Well they might!!!!
It all depends on where you put the cut off and which test you use. This is
something that has not yet been discussed. With a continuous value such as
d-dimer level (or CKMBmass, or amylase etc. etc.) the ability of the test to
rule out or rule in depends on where you put the marker. For example a
CKMBmass on our testing less than 5 is pretty good at ruling out AMI (or
less than 3 rise & with reference to timings), but at this level it has poor
rule in value (you probably need to get above 8).
There is also quite a lot of variation between different tests. It is not
fair to compare SimpliRED against lab based quantitive assays (except in a
trial) for example.
Anyway we constantly interpret such tests as yes/no values. For example a
d-dimer of 270 is negative, whereas one of 279 is positive. Rubbish!! it
just means that one of those results makes a DVT/PE slightly more likely. It
is definately not positive/negative.
So in our example if we set the d-dimer cut off at 5 then it would work,
though it would be no use in practice at all.(very sensitive but not
specific enough). That is the thrust behind using the results of thes tests
in the context of a clinical risk stratification. By manipulating the
pretest probability in a structured manner you get much more benefit out of
your test.
Wouldn't it be nice if all continuous data tests like these were given with
two values, a rule in and a rule out level. And in the middle it is
equivocal and you need to do more work, or another test. Interestingly many
levels are set at the point at which they are most "accurate", i.e. a
compromise of sensitivity against specificity. Where is the sense in that!
It does not reflect our clinical practice.
Simon
Hi Ho Silver (hobby horse)
Simon Carley
SpR in Emergency Medicine
[log in to unmask]
Evidence based emergency medicine
http://www.bestbets.org
----- Original Message -----
From: "Dunn Matthew Dr. ACCIDENT & EMERGENCY - SwarkHosp-TR"
<[log in to unmask]>
To: <[log in to unmask]>
Sent: Thursday, October 18, 2001 10:07 AM
Subject: Re: re-d dimers
> To clarify my previous post: our haematologists tell us that negative
> d-dimers do rule out a DVT regardless of clinical findings. This seems to
> differ from other respondents.
>
> Matt Dunn
|