Re: Cholesterol Levels Are Falling, But Red Flags Are Rising
- From: "Robert" <RobertsSong@xxxxxxxxxxx>
- Date: Thu, 13 Oct 2005 13:32:30 -0700
<James216440@xxxxxxxxx> wrote in message
> Robert,
>
> I admire your confidence in the numbers you personally produce. Spoken
> like a true lab guy. My people often made the same kinds of strong
> assertions on how good their work was. And many others outside my
> particular organization were equally adamant on their work. But as I
> said I have been personally too involved in way too many round robin
> studies on a whole variety of analyses to believe anyone is nearly as
> good in terms of absolute accuracy as you claim.
In lab cvs are close
> to meaningless. And even showing you can get the correct results on an
> NBS standard do not really prove much at all. If analytical chem were
> that good many of my publications would never have happened as there
> would have been no story to tell.
The story involving cholesterol was a high CV in the begining which is why
if you look at the GAO report showed the progression to take. There were no
national standards and therefore no uniformity in results. The results were
evaluated on an "in site" reference range depending on the type of
instrument you had and the population of patients you saw. That is how you
establish the reference range.
It is not a matter of opinion. Federal law requires testing for accuracy of
test results in clinical laboratories 4 times per year.
This is not a claim. They are tested.
Most of the studies are old and used to promote the standard which is in
place today. Place the national standard and test the labs 4 times a year.
You don't need further studies as the PT by CAP and others are there.
I don't know what kind of job you have but does it involve being tested 4
times a year by outside organizations and if you don't pass then you can't
perform that task?
The other point that you are missing involves the interpretation of testing.
If we were to replace an instrument with another one we need to do
correlation studies involving both instruments. Test the old and new
instruments and check to see if there is a statistical difference in both
sets by T test statistics. This only tells you if there is a mathematical
statistical difference and not a clinically significant difference. As
clinical testing goes it is the clinically significant range that is
important and not just a statistical one. The 200 mg/dl standard was chosen
because of it's clinical significant range.
You can not compare other testing in your round robin as there is no
national standards uniformity. If they are the same instrument or same
organizations you can calibrate your instruments to make them agree within 2
STD's.
We also check instruments in our own hospital as we often have two or three
of the same instrument and have them all match. We can not have them give
differing results depending on what floor they are on. Again the 2SD's and
clinically significant differences is followed.
And some of the best of those
> publications were on analyses that have been done for over 100 years
> and have excellent NIST standards to run to show performance.
>
> My K numbers said I should have been in fibrilation at the moment the
> samples were taken. The results were called in. My doctor called me
> at home and I was in the middle of a quite strenuous project I simply
> could not drop without adverse consequences. I had a lot of wet glue
> drying on wood and faced hours of clean up if I simply went immediately
> to the lab which is 30 miles away for a retest. We agreed there was
> zero chance the K number was correct and that getting a retest the next
> day was fine. The lab drew both samples. I have had a needle stuck in
> my arm often enough to have some idea if it is being done properly.
> Both were. The lab simply screwed up the test.
I don't have any insight on that particular case. I can tell you that our
policy requires counseling for one major error. On a second major error then
a written warning to documented and the third major error per year the
employee is terminated. We have a three strike rule.
Mistakes can and do happen but to say that they are the norm is not true and
as I have said the recollecting is needed. The lab could have just as easily
rerun the old sample to see if they got the same high result. This is needed
to find out if it is machine error or not. The laboratory would know exactly
at what step the error occured whether wrong label, hemolysis of sample,
machine error or data entry error.
We can run ten samples with barcodes that have name and test to perform info
and the label and put them in the machine and the machine reads the label
and runs the test on them and it is interphased with a host computer. All we
do is look at the computer printout. If you see all deadly high K's coming
out then you have a problem. Most of the time it's only one patient and then
you can repeat that one patient to see if you get a high result.
That's pretty much standard. We also have the option of not reporting out
the result at all and simply state specimen unsatisfactory. One screw does
not make the rule. Screw-ups in general is what makes up policies and
protocols on how to check for that. Repeats are one of those.
How I have no idea.
> It happens face it. By the way the lab was at the branch of the
> teaching hospital that consistantly got the low cholesterol numbers on
> me. It was not some back alley lab run by people who have no training
> or skills other then filling out the billing info. Not that there are
> not many of the latter. They come and go like flies in the summer at a
> horse farm.
All clincial laboratories are monitored and teaching vs small hospital has
nothing to do with it. Errors can happen in all settings.
>
> The lab business is getting better with time. No question about that.
> The days of labs reporting data when they do not even have access to
> the instruments required to perform the tests they claim they performed
> are slowly ending. Commercial labs have historically been a very dirty
> business filled with outright fraud. But these days there are enough
> certification processes in place to force a little honesty into the
> hearts of the biggest thieves. My lab was audited routinely by FDA,
> EPA, the Nuc Industry, A2LA and many individual clients over a wide
> range of industrys. I liked having a steady stream of auditors come
> through the place as it reduced my day to day management job immensly
> and freed me up to do my real job which was to grow the business. My
> people all knew they were going to face many critical auditors every
> single year and if the documentation, labeling, standardization
> records, customer complaint issues and on and on were not in order
> every day they would be in big trouble. We even had fill dates and
> expiration labels on our acetone wash bottles used in glassware
> cleanup. And with all this we also found an example every year or so
> where we had simply blown an analysis. Then I got to call the customer
> personally and tell him what we had done and why he was going to get an
> ammended report. Sometimes it was an instument problem. More often it
> was human error. I am sure we did not find every significant error we
> generated. I did catch two cases of individual fraud. Fired both of
> them. The way I found both was the data was too good to be believable.
> In one case I submitted blind knowns to catch the idiot. In the other
> simply had another person try to repeat the analysis. These are the
> simple thieves to catch. Anyone with a tiny amount of sense can cover
> his tracks better then this.
>
Huuuuum!!!! I don't know about the situation you are in but let me say that
some labs hire all kinds of untrained people and have the licensed personal
oversee their functions as I have done when I first started out. These are
fresh high school people without proper training who really don't know what
they are doing put straight into a clinical laboratory. It is cheaper to
hire these people compared to the more expensive professionals so you get
what you pay for.If these are professionals then you should have the state
regulatory or national registry pull their licenses.
In our state it required we declare, as in any profession, whether we have
been convicted of a felony or misdemeanor every year.
Fraud is obviously what they are looking for.
The cost of healthcare is up and some of the proposed cuts involve cutting
some requirements.
.
- References:
- Re: Cholesterol Levels Are Falling, But Red Flags Are Rising
- From: James216440
- Re: Cholesterol Levels Are Falling, But Red Flags Are Rising
- From: Robert
- Re: Cholesterol Levels Are Falling, But Red Flags Are Rising
- From: James216440
- Re: Cholesterol Levels Are Falling, But Red Flags Are Rising
- Prev by Date: If you or anyone you know has Cardiomyopathy
- Next by Date: problems sleeping on vacation
- Previous by thread: Re: Cholesterol Levels Are Falling, But Red Flags Are Rising
- Next by thread: OT Nashville Steak House
- Index(es):
Relevant Pages
|
