-
- admin
- Site Admin
-
- Posts: 647
- Joined: Wed Aug 11, 2004 12:19 am
By Anonymous on Wednesday, June 2, 2004 - 09:33 am:
We have a small analytic lab employing four HPLC
technicians. New hires are required to run a proficiency
test consisting of 4 calibrators ( to establish the
calibration curve using linear regression software), 3
controls (blank, low and high) and 10 samples. The
samples have been previously tested by a staff
technician and so we have an existing target value. To
pass the test, we require that all calibrators and controls
be within +/- 20% of their face (spiked) values. We also
require that all sample values be within +/- 20% of the
previously reported test values. We have never had a
problem with this standard until now. Technicians have
generally passed after 1-3 attempts. We declined to clear
a technician after he failed 7 test runs (failures were due
to 1 or more sample values being out of range). He is
challenging our decision, claiming that our criteria for
passing is "flawed." He says that the standard should be
that sample values must be within 20% of the mean of
the 2 sample values, and that if we had used this criteria,
his values would have passed. Given that our standard
has always worked, is there anything wrong with it?
-------------------------------------------------------------------------------------------------------
By Anonymous on Wednesday, June 2, 2004 - 10:35 am:
I would ask the question, If all but 1 person can pass the test as written, what is that 1 person doing wrong (or differently). It is possible that the standard has changed or a previously unknown variable has crept into the method.
-------------------------------------------------------------------------------------------------------
By Anonymous on Saturday, June 5, 2004 - 09:42 am:
Have an experienced staffer re-run the test just to be sure the target hasn't changed. If it *has*, then see if the new guy got the correct results after all. If it hasn't, then he's out (you can't change the pass/fail criteria retroactively -- unless you're in politics).
-------------------------------------------------------------------------------------------------------
By Tim on Tuesday, June 8, 2004 - 03:29 am:
I assume that you have been closely supervising this person when they failed after the first couple of attempts? If not, how do you know that it is not just something fundamental in what they are doing that is wrong?
I would also concur with the previous two posters that you should get an experienced person to double check the target hasn't changed and if it hasn't, then this person is doing something fundamentally wrong.
We have a small analytic lab employing four HPLC
technicians. New hires are required to run a proficiency
test consisting of 4 calibrators ( to establish the
calibration curve using linear regression software), 3
controls (blank, low and high) and 10 samples. The
samples have been previously tested by a staff
technician and so we have an existing target value. To
pass the test, we require that all calibrators and controls
be within +/- 20% of their face (spiked) values. We also
require that all sample values be within +/- 20% of the
previously reported test values. We have never had a
problem with this standard until now. Technicians have
generally passed after 1-3 attempts. We declined to clear
a technician after he failed 7 test runs (failures were due
to 1 or more sample values being out of range). He is
challenging our decision, claiming that our criteria for
passing is "flawed." He says that the standard should be
that sample values must be within 20% of the mean of
the 2 sample values, and that if we had used this criteria,
his values would have passed. Given that our standard
has always worked, is there anything wrong with it?
-------------------------------------------------------------------------------------------------------
By Anonymous on Wednesday, June 2, 2004 - 10:35 am:
I would ask the question, If all but 1 person can pass the test as written, what is that 1 person doing wrong (or differently). It is possible that the standard has changed or a previously unknown variable has crept into the method.
-------------------------------------------------------------------------------------------------------
By Anonymous on Saturday, June 5, 2004 - 09:42 am:
Have an experienced staffer re-run the test just to be sure the target hasn't changed. If it *has*, then see if the new guy got the correct results after all. If it hasn't, then he's out (you can't change the pass/fail criteria retroactively -- unless you're in politics).
-------------------------------------------------------------------------------------------------------
By Tim on Tuesday, June 8, 2004 - 03:29 am:
I assume that you have been closely supervising this person when they failed after the first couple of attempts? If not, how do you know that it is not just something fundamental in what they are doing that is wrong?
I would also concur with the previous two posters that you should get an experienced person to double check the target hasn't changed and if it hasn't, then this person is doing something fundamentally wrong.
