Keith and The Girl is a free comedy talk show and podcast
Check out the recent shows
Click here to get Keith and The Girl free on iTunes.
Click here to get the podcast RSS feed. Click here to watch all the videos on our YouTube channel. |
04-28-2009, 10:51 PM | #71 (permalink) | ||
Senior Member
Join Date: Jan 2006
Posts: 866
|
Quote:
Sensitivity = Pr(T+|D+) = Probability of testing positive given you have the disease. Specificity = Pr(T-|D-) = Probability of testing negative given you do not have the disease. PPV = Pr(D+|T+) = Probability of disease given you tested positive NPV = Pr(D-|T-) = Probability of no disease given you tested negative or, expressed in another way: ...........................sensitivity x prevalence PPV = ------------------------------------------------------------------- ..........sensitivity x prevalence + (1 - specificity) x (1 - prevalence) ...........................specificity x (1 - prevalence) NPV = ------------------------------------------------------------------ .........(1 - sensitivity) x prevalence + specificity x (1 - prevalence) Note that unlike sensitivity and specificity, the PPV and NPV are dependent upon the prevalence of a 'disease'. A test can have 99% sensitivity and 99% specificity, but only detect 50% of the truly positive cases (possible in rare diseases). I'll admit that it's odd to think of pregnancy as a disease and that the prevalence for 'short-duration' conditions is not as easy to comprehend (or estimate) as in chronic diseases like lung cancer, but try to stay with me. Using Baye's theorem, we can calculate the PPV and NPV if given the sensitivity and specificity of a test, and if an estimate of the disease prevalence is available. Let's work through an example. Ex. Suppose the sensitivity and specificity of a test is 0.95 and the prevalence of pregnancy for US teenage girls is 0.12. Using the formulas above, PPV = 0.72 (72%) NPV = 0.99 (99%) However, you're actually interested in Pr(D+|T-,T-,T+), the probability of being pregnant given two negatives and one positive test. Since each test is independent of the other, this probability becomes: .................................Pr(T-|D+)*Pr(T-|D+)*Pr(T+|D+)*Pr(D+) Pr(D+|T-,T-,T+)= -------------------------------------------------------------------------- ..........................Pr(T-|D+)*Pr(T-|D+)*Pr(T+|D+)*Pr(D+) + Pr(T-|D-)*Pr(T-|D-)*Pr(T+|D-)*Pr(D-) or ...........................(1- sensitivity) x (1- sensitivity) x sensitivity x prevalence Pr(D+|T-,T-,T+)= ------------------------------------------------------------------ .........{(1- sensitivity) x (1- sensitivity) x sensitivity x prevalence} + {specificity x specificity x (1- specificity) x (1-prevalence)} and for our made up values: Pr(D+|T-,T-,T+)= 0.01 and similarly Pr(D-|T-,T-,T+)= 0.99 So, under these circumstances I'd say it would be highly unlikely to be pregnant. HOWEVER, remember these are made up sens., spec., and prevalence estimates. More importantly, this assumes each test was administered properly. A somewhat old meta analysis (1998) in which investigators reviewed 16 over the counter pregnancy tests, found sensitivities to be as low as 0.52 (varied drastically among volunteer groups and actual consumer groups). Using this number for the sensitivity in our example changes things drastically (PPV=0.59, NPV=0.94, Pr(D+|T-,T-,T+)= 0.27, and Pr(D-|T-,T-,T+)= 0.73). Here is the link to the article as well as the researchers' conclusions. Arch Fam Med -- Diagnostic Efficiency of Home Pregnancy Test Kits: A Meta-analysis, September 1998, Bastian et al. 7 (5): 465 Quote:
Last edited by golgi body; 04-29-2009 at 01:15 AM. |
||
(Offline) |
06-16-2009, 05:43 PM | #72 (permalink) |
Senior Member
Join Date: May 2008
Location: OKC, OK
Posts: 234
|
I'm having the same kind of issue.
My ex said that she might be preggo, because she's noticing sensitivity and nausea, etc, the whole nine. she's going to get a test today, but I'm wondering if it's mine. Here's why. She and I last had sex in April. She just cheated on me 2 sundays ago. If she is pregnant, is there a possibility that it is this other guys and not mine? |
(Offline) |
06-16-2009, 06:59 PM | #73 (permalink) | |
Senior Member
Join Date: Aug 2006
Location: I live in Southern California.
Posts: 3,055
|
Quote:
|
|
(Offline) |
Keith and The Girl is a free comedy talk show and podcast
Check out the recent shows
Click here to get Keith and The Girl free on iTunes.
Click here to get the podcast RSS feed. Click here to watch all the videos on our YouTube channel. |
06-17-2009, 07:21 AM | #75 (permalink) | |
Senior Member
Join Date: Nov 2008
Location: The Free State
Posts: 580
|
Quote:
If she is pregnant, an early ultrasound will be able to date the pregnancy to within 12 hours. Pretty easy to determine whether your sperm could possibly have been there during that window yes? |
|
(Offline) |
06-17-2009, 08:12 AM | #76 (permalink) | |
Senior Member
Join Date: Mar 2009
Location: Nude Hampster
Posts: 1,971
|
Quote:
(829): im not a father (843): about that... |
|
(Offline) |
Currently Active Users Viewing This Thread: 1 (0 members and 1 guests) | |
|
|