Paul Collinson writes
<<ow here is something that will ignite controversy
Is there is =93best buy=94 for TFT=92s
My endocrinologist is quoting at me the guidelines published by the
British endocrine society/RCP and considers that two tests should be
done on all patients.
His view is TSH plus fT3
I read these as saying that diagnosis should be conformed by a free
hormone e.g. TSH then fT3 or fT4 if the TSH is not =93normal=94.
Questions
Does everybody measure 2?
Has anyone audited what is the best policy ( I mean real audit, not a
survey)?
Is there any outcome data?
What is published?=20
--=20
Paul Collinson>>
>>>>>>
I agree with your endocrinologist.
TFTS are done in a number of contexts and on samples
drawn from different populations - I work closely
with an endocrinologist who has a large hypopit workload.
Thus generalisations are dangerous and the figure of 3% from
Australia although helpful is not applicable everywhere.
We did a study using an agreed algorithm and simulated doing
TSH as a front line test. We found that we had to do a second line
test on around 60% of samples. By the time you had allowed for the
cost of finding samples twice and all the general thrashing around
involved we were deeply unimpressed with the reagent savings.
Two tests do not cost twice as much as one - the difference is much
smaller. (We ended up with a situation that was broadly cost neutral because we
negotiated a larger reagent discount!). We also were struggling to provide
an appropriate turnaround time ( and delaying the most abnormal results
the most!).
I accept that TSH alone is adequate in managing patients on thyroxine
replacement. However I manage 850 such patients and the combination of results
(total T4 and TSH) are often informative even in that group ( suggesting things
like
non compliance or non-thyroidal illness).
American endocrine organisations have come out strongly in favour of
two tests and I absolutely agree with them especially for diagnostic tests
(ie the first time around).
Sites that only do TSH simply do not know how many diagnoses they miss
(including TSH omas and other rarieties) and whilst the numbers may be small
these treatable conditions will certainly be missed by the lab.
Two tests also act as form of internal QA and we cetainly get clinically
implausable
results that turn out to be lab errors on repeat (short samples or whatever).
In some cases
initially normal TSHs turn out to be wrong. Blunder rates may run in most
labs at between 0.01% and 0.1% or higher (historically a figure of 1% has been
found).
My endocrinologist and I agree that we would not like to start life long
therapy on the basis of one test done in singlicate. For that reason it appears
more useful to do a TSH and Total T4 in singlicate rather than a TSH in
duplicate.
There is a difficult tradeoff here between an obvious and "easy" cost save
and a quality improvement that is real but hard to price.
Maybe there is no "right" answer, just a judgement that should be made in the
light
of the population served and in close local discussion with the relevant
clinicians.
Certainly very large studies in many different settings would have to be made to
establish
the answer and as innovation ensures that tests should get cheaper over time it
may
become increasingly unimportant.
More interesing questions (to my mind) include strategies to eliminate
worthless repeat tests
done within days or weeks of each other (often one by a GP and one by the
hospital) and
trying to eliminate situations where thyroid function has not been for far too
long
in the course of a illness resulting in delayed diagnosis.
James Falconer Smith.
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|