Log in
Register|
Compliance ★ India, 2014-02-05 09:30 (4127 d 18:41 ago) Posting: # 12338 Views: 5,828 |
|
|
Dear All, Is it advisable to run parallel study on two different make of instruments. I feel it is okay if we doing 2 way cross over but bit confuse about parallel design study analysis. Please share your thought on this. Regards, Compliance |
|
ElMaestro ★★★ Denmark, 2014-02-05 09:37 (4127 d 18:34 ago) @ Compliance Posting: # 12339 Views: 4,881 |
|
|
Hi compliance, ❝ Is it advisable to run parallel study on two different make of instruments. I feel it is okay if we doing 2 way cross over but bit confuse about parallel design study analysis. Shouldn't be an issue from an approval perspective as long as the use of both instruments is protocolled and validated. But it might increase variability and thus the width of the CI. If we aren't talking instruments for bioanalysis but e.g. a medical device used to quantify a PD-endpoint then this situation would be reasonably common for a multi-center trial. — Pass or fail! ElMaestro |
|
Compliance ★ India, 2014-02-05 14:10 (4127 d 14:01 ago) @ ElMaestro Posting: # 12342 Views: 4,918 |
|
|
Dear ElMaestro, ❝ Shouldn't be an issue from an approval perspective as long as the use of both instruments is protocolled and validated. Yes we do validate method on both of the instrument. ❝ But it might increase variability and thus the width of the CI. Here my concern is. i am thinking that till date the evaluation of the same subject data is not evaluated on two different make of instrument. I mean till date none of the CRO (with whom we are working) has done the evaluation of ISR on either instrument. I am worried does the out come of the subject may change if analysed on two different make. This is the parallel study and inter subject ratio play a roll in this. if the either instrument in not producing reasonably same results, it will be great trouble. Regards, Compliance |
|
Ohlbe ★★★ France, 2014-02-05 15:28 (4127 d 12:43 ago) @ Compliance Posting: # 12344 Views: 4,971 |
|
|
Dear Compliance, If you have some concerns you may perform a cross-validation between the two makes, first by analysing QC samples from the same batch on both machines, then by analysing trial samples on both machines and comparing the results. — Regards Ohlbe |
|
Compliance ★ India, 2014-02-07 05:38 (4125 d 22:34 ago) @ Ohlbe Posting: # 12360 Views: 4,918 |
|
|
Dear Ohlbe, ❝ first by analysing QC samples from the same batch on both machines, then by analysing trial samples on both machines and comparing the results. Exactly same i thought and i am in process to do so. Your feed back helps me a lot. Regards, Compliance |
|
Helmut ★★★   Vienna, Austria, 2014-02-07 14:46 (4125 d 13:26 ago) @ Compliance Posting: # 12363 Views: 5,013 |
|
|
Hi Compliance, sorry for cutting in. If you compare back-calculated concentrations from the machines you have to take into account that both are (uncertain) estimates. Therefore, conventional regression (assuming error-free regressors) is not suitable. Consider Deming-regression (aka orthogonal regression) and/or a Bland-Altman plot (aka Tukey mean-difference plot). See also this thread. Good luck! — Dif-tor heh smusma 🖖🏼 Довге життя Україна! ![[image]](https://static.bebac.at/pics/Blue_and_yellow_ribbon_UA.png) Helmut Schütz ![[image]](https://static.bebac.at/img/CC by.png) The quality of responses received is directly proportional to the quality of the question asked. 🚮 Science Quotes |
Ing. Helmut Schütz