Re: MSTS Risk
From: Terry J DuBose (tjdubose@juno.com)
Tue Feb 13 20:29:02 2001
Wayne, the Web page I pointed to earlier discussed the fact that the Odds
Ratio is a bit counter intuitive... or a bit odd, as you put it. Check
it out:
http://www.jr2.ox.ac.uk/Bandolier/band25/b25-6.html
--
Terry J DuBose
Little Rock, AR 72205
On Tue, 13 Feb 2001 20:18:36 -0600 Wayne.Persutte@UCHSC.edu writes:
> Terry, Gene, Allan, et al.
>
> Does this mean that a risk of 1 in 500 is the same as an odds ratio
> of
> 1:499??
>
> If I am wrong, please forgive my ignorance, but I believe that
> reporting
> odds ratios in interpreting triple screen information, instead of
> risk (as
> is the more common vernacular in prenatal diagnosis) is a little odd
> (no pun
> intended). Although I have seen it before, it caught me a bit off
> gaurd when
> I saw a patient today with a 1 in 1 'risk' (as was presented to me);
> which
> made no sense. I now know that the genetic counselor intended to
> imply the
> odds ratio. Does anyone know why we use OR instead of risk? Does it
> simply
> (as Gene stated) have its roots in Bayesian analysis which is
> commonly used
> by geneticists and genetic counselors?
>
> Gene you lost me here, "Odds are different from actual risk, with
> one
> exception: when they are 1:1. If the odds are "even", or 1 to 1,
> you have a
> total of 2 different outcomes, and the risk is 1 out of 2, or 50%."
> First
> you state that the odds of 1:1 are the same as a risk of 1:1, then
> you seem
> to contradict yourself? While I think I am beginning to understand
> what you
> are saying, I am afraid you threw me a cuvreball here.
>
> Wayne
>
> -----Original Message-----
> From: Allen Worrall [mailto:jworrall@alaska.net]
> Sent: Tuesday, February 13, 2001 1:25 PM
> To: Multiple recipients of list ULTRASOUND
> Subject: Re: MSTS Risk
>
> Risk is, for example, 1 affected pregnancy in 270 pregnancies. In
> other
> words the lower or second number includes the affected pregnancy. An
> odds
> ratio is, for example, 1 affected pregnancy in x normal
> pregnancies. The
> second number does not include the affected pregnancy. Thus, to
> convert risk
> to odds, (when the risk is 1:270), reduce the second number by 1, so
> the
> odds ratio is now 1:269.
>
> So, if the risk is 1:1, I guess the odds would be 1:0, meaning there
> are no
> normal pregnancies.
>
> I am sure my explanation is simplistic.
>
> Allen Worrall
