Log in
Register|
sam ★ India, 2013-08-01 08:15 (4300 d 14:13 ago) Posting: # 11138 Views: 27,373 |
|
|
Dear All, Recently I carried out one study in the ER formulation for USFDA. The results are very good for the Cmax and AUCinf but the study fails marginally in lower CI in AUCt that is 79.28. The failure in this study is due to the presence of two outlier in the whole study. If we remove these outliers the study qualify the BE criteria for all the three parameters. But as we all know that USFDA don’t encourage the outliers test and its exclusion from the study. Now we have only option to save our study otherwise we have to dump the whole study. The only option is if we remove the AUCt and AUCinf from the from two subjects whose %AUCextrap value is more than 20% and keeping Cmax and Tmax untouched. Our study qualifies the BE criteria of 80-125%. So please suggest if we can do in same line as I am thinking or there are any other ways to save our study. Hope for your positive reply. Sam Edit: Category changed. [Helmut] |
|
ElMaestro ★★★ Denmark, 2013-08-01 12:29 (4300 d 09:58 ago) @ sam Posting: # 11139 Views: 25,980 |
|
|
Hi Sam, ❝ (...) the study fails marginally in lower CI in AUCt that is 79.28. The failure in this study is due to the presence of two outlier in the whole study. If we remove these outliers the study qualify the BE criteria for all the three parameters. But as we all know that USFDA don’t encourage the outliers test and its exclusion from the study (...) So please suggest if we can do in same line as I am thinking or there are any other ways to save our study. If you can for-cause audit the study and ID justifiable reasons for exclusion of the subjects in question then that would most likely be ok (pukey-pukey in period 1 but not in period 2 and stuff like that). But it takes real effort to do it, and the audit should not just focus on the values/subjects that you believe caused study failure but also on the subjects that were in your views within the normal range. A little off topic, but given the recent discuss in another thread, this example is one where the concept of a repeat trial may become relevant. If your lower CI limit was 79.28% then chances are that the true T/R (geometric mean, observed) was likely to be reasonably close to 100% and that more subjects in the trial therefore would have saved it. Got some numbers for us? — Pass or fail! ElMaestro |
|
sam ★ India, 2013-08-01 13:21 (4300 d 09:07 ago) @ ElMaestro Posting: # 11140 Views: 25,974 |
|
|
Hi John ❝ A little off topic, but given the recent discuss in another thread, this example is one where the concept of a repeat trial may become relevant. If your lower CI limit was 79.28% then chances are that the true T/R (geometric mean, observed) was likely to be reasonably close to 100% and that more subjects in the trial therefore would have saved it. Got some numbers for us? First of all thanks for the prompt reply, The ratio for AUCt and AUCinf are 83.82 and 84.91 and the corresponding CIs are (79.28% - 88.61%) and (80.68% - 89.35%) respectively. The subjects evaluated for this trial is 124. Best Regards sam |
|
ElMaestro ★★★ Denmark, 2013-08-01 13:39 (4300 d 08:48 ago) @ sam Posting: # 11141 Views: 25,878 |
|
|
Hi Sam, ❝ The ratio for AUCt and AUCinf are 83.82 and 84.91 and the corresponding CIs are (79.28% - 88.61%) and (80.68% - 89.35%) respectively. The subjects evaluated for this trial is 124. Oh wow. It is slightly surprising, I thought the PEs would be closer to unity, but be that as it may. You might be able to demonstrate BE by increasing the number of subjects slightly, all other factors equal. — Pass or fail! ElMaestro |
|
sam ★ India, 2013-08-01 14:01 (4300 d 08:26 ago) @ ElMaestro Posting: # 11143 Views: 25,962 |
|
|
Hi ElMaestro, ❝ It is slightly surprising, I thought the PEs would be closer to unity, but be that as it may. You might be able to demonstrate BE by increasing the number of subjects slightly, all other factors equal. Thanks for your suggestion. I am also sure that if I repeat the study the study will qualify the BE criteria. But I have seen few cases in which if the %AUCextrap is more than 20% then the AUCt, AUCinf and all other secondary parameters were not calculated. Also I have few more things in some literatures. Following cases Kel, T1/2, AUC0-inf, AUCt, Kel start and Kel stop will not be reported
