensitive.
    Sensitivity is calculated by the following formula:
    True positives/True positives+False negative
    An ideal test will have a high sensitivity meaning it will catch all people with a disease. If a test has a low sensitivity, it wouldn't be reliable because as a clinician based in false negative, you will tell your patient that he/she does not have the disease while in reality he/she is diseased. It is like letting loose a criminal.
    Specificity: Suppose 10 persons were tested for a disease and all 10 people came positive for that disease. In reality, only 8 people had the disease. The two people who came positive do not have the disease in reality. They have some other condition that gives a false positive.
    If a test has a low specificity, it is not relaiable because it will give false positive results.
    The formula for specificity is:
    True negative/true negative+False positive
    An example of a test that is not very specific is VDRL and RPR (syphilis diagnostic tests). These tests show positive results for SLE (systemic lupus erythematosus).
    Any comments are welcomed.
    Good luck in your studies.

  2. #2
    stephew is offline Moderator Guru 512 points
    Join Date
    Jan 2003
    Posts
    21,867
    Downloads
    0
    Uploads
    0
    Thanks
    0
    Thanked 0 Times in 0 Posts
    know the different beween this and positive and negative predictive value. they love that and its very usefyul to use.
    Steph
    If you get a warning, put on yer manpants and stop whining about it.

  3. #3
    qaiser is offline Newbie 510 points
    Join Date
    Nov 2006
    Posts
    16
    Downloads
    12
    Uploads
    0
    Thanks
    0
    Thanked 0 Times in 0 Posts
    Stephew,
    Thanks for your input. Yes perdictive value is an important topic indeed. My next post is about predictive value. By the way what was those lines about warning below your reply? I did not quite get that. Thanks again for your reply. Good luck in your studies.

  4. #4
    Sree Cheruku is offline Super Moderator 534 points
    Join Date
    Dec 2004
    Location
    Cleveland, OH
    Posts
    2,394
    Downloads
    10
    Uploads
    0
    Thanks
    0
    Thanked 0 Times in 0 Posts



    MANPANTS! duh.. don't you own a pair? They help you not whine when you get a warning.
    Last edited by Sree Cheruku; 12-27-2006 at 07:45 PM.
    [X]Board Certified in Anesthesiology
    [X]PGY-5 Critical Care Medicine [*] PGY-6 Cardiothoracic Anesthesiology


  5. #5
    redeye's Avatar
    redeye is offline Member 510 points
    Join Date
    Nov 2006
    Posts
    235
    Downloads
    0
    Uploads
    0
    Thanks
    0
    Thanked 0 Times in 0 Posts
    Quote Originally Posted by qaiser View Post
    Stephew,
    Thanks for your input. Yes perdictive value is an important topic indeed. My next post is about predictive value. By the way what was those lines about warning below your reply? I did not quite get that. Thanks again for your reply. Good luck in your studies.
    Yes, everyone always seems to emphasize sensitivity and specificity when discussing tests, but prevelence to me is the most important and most ignored.

    Take a test that is both 99% sensitivity and 99% selectivity (sounds great)

    If the prevelence for the disease you are testing for is 1/10,000 then you will have a ~1% probability that a positive result is an actual positive (99% of the patients will be false positive). So the predictive value is only 1%.
    Last edited by redeye; 12-27-2006 at 08:36 PM.

  6. #6
    qaiser is offline Newbie 510 points
    Join Date
    Nov 2006
    Posts
    16
    Downloads
    12
    Uploads
    0
    Thanks
    0
    Thanked 0 Times in 0 Posts
    Hi Redeye!
    Thanks for your input. Good luck in your studies.

  7. #7
    redeye's Avatar
    redeye is offline Member 510 points
    Join Date
    Nov 2006
    Posts
    235
    Downloads
    0
    Uploads
    0
    Thanks
    0
    Thanked 0 Times in 0 Posts
    Quote Originally Posted by qaiser View Post
    Hi Redeye!
    Thanks for your input. Good luck in your studies.
    Thank you. Good luck!!!

Similar Threads

  1. Specificity q1 N*
    By Roxanita in forum Behavioral Science Forum
    Replies: 4
    Last Post: 12-06-2006, 11:29 PM
  2. sensitivity and specificity
    By Asclepius1 in forum USMLE Step 1 Forum
    Replies: 1
    Last Post: 05-15-2005, 04:04 PM
  3. Specificity
    By Anonymous in forum USMLE Step 1 Forum
    Replies: 1
    Last Post: 05-15-2004, 06:05 PM
  4. MORE test advice...during the test, read the answer choices
    By tommyk in forum USMLE Step 1 Forum
    Replies: 0
    Last Post: 02-09-2004, 07:45 AM
  5. sensitivity or specificity, please explain
    By Anonymous in forum USMLE Step 1 Forum
    Replies: 1
    Last Post: 11-19-2003, 01:11 PM

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •  

Site Meter