When we finished the experiment according to the SOP and study plan, the results is riddiculous and the QC sample is beyond the criteria due to some strange reason from the instrument, saying that the instrument is not stable in long time running, but it is validated and works well in short time tests.
We reperfomed the analysis again in another apparatus which showed good function in both short and long time running and then the result is good.

My questions are
Do I need to report or mention the result of the first test in my final report.
Is this a deviation from study protocol or any SOP.
Can you suggest what kind of documents of procedures and records should be done?

Thank you very much for any sharing of your oppinon.

Welcome to The Cove! :bigwave: This is a good first post.

To what standard are you operating? Requirements for recording all testing can differ.

If you work to FDA requirements, my understanding is you would have to maintain the original testing data, and determine the cause of error. Automotive customers might want an 8D to determine error and correct the cause(s).

There is no way for us to tell here what went wrong; you would need to do that, but I have some ideas:

1) You mentioned a different apparatus was used for the second test. Did you check the first one for calibration? Was a quick calibration check performed before the first test? (It's not always required, I am just asking)

2) Did different people perform the tests?

3) Were instructions used?

4) Was the apparatus subject to different environmental conditions like heat, cold or vibration?

5) Is there a warm up period for the apparatus?

6) Is the original apparatus clean and undamaged by a bent or burred edge? Even a bent needle on a dial indicator can produce improper readings. I wonder if something became loose on the original apparatus mounting rig.

7) Was math involved? If so, see #3; was a calculator used? Are notes available?

I hope this helps!

If you work to FDA requirements, my understanding is you would have to maintain the original testing data, and determine the cause of error. Automotive customers might want an 8D to determine error and correct the cause(s).

---YES, the GLP principles required that all testing data maintained. I am confused on whether the first test be written into the final report.

There is no way for here us to tell what went wrong; you would need to do that, but I have some ideas:
1) You mentioned a different apparatus was used for the second test. Did you check the first one for calibration? Was a quick calibration check performed before the first test? (It's not always required, I am just asking)
2) Did different people perform the tests?

---Yes, all the checks have been done before the first test, and other people in the lab found also the same trouble in the first instrument. It passed all the validation test and showed good performance in short-time test but just functions unstably during long-time(12hours more) run. We will contact the instrument supplier for solving that problem.

3) Were instructions used?
4) Was the apparatus subject to different environmental conditions like heat, cold or vibration?

---The analyst followed all the instructions and the apparatus was subject to normal environmental conditions.

5) Is there a warm up period for the apparatus?
---It is not a new one and has been working in our lab in 4-5 years.

6) Is the original apparatus clean and undamaged by a bent or burred edge? Even a bent needle on a dial indicator can produce improper readings. I wonder if something became loose on the original apparatus mounting rig.

---Yes, there must be something wrong with the instrument. Just that we are now unable to find out it and will leave it to our supplier.

7) Was math involved? If so, see #3; was a calculator used? Are notes available?

---Few possible it is related to math since the result is directly read out by the instrument.

I hope this helps!

Jennifer, Thanks a lot for your reply.

min1999,

I do not think the original data needs to be included in the report, but do keep a record of the experiment in your files. I also think it's a good plan to have the instrument's supplier check out the unit, help you troubleshoot the problem.

It is getting late, so if no one adds anything I can bring this back up tomorrow for another view and comment.

...the results is riddiculous and the QC sample is beyond the criteria due to some strange reason from the instrument, saying that the instrument is not stable in long time running, but it is validated and works well in short time tests.

Have you revised your validation process and criteria so as to test for and disqualify this sort of longer-time instability?

The last thing a lab needs is a V&V process that assigns a "validated" credential to test equipment based on incomplete or incorrect evaluation of the equipment's capability to perform its required task over the required time in the required environment.

One way to formally disqualify the original data, retrospectively, is to formally determine, retrospectively, that the test equipment used was not correctly validated.

Does anyone else care to weigh in on this?

Have you revised your validation process and criteria so as to test for and disqualify this sort of longer-time instability?



Thanks a lot, MIREGMGR

I agree with you on the opinion of the reevaluation of the validation process in similar case. But Just that sometimes the instrument is complicated and trouble arises only in some special time or experiment. It may be not practical to validate each tiny part of the instrument in the daily qualification. However We will consider that and try to cover it in our yearly validation.
Now I still cant disqualify the original data by the wrong validation process. It is just the trouble during this analysis.

Btw,Can you introduce more details on the "V&V process"

"V&V" is just shorthand for verification and validation.

i just chanced upon this post; i am aware that its too late and hour of need has long passed away.;
but nevertheless let me share my :2cents: of experience on the problem-scenarios here [especially since there are few things which are not covered on the subject so far]

summary of posts so far
1. 8d process requirement
2. invalidating the data
3. revisiting the V&V process of the equipment

i am assuming that you are referring to problems faced by you in analytical instruements...

1. essentially i would like to add here that 'analytical method' also needs to be investigated. and if that specific make-of-instrument is found not-suitable for that specific method then it should be duly specified in the method

2. although you didn't mention that during investigation process you must have re-considered all the results generated using that instrument [even when it apparently workied fine] i assume that you had duly evaluated them as well, when the instrument was giving expected results.!

3. there should be an incident/deviation initiated for the 'observation/discovery' on the malfunctioning of the equipment; and then followed by CAPA process. this is similar to the 8D process with a lot of emphasis on the documentation/records of the detailed evaluation process.

4. as per the reporting, these failure/faulty results may not be considered for evaluation of the study[as the post apparently suggests that the reason being nailed down as instrument!], but they are certainly included as the deviations/incidents in the final report.

reqeust min1999 to share the follow up of the a/p implemented for the said problem and their experience with regulators/auditors

thanks
valiveti