Edit: Please don’t copy/paste from somewhere else without checking the Preview. Your post contained non-printable characters. [Helmut] |
|
Helmut ★★★   Vienna, Austria, 2013-08-01 14:52 (4300 d 07:36 ago) @ sam Posting: # 11145 Views: 26,252 |
|
|
Hi Sam, ❝ I am also sure that if I repeat the study the study will qualify the BE criteria. I am not sure about that (considering your bad PE, the low CVs, and the high sample size). ❝ But I have seen few cases in which if the %AUCextrap is more than 20% then the AUCt, AUCinf and all other secondary parameters were not calculated. ❝ Following cases Kel, T1/2, AUC0-inf, AUCt, Kel start and Kel stop will not be reported ❝ • Those subject exhibiting regression (Rsq) less than 0.8… That’s a stupid criterion. Any cut-off based on the R²adj doesn’t make sense. ❝ …i.e. non terminal log linear phase Kel, T1/2 and AUC0-inf will not be reported for that subject./period. Make sense if λz cannot be reliably estimated (e.g., less than three data points). ❝ • For subjects /period exhibiting AUC_%Extrap_obs is greater than 20% .ie. AUC0-t/AUC0-inf ratios below 80%, Kel, T1/2 and AUC0-inf will not be reported. Doubful for the FDA and definitely not acceptable for EMA (see my previous post). ❝ Kindly let me know if any of the cases applies here. IMHO, no. — 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 |
|
d_labes ★★★ Berlin, Germany, 2013-08-01 13:59 (4300 d 08:29 ago) @ sam Posting: # 11142 Views: 26,068 |
|
|
Dear sam! ❝ Hi John So far as I know has our captain the name Großer Meister  .❝ The ratio for AUCt and AUCinf are 83.82 and 84.91 and the corresponding CIs are (79.28% - 88.61%) and (80.68% - 89.35%) respectively. The subjects evaluated for this trial is 124. That smells like forced bioequivalence. Not sure if FDA assessors like that (better quite sure they don't). — Regards, Detlew |
|
Helmut ★★★ Vienna, Austria, 2013-08-01 15:31 (4300 d 06:57 ago) @ d_labes Posting: # 11148 Views: 25,819 |
|
|
Hi Detlew! ❝ That smells like forced bioequivalence. Not even forced bioequivalence – rather gambling with the safety of subjects (see the end of this post). — Dif-tor heh smusma 🖖🏼 Довге життя Україна! Helmut Schütz The quality of responses received is directly proportional to the quality of the question asked. 🚮 Science Quotes |
|
Helmut ★★★ Vienna, Austria, 2013-08-01 14:32 (4300 d 07:56 ago) @ sam Posting: # 11144 Views: 26,381 |
|
|
Hi Sam, ❝ […] ER formulation for USFDA. The results are very good for the Cmax and AUCinf but the study fails marginally in lower CI in AUCt that is 79.28. The failure in this study is due to the presence of two outlier in the whole study. If we remove these outliers the study qualify the BE criteria for all the three parameters. But as we all know that USFDA don’t encourage the outliers test and its exclusion from the study. ❝ […] The only option is if we remove the AUCt and AUCinf from the from two subjects whose %AUCextrap value is more than 20% and keeping Cmax and Tmax untouched. Our study qualifies the BE criteria of 80-125%. Some points:
— Dif-tor heh smusma 🖖🏼 Довге життя Україна! Helmut Schütz The quality of responses received is directly proportional to the quality of the question asked. 🚮 Science Quotes |
![[image]](img/uploaded/IAsmall.png)
guidance, p22
|
sam ★ India, 2013-08-01 14:55 (4300 d 07:32 ago) @ Helmut Posting: # 11146 Views: 25,935 |
|
|
Hi Helmut, Thanks for your valuable suggestions: ❝ • You are right about FDA’s thinking about ‘outliers’. BTW, did these two subjects really qualify by a statistical test or was your criterion >20% res. AUC? In my study there were five subjects whose data for the test products are not in concurrence to the other subjects data. After thorough review we applied test of outlier and among the five subjects two qualifies as outliers. Among the two outliers one subject has AUCextrap value more than 20% and other subject’s whose extrapolation value is more than 20% is among the five whose data are aberrant compared to the other subjects data. ❝ • You can try to start an argument based on the fact that the formulation was ER. In BE we are interested in comparing the absorption (= formulation + drug specific) not distribution/elimination (= drug specific). With ER the slowest phase might be absorption – not elimination. This is true once you crossed