| Author | Message | ||
     Gavriel Levin (levin) Advanced Member Username: levin Post Number: 25 Registered: 1-2006 |
Hi Francesco How small pharma companies behave? Well, they do hire and use classical methods - because without it they just can't have regulatory approvals. If they don't have a lab, they send out for analysis. This though, opens the door to other ghosts. What do you do when there is a discrepancy between the results reported by the service lab and your customer's lab? Whose liability? Say the damage to you is 200K$, (could easily be more) you paid the lab say 2,000 for the service, can they be liable for the 200K$ dmagae to you? Being small does not excuse from doing all the needed tests. Back in 1997 I had dealings with a nutraceutical company that (and they are not as strinegnt with these products as with pharmaceuticals) was using an external lab for many of their required analytical services. They ended finally putting their own lab, because of such issues, as well as timing, etc. Life is not a rose garden. Gabi Levin Brimrose [email protected] | ||
| Gavriel Levin (levin) Advanced Member Username: levin Post Number: 24 Registered: 1-2006 |
Hi Francesco How small pharma companies behave? Well, they do hire and use classical methods - because without it they just can't have regulatory approvals. If they don't have a lab, they send out for analysis. This though, opens the door to other ghosts. What do you do when there is a discrepancy between the results reported by the service lab and your customer's lab? Whose liability? Say the damage to you is 200K$, (could easily be more) you paid the lab say 2,000 for the service, can they be liable for the 200K$ dmagae to you? Being small does not excuse from doing all the needed tests. Back in 1997 I had dealings with a nutraceutical company that (and they are not as strinegnt with these products as with pharmaceuticals) was using an external lab for many of their required analytical services. They ended finally putting their own lab, because of such issues, as well as timing, etc. Life is not a rose garden. Gabi Levin Brimrose [email protected] | ||
| Francesco Beccari (francesco_k) Junior Member Username: francesco_k Post Number: 7 Registered: 11-2006 |
mmm i c but then how small pharma behave? I guess it is a common problem to all those who cannot analyse their products by classical methods. | ||
| Gavriel Levin (levin) Advanced Member Username: levin Post Number: 23 Registered: 1-2006 |
Francesco, I would not be so quick to jump to conclusions about legal issues. Obtaining an FDA (or similar authorities such as the EU or other countries) is not such a quick and dirty effort. My recommendation is that before you decide on how to invest money into any instrument, study the requirements in either Pharmacopea (US, British, European, etc.) Firstly I don't think that one sample analyzed by an outside lab will suffice to say - everything that is like it is defintely and legaly the same. Secondly, there are also requirements for assay, and impurities, these can not be done by NIR. If you write to me I will explain in more details. I have worked with pharma companies on NIR since 1997. Things aren't that simple. One of the complaints I heard from my customers in pharma companies is that you can't trust suppliers' certificates to the fullest extents. [email protected] | ||
| Francesco Beccari (francesco_k) New member Username: francesco_k Post Number: 5 Registered: 11-2006 |
Hi Gavriel there are no stupid questions, only stupid answers!, so I'll try to answer properly. No, atm we only use a melting point for quick test, because we are not yet qualified for pharma products distribution. From now on, we'd like to move to this, and so we are committed to an ID. NIR is of course the solution. So maybe now you understand why I REALLY wish to avoid further job and rely on supplier's CoA. As told, this is not a problem from the point of view of the feasibility, but only from a regulatory point of view, although James' post is probably the solution. CIAO | ||
| Gavriel Levin (levin) Advanced Member Username: levin Post Number: 22 Registered: 1-2006 |
Hi Francesco, Simple and possily stupid - how do you identify your raw materials today? MID infra red with FTIR? Do you prepare samples with KBr amd press them? If you have a qualified lab to do this it is all you need to do for RM identification. Create a liabrary from MID IR identified materials, verify their assay is OK and if reuqired by law the levels of impurities, and then when your library contains acceptable number of samples from enough batches, then you validate the library with an adequate number of smaples from batches not included in your library. Now, how do you do assay today? Do you ship to an outside vendor? Gabi Levin Brimrose [email protected] | ||
| James Malone (jmalone) New member Username: jmalone Post Number: 4 Registered: 8-2006 |
Francesco: You don't need a lab to make a discriminant model. Many NIR and some Raman spectrometers can operate outside the confines of a lab. You could use one of these to collect spectra from incoming materials and build a discriminant model that says "this sample is very much like (or unlike!) previous samples with the same label from the same supplier". This will not be a definitive analysis such as the CoA, but it will give you some legal protection if you can show that you can identify differences between lot numbers of the same material. For more protection, I would a have a "typical" sample of each material from the supplier analyzed by a certified lab. Then, using the spectral data on incoming lots of that material, you could state the degree of "sameness" to the certified analytical results. Hope this helps. | ||
| Francesco Beccari (francesco_k) New member Username: francesco_k Post Number: 4 Registered: 11-2006 |
Thanks Howard - ya actually mine is a LEGAL issue, and my concern is only limited to the assurance that I can rely on the CoA from supplier. | ||
| Howard Mark (hlmark) Senior Member Username: hlmark Post Number: 57 Registered: 9-2001 |
Francesco - this is more of a legal issue than a technical one, I think. If you're only trying to identify the material then you don't need reference lab values for the composition. You would, however want to ensure that the samples you include in your modelling procedure cover the range of compositions you'll encounter in the future, so that your model doesn't erroneously reject valid samples. Given all this, my opinion is that you could rely on the COAs from the supplier, but that's only my opinion. If regulatory issues are involved then you may want to get confirmation that those COAs will be acceptable to the authorities, to demonstrate that your sample set includes the range of compositions you need. Howard \o/ /_\ | ||
| Francesco Beccari (francesco_k) New member Username: francesco_k Post Number: 2 Registered: 11-2006 |
Hi All can anybody please answer to this problem: I need to build an NIR discriminant model by using samples of raw materials coming from my supplier, and I don't have my own lab. Can I use the Certificate of Analysis coming from my supplier and rely on his data? Or, do I have to re-analyse them? The EurPh and the EMEA docs don't tell who should provide the CoA, but I would really enjoy an answer from you Thanks Francesco |