flip-flop PK (ka = kel). Even for ka > kel the apparent (!) elimination will be contaminated by absorption. Compare half lives of your study with literature data of immediate release formulations. Likely your half lives are longer. Therefore, AUC∞ is a more relevant metric for extent of absorption than AUCt. Should I remove the AUCt parameter only for that subjects which have AUCextrap value more than 20% and write a comment on the report as per our suggestions. ❝ • Your sample size was terribly high given the CVs of 25–27%. Since the ratio in the pilot study was around 84% and considering the ISCV of around 26%. The actual sample size come to be around 138. Best Regards Sam Edit: Standard quotes restored. [Helmut] |
|
Helmut ★★★ Vienna, Austria, 2013-08-01 15:18 (4300 d 07:09 ago) @ sam Posting: # 11147 Views: 26,039 |
|
|
Hi Sam, ❝ In my study there were five subjects whose data for the test products are not in concurrence to the other subjects data. Can you please be more specific. What do you mean by “not in concurrence to the other subjects data”? ❝ Should I remove the AUCt parameter only for that subjects which have AUCextrap value more than 20%… No. You should evaluate the study based on the statistical analysis plan / clinical study protocol. Everything else (post hoc exclusion) is cherry-picking and not acceptable. You can exclude this subjects only in an additional evaluation and present it as a sensitivity analysis. ❝ … write a comment on the report as per our suggestions. I would try that (i.e., AUCt is not a reliable/relevant metric of absorption for ER formulations). Back it up with a comparison of half lives. ❝ Since the ratio in the pilot study was around 84% and considering the ISCV of around 26%. The actual sample size come to be around 138. Can you elaborate? With T/R of 84% and a CV of 26% you would have needed 342 (!) subjects to obtain 80% power (and still 260 for 70% power – which is the lowest any serious IEC should accept, IMHO). Was your target power only 46% – knowingly running another “casino-type” study? Bad. — Dif-tor heh smusma 🖖🏼 Довге життя Україна! Helmut Schütz The quality of responses received is directly proportional to the quality of the question asked. 🚮 Science Quotes |
|
sam ★ India, 2013-08-01 15:40 (4300 d 06:47 ago) @ Helmut Posting: # 11149 Views: 25,717 |
|
|
Hi Helmut, ❝ Can you please be more specific. What do you mean by “not in concurrence to the other subjects data”? Means the T/R ratio of these subjects are in between 23-35%. But other subjects have a ratio from 60-100% ❝ You can exclude this subjects only in an additional evaluation and present it as a sensitivity analysis. Please suggest some test which we can apply for the exclusion of AUCt parameter for only two subjects. ❝ I would try that (i.e., AUCt is not a reliable/relevant metric of absorption for ER formulations). Back it up with a comparison of half lives. ok ❝ Can you elaborate? With T/R of 84% and a CV of 26% you would have needed 342 (!) subjects to obtain 80% power (and still 260 for 70% power – which is the lowest any serious IEC should accept, IMHO). Was your target power only 46% – running another “casino-type” study? Bad. I Really apologize for the mistake, the ratio is 84.71 and CV for the pilot study is 19.11. |
|
jag009 ★★★ NJ, 2013-08-01 22:04 (4300 d 00:24 ago) @ sam Posting: # 11151 Views: 25,816 |
|
|
❝ ❝ Can you elaborate? With T/R of 84% and a CV of 26% you would have needed 342 (!) subjects to obtain 80% power (and still 260 for 70% power – which is the lowest any serious IEC should accept, IMHO). Was your target power only 46% – running another “casino-type” study? Bad. ❝ ❝ I Really apologize for the mistake, the ratio is 84.71 and CV for the pilot study is 19.11. With ISCV=19.11, T/R ratio 84.71, n~140 at 80% power so you were within range with your pivotal study. If outlier test confirms then one option is to do a rechallenge study -> Run another small study with the questionable subjects along with a few controls. I assume you tweaked the formulation based on the pilot study data (Ratio 84.71) before going pivotal? Going into a pivotal study with the same formulation is a bit risky IMHO... John |
|
Helmut ★★★ Vienna, Austria, 2013-08-02 05:32 (4299 d 16:56 ago) @ jag009 Posting: # 11158 Views: 25,782 |
|
|
Hi John, ❝ If outlier test confirms then one option is to do a rechallenge study -> Run another small study with the questionable subjects along with a few controls. Do you have a reference? If I recall it correctly the sample size should be six or 20% of subjects – whichever is larger. 26 ≠ small, IMHO.  — Dif-tor heh smusma 🖖🏼 Довге життя Україна! Helmut Schütz The quality of responses received is directly proportional to the quality of the question asked. 🚮 Science Quotes |
|
jag009 ★★★ NJ, 2013-08-02 07:10 (4299 d 15:18 ago) @ Helmut Posting: # 11159 Views: 25,608 |
|
|
Hi Helmut, ❝ Do you have a reference? If I recall it correctly the sample size should be six or 20% of subjects – whichever is larger. 26 ≠ small, IMHO. 26 < 124. You know what I mean by small (relative wise)... I need to dig up that particular report of mine (not mine, someone ran it before my time). John |
|
Helmut ★★★ Vienna, Austria, 2013-08-02 16:44 (4299 d 05:43 ago) @ jag009 Posting: # 11168 Views: 25,582 |
|
|
Hi John, ❝ ❝ Do you have a reference? If I recall it correctly the sample size should be six or 20% of subjects – whichever is larger. 26 ≠ small, IMHO. ❝ ❝ 26 < 124. You know what I mean by small (relative wise)... You noted the irony smiley? Of course if one is used to run studies in 100+ subjects 26 are peanuts. ❝ I need to dig up that particular report of mine (not mine, someone ran it before my time). Yes, please. AFAIK, the redosing sample size was never an official recommendation by the FDA. I have seen it in a few documents under FOI. Maybe I have them somewhere, but since they are all scanned PDFs no way of searching. — Dif-tor heh smusma 🖖🏼 Довге життя Україна! Helmut Schütz The quality of responses received is directly proportional to the quality of the question asked. 🚮 Science Quotes |
|
sam ★ India, 2013-08-02 09:00 (4299 d 13:28 ago) @ jag009 Posting: # 11160 Views: 25,642 |
|
|
Hi John ❝ If outlier test confirms then one option is to do a rechallenge study -> Run another small study with the questionable subjects along with a few controls. Will the USFDA accept the data after exclusion of the subjects if the re-dosing confirms the outlier. Regards Sam |
|
Helmut ★★★ Vienna, Austria, 2013-08-02 05:22 (4299 d 17:06 ago) @ sam Posting: # 11157 Views: 26,038 |
|
|
Hi Sam, ❝ Please suggest some test which we can apply for the exclusion of AUCt parameter for only two subjects. No idea. ❝ ❝ Can you elaborate? With T/R of 84% and a CV of 26% you would have needed 342 (!) subjects to obtain 80% power (and still 260 for 70% power – which is the lowest any serious IEC should accept, IMHO). Was your target power only 46% – running another “casino-type” study? Bad. ❝ ❝ I Really apologize for the mistake, the ratio is 84.71 and CV for the pilot study is 19.11. Really? If your ratio and CV are just 0.02% (!!) worse, your power would already be <80%. Remember that outcomes of pilot studies are estimates, not set in stone. Planing for 80% power – hoping for exactly matching the pilot and face not a single drop-out in such a large study – is gambling as well.*
— Dif-tor heh smusma 🖖🏼 Довге життя Україна! Helmut Schütz The quality of responses received is directly proportional to the quality of the question asked. 🚮 Science Quotes |
|
sam ★ India, 2013-08-02 09:02 (4299 d 13:26 ago) @ Helmut Posting: # 11161 Views: 25,588 |
|
|
❝ Really? If your ratio and CV are just 0.02% (!!) worse, your power would already be <80%. Remember that outcomes of pilot studies are estimates, not set in stone. Planing for 80% power – hoping for exactly matching the pilot and face not a single drop-out in such a large study – is gambling as well.* Its real. ❝ * In the pivotal study PE ↓ 83.82%, CV ↑ 27%, and you had 14 drop-outs… very correct. Also confirm me if i do redosing study and it confirms the outlier. should i present the data without that subjects as discussed by John. Regards Sam |
|
luvblooms ★★ India, 2013-08-02 10:37 (4299 d 11:50 ago) @ sam Posting: # 11163 Views: 25,664 |
|
|
Dear Sam ❝ Its real. I must say a very very BOLD decesion to go ahead with the study!! ❝ Also confirm me if i do redosing study and it confirms the outlier. should I present the data without that subjects as discussed by John. Now coming to your redosing study, you can do it but it is very difficult to justify and sometime inconclusive as well. We have done it once for USFDA and really regreted it. In our study we found a unique situation, where the one volunteer who behaved abruptly in the first study was ok but other 2 volunteers decided to act wildly and the total study was inconclusive. But since you have already taken some bold steps, you can try the redosing process and if lucky, submit the data with redosing study but you need to define it properly in the redosing protocol. And If I am right, you have to submit the full study results. I mean to say
BTW, was the study fasting or fed? And what was the outcome of other study (fasting/fed)? Better — ~A happy Soul~ |
|
sam ★ India, 2013-08-02 11:01 (4299 d 11:27 ago) @ luvblooms Posting: # 11164 Views: 25,644 |
|
|
Hi Luvblooms, ❝ I must say a very very BOLD decision to go ahead with the study!! I don’t have other option as the management wanted to go with this study ❝ But since you have already taken some bold steps, you can try the redosing process and if lucky, submit the data with redosing study but you need to define it properly in the redosing protocol. Suppose I am doing re-dosing study and it meets all the acceptance. I am worried will USFDA accept the dossier as they don’t encourage the same. ❝ And If I am right, you have to submit the full study results. ❝ I mean to say ❝ a. Original study data including outlier ❝ b. You cannot remove the outlier from study, If you are able to find your proposed outlier as real outlier in redosing study, the data of the outliers in the original study to be replaced with the data obtained in the redosing study and then BE to be calculated. I don’t think the redosing study is performed to get the new data which can be used for substitution of the old data, as per my understanding it is only done to prove that the detected outlier is really outlier and the outlier will proved than you have the submit the data with and without outlier. ❝ BTW, was the study fasting or fed? And what was the outcome of other study (fasting/fed)? Its fed study and we are waiting the fasting study results. Best Regards Sam |
|
Helmut ★★★ Vienna, Austria, 2013-08-02 16:53 (4299 d 05:35 ago) @ sam Posting: # 11169 Views: 25,642 |
|
|
Hi Sam, ❝ I don’t have other option as the management wanted to go with this study Splendid! Given your results chances that the product does not show a lower BA than the reference are ~0.00005%. Four questions:
— Dif-tor heh smusma 🖖🏼 Довге життя Україна! Helmut Schütz The quality of responses received is directly proportional to the quality of the question asked. 🚮 Science Quotes |
|
jag009 ★★★ NJ, 2013-08-02 19:25 (4299 d 03:03 ago) @ Helmut Posting: # 11174 Views: 25,611 |
|
|
Hi Helmut! ❝ ❝ I don’t have other option as the management wanted to go with this study ❝ ❝ Splendid! Given your results chances that the product does not show a lower BA than the reference are ~0.00005%. Four questions: ❝ • Would you take the drug yourself? ❝ • Would you give it to your wife and/or children? ❝ • Would your boss take it him/herself? ❝ • Would your boss give it to his wife / her husband and/or children? Hey they pay for for my bills, food, sports car etc etc  Remember I told you about my case a month or so ago? I had to deal with a drug product having ISCV~60% and T/R Ratio ~120% from a pilot study! They wanted me to go for a pivotal and also low-balled my sample size recommendation to ~70% of the computed sample size! Found the rechallenge study... Yes it was accepted by the FDA and the original results were very similar to what Sam has (79.xx% as the lower limit of the 90% CI, in my case it was Cmax). Here goes: Original fasting study was n=47, rechallenge was n=8 with 5 questionable subjects + 3 controls; The 5 questionable subjects were suspected as outliers because their Cmax ratios were greater than ± 2 standard deviations from the mean ratio. AUCs were fine. All subjects were from the original study population. The rechallenge study was successful so the original study data was re-evaluated with n=42 (minus the 5 weirdos) and bioequivalence was concluded. How did we conclude that the 5 weirdos were enigmatic? Well their ratio observed from the rechallenge study were within ± 2 SD of the mean observed from the original study. Note that this assessment was made for Cmax and AUCs ratios. So n=8/n=47 → ~ 17% of the original study population. The final report has 3 parts: Final bioequivalent results after removal of outliers (n=42) with explanation of why the removal was justified, original results (n=47), rechallenge results (n=8) Mind you this study didn't fail because of incomplete AUCt though. The drug product was not highly variable, CV was about 18%. Why the criteria was ± 2 SD? I don't have the real justification and there's none written on the report. Now I don't know if there was communication with FDA(OGD) before we went ahead to conduct the rechallenge study though since this happened before I joined the company. In Sam's case, if it is in my shoes, I would go ask FDA first and tell them why I suspect the AUCt issue was attributed to subject variation rather than inadequate sampling time period since AUCt was sufficiently captured in almost all subjects (yes present the AUCt/AUCinf ratio, show a table of the half-life values, throw all possible presentations that you can think of including a kitchen sink). I would do this rather than spending the $ on a rechallenge study and later on receiving a middle finger salute from the FDA (Mean haha tough luck, we don't accept it!). Maybe FDA will buy your reason(s) for the AUCt problem and accept the analysis excluding those subjects. Hope this helps John |
|
Helmut ★★★ Vienna, Austria, 2013-08-02 20:11 (4299 d 02:17 ago) @ jag009 Posting: # 11176 Views: 25,669 |
|
|
Hi John! ❝ ❝ Splendid. Given your results chances that the product does not show a lower BA than the reference are ~0.00005%. Four questions: […] ❝ Hey they pay for for my bills, food, sports car etc etc I’m fine with most of it. Still I think that if we are not soooo much interested in ethics a – really better paid – job would be investment banker. BTW, not by any chance the first (!) question I was asked in my first workshop in India was: “Would you advise to open up a new CRO to make a lot of money?” Telling! ❝ Remember I told you about my case a month or so ago? I had to deal with a drug product having ISCV~60% and T/R Ratio ~120% from a pilot study! They wanted me to go for a pivotal and also low-balled my sample size recommendation to ~70% of the computed sample size! This story? Did you tell the guy in his Armani suit buying a new Lamborghini once its ash tray if full what that means in terms of power? Thanks for your comprehensive example. ±2SD covers 95.45% of data – so such a cut-off is not completely out of the blue. BTW, Canada considers an outlier if the studentized intra-subject residual is >±3SD (i.e., covering 99.73% of data). — Dif-tor heh smusma 🖖🏼 Довге життя Україна! Helmut Schütz The quality of responses received is directly proportional to the quality of the question asked. 🚮 Science Quotes |
|
jag009 ★★★ NJ, 2013-08-02 22:11 (4299 d 00:17 ago) @ Helmut Posting: # 11178 Views: 25,536 |
|
|
Hi Helmut, ❝ This story? Did you tell the guy in his Armani suit buying a new Lamborghini once its ash tray if full what that means in terms of power? Oh he knows but what can he do? He has to please the gods above him too. By the way he drives a porsche... ❝ Thanks for your comprehensive example. ±2SD covers 95.45% of data – so such a cut-off is not completely out of the blue. BTW, Canada considers an outlier if the studentized intra-subject residual is >±3SD (i.e., covering 99.73% of data). But Canada does accept removal of outliers, correct? I don't think FDA does and will ask for a rechallenge study if you can convince them that there are outliers. John |
|
Helmut ★★★ Vienna, Austria, 2013-08-02 22:56 (4298 d 23:32 ago) @ jag009 Posting: # 11180 Views: 25,632 |
|
|
Hi John, ❝ ❝ This story? Did you tell the guy in his Armani suit buying a new Lamborghini once its ash tray if full what that means in terms of power? ❝ Oh he knows but what can he do? Convert to a different faith. ❝ He has to please the gods above him too. I recommend the Church of the Flying Spaghetti Monster. ❝ By the way he drives a porsche... A Stretch Volkswagen Beetle on steroids? ❝ But Canada does accept removal of outliers, correct? Yes. A strategy should be stated in the protocol and re-testing of subjects is not recommended. ❝ I don't think FDA does and will ask for a rechallenge study if you can convince them that there are outliers. Duno. You are the expert. — Dif-tor heh smusma 🖖🏼 Довге життя Україна! Helmut Schütz The quality of responses received is directly proportional to the quality of the question asked. 🚮 Science Quotes |
|
jag009 ★★★ NJ, 2013-08-03 00:23 (4298 d 22:04 ago) @ Helmut Posting: # 11181 Views: 25,635 |
|
|
Hi Helmut, ❝ ❝ Oh he knows but what can he do? ❝ ❝ Convert to a different faith. ❝ ❝ ❝ He has to please the gods above him too. His god is printed in green color ❝ ❝ By the way he drives a porsche... ❝ A Stretch Volkswagen Beetle on steroids? That's the typical joke from a sales person when you visit a BMW/Merc dealership here. ❝ ❝ I don't think FDA does and will ask for a rechallenge study if you can convince them that there are outliers. ❝ ❝ Duno. You are the expert. No expert, just want to help out |
|
Helmut ★★★ Vienna, Austria, 2013-08-03 12:11 (4298 d 10:17 ago) @ jag009 Posting: # 11184 Views: 25,473 |
|
|
Hi John, ❝ ❝ You are the expert. ❝ ❝ No expert, just want to help out Don’t hide your light under a bushel! With 340 posts you are both Americas’ champ. I appreciate your firsthand accounts. My experience with the FDA is limited. — Dif-tor heh smusma 🖖🏼 Довге життя Україна! Helmut Schütz The quality of responses received is directly proportional to the quality of the question asked. 🚮 Science Quotes |
|
jag009 ★★★ NJ, 2013-08-03 00:29 (4298 d 21:58 ago) @ Helmut Posting: # 11182 Views: 25,529 |
|
|
Hi Helmut, ❝ Still I think that if we are not soooo much interested in ethics a – really better paid – job would be investment banker. BTW, not by any chance the first (!) question I was asked in my first workshop in India was: “Would you advise to open up a new CRO to make a lot of money?” Too many regulations being/in the process of being implemented. Go start somewhere else like some small countries in Africa... OTOH (On the other hand), I heard that it is next to impossible to run studies in china unless the drug of interest exists and is currently marketed in China. The chinese government will not approve your trial or hold onto your request forever. However, I have heard shady stories that some will run the studies anyway. As long as no one dies, the government doesn't follow up... → Meaning no submission studies since FDA/others will come and audit. John |
|
ElMaestro ★★★ Denmark, 2013-08-03 01:08 (4298 d 21:19 ago) @ jag009 Posting: # 11183 Views: 25,546 |
|
|
Oh dear, ❝ I heard that it is next to impossible to run studies in china unless the drug of interest exists and is currently marketed in China. The chinese government will not approve your trial or hold onto your request forever. However, I have heard shady stories that some will run the studies anyway. As long as no one dies, the government doesn't follow up... → Meaning no submission studies since FDA/others will come and audit. Somebody, quick, please grab a baseball bat and put me out of my misery. — Pass or fail! ElMaestro |
|
jag009 ★★★ NJ, 2013-08-04 01:00 (4297 d 21:27 ago) @ ElMaestro Posting: # 11191 Views: 25,421 |
|
|
Ho ho ho, ❝ Somebody, quick, please grab a baseball bat and put me out of my misery. That's nothing. Did you hear the latest fiasco in China with the police arresting GSK management and others because of bribery involving $ and sex? John |
|
luvblooms ★★ India, 2013-08-05 08:02 (4296 d 14:26 ago) @ jag009 Posting: # 11196 Views: 25,337 |
|
|
Dear John ❝ Original fasting study was n=47, rechallenge was n=8 with 5 questionable subjects + 3 controls; The 5 questionable subjects were suspected as outliers because their Cmax ratios were greater than ± 2 standard deviations from the mean ratio. AUCs were fine. All subjects were from the original study population. The rechallenge study was successful so the original study data was re-evaluated with n=42 (minus the 5 weirdos) and bioequivalence was concluded. How did we conclude that the 5 weirdos were enigmatic? Well their ratio observed from the rechallenge study were within ± 2 SD of the mean observed from the original study. Note that this assessment was made for Cmax and AUCs ratios. Shouldn't it be Mean ± 3 SD as per the 68–95–99.7 rule or three-sigma rule or empirical rule????  — ~A happy Soul~ |
|
jag009 ★★★ NJ, 2013-08-05 22:14 (4296 d 00:14 ago) @ luvblooms Posting: # 11215 Views: 25,259 |
|
|
Hi Luvbloom, ❝ Shouldn't it be Mean ± 3 SD as per the 68–95–99.7 rule or three-sigma rule or empirical rule???? I dunno. The study was done and submitted before my time here... John |
Ing. Helmut Schütz