Log in
Register|
sarawut ☆ Thailand, 2012-05-08 14:36 (5146 d 16:23 ago) Posting: # 8531 Views: 9,152 |
|
|
Dear all, I have found a BE study in which the subjects were separated into 5 groups (I think by subject convenience, there were 5 starting dates) and the washout intervals were not identical (not the same in each subject but at least 7 days) (also I think by subject's convenience). It is a standard 2x2x2 BE design for a total of 44 subjects with balanced arm. Do you think it is acceptable if the test of group effect is not significant. Also is it possible to test group effect in this study (5 groups)? Any thought is welcome. Sarawut |
|
Helmut ★★★   Vienna, Austria, 2012-05-08 15:40 (5146 d 15:19 ago) @ sarawut Posting: # 8532 Views: 6,888 |
|
|
Dear Sarawut! ❝ […] the washout intervals were not identical (not the same in each subject but at least 7 days) (also I think by subject's convenience). Personally I have seen a maximum of three groups – but I know that more have been used, especially for HVDs in the ages before scaling (sample sizes >100). Unequal washouts (let’s say 8 and 10 days if the protocol specified at least 7 days) are not that uncommon (e.g., if one of the treatment days would be a public holiday). Never heard about any problems with regulators. ❝ It is a standard 2x2x2 BE design for a total of 44 subjects with balanced arm. ❝ Do you think it is acceptable if the test of group effect is not significant. ❝ Also is it possible to test group effect in this study (5 groups)? Three questions. Answers: yes, yes, and why not? See also this post including a reference. — 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, 2012-05-08 17:02 (5146 d 13:56 ago) @ Helmut Posting: # 8534 Views: 6,839 |
|
|
Dear Helmut! ❝ ❝ It is a standard 2x2x2 BE design for a total of 44 subjects with balanced arm. ❝ ❝ Do you think it is acceptable if the test of group effect is not significant. ❝ ❝ Also is it possible to test group effect in this study (5 groups)? ❝ ❝ Three questions. Answers: yes, yes, and why not? I see only 2. Maybe I have to clear my glasses  ?Or can't see the trees for all the wood  .— Regards, Detlew |
|
Helmut ★★★ Vienna, Austria, 2012-05-08 17:38 (5146 d 13:21 ago) @ d_labes Posting: # 8535 Views: 6,841 |
|
|
Dear Detlew! ❝ ❝ Three questions. Answers: yes, yes, and why not? ❝ ❝ I see only 2. Maybe I have to clear my glasses
 …— 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 |
|
Ohlbe ★★★ France, 2012-05-09 13:15 (5145 d 17:44 ago) @ Helmut Posting: # 8538 Views: 6,782 |
|
|
Dear Helmut, The original text was ❝ It is a standard 2x2x2 BE design for a total of 44 subjects with balanced arm. not ❝ It is a standard 2x2x2 BE design for a total of 44 subjects with balanced arm? "It is", not "Is it". I agree with Detlew Labes... SCNR Regards Ohlbe — Regards Ohlbe |
|
Helmut ★★★ Vienna, Austria, 2012-05-09 14:28 (5145 d 16:31 ago) @ Ohlbe Posting: # 8541 Views: 6,782 |
|
|
Dear Ohlbe & Detlew! ❝ "It is", not "Is it". I agree with Detlew Labes... Oh no! I should clean my contact lenses. — 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 |
|
sarawut ☆ Thailand, 2012-05-08 18:49 (5146 d 12:09 ago) @ Helmut Posting: # 8536 Views: 6,923 |
|
|
Dear Helmut, Thanks for your response. Personally, I have seen the study with 2 equal groups because of the limitation of facility. That is reasonable coz to handle more than 24 subjects at the same day is quite tough. However for this study, I am just curious that why they didn't try to a maxium capacity of their facility. The number of subjects in each 5 groups are not the same. Each subject could have washout interval for 7, 8, 9, 10 days. I have just seen it. But it seem OK if there is no group effect.  Sarawut |
|
boonchai_l ☆ Thailand, 2012-05-16 14:19 (5138 d 16:40 ago) @ sarawut Posting: # 8573 Views: 6,772 |
|
|
SA-WAS-DEE Sarawut, Let see announcement number สธ1009.4/ว.11243 from Thai FDA date 13 Sep 2011.
Accoulding to your question, ❝ Do you think it is acceptable if the test of group effect is not significant. The answer is in #3. Regards, Boonchai_L |
|
Helmut ★★★ Vienna, Austria, 2012-05-16 15:01 (5138 d 15:58 ago) @ boonchai_l Posting: # 8574 Views: 6,679 |
|
|
สวัสดี Boonchai! ❝ Let see announcement number สธ1009.4/ว.11243 from Thai FDA date 13 Sep 2011. Interesting. If the announcement is available online can you provide us the URL?
— 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 |
|
boonchai_l ☆ Thailand, 2012-05-16 15:37 (5138 d 15:22 ago) @ Helmut Posting: # 8575 Views: 6,658 |
|
|
Ha Ha Ha amazing for your สวัสดี HS!, ❝ Interesting. If the announcement is available online can you provide us the URL? http://drug.fda.moph.go.th/zone_bioequivalence/files/doc210.pdf I will reply your comment tomorrow because I have to go now. Regards, Boonchai_l |
|
Helmut ★★★ Vienna, Austria, 2012-05-16 15:55 (5138 d 15:04 ago) @ boonchai_l Posting: # 8576 Views: 6,625 |
|
|
Dear Boonchai! ❝ Ha Ha Ha amazing for your สวัสดี HS!, Yeah – was in your nice country more than once… Took me ages to pronounce even simple words like ชา – almost – correctly.  ❝ http://drug.fda.moph.go.th/zone_bioequivalence/files/doc210.pdf Wow, a scan… Bad luck for people depending on ![[image]](img/uploaded/image219.png) Translate.— 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 |
|
boonchai_l ☆ Thailand, 2012-05-16 21:02 (5138 d 09:57 ago) @ Helmut Posting: # 8578 Views: 6,710 |
|
|
Dear HS, ❝ Yeah – was in your nice country more than once… Took me ages to pronounce even simple words like ชา – almost – correctly. Ha Ha Ha, it is usual for foreigners and thanks for the admiration my country. Refer to that announcement, let me add more 1 thing. 4. All groups must have equal number of subjects. BTW, refer to #3, why I said stupid? I have an example. If you are going to run BE study with high CV drug on 2x2 standard design and the number of subject, 80 to reach the desired power and your max capacity of the facility is 34. You decide to separate them into 3 groups, 27, 27 and 26 (because of #1). During the study, you have dropout 1, 2 and 4 subjects in each group, respectively. So you analyze 26, 25 and 22 in each and 73 subjects in total (Is it in conflict with #4?). However, you pool them to evaluate BE and the result shows equivalence with 84% power but the ANOVA found the significant group effect (It is possible to occur because the group effect is nested with subject effect, you never know which subject is good, poor, fast or slow metabolizer, assuming you bad luck in this case, the first group was filled with the slow metabolizers and the second group was filled with the fast metabolizers). So you can not pool them according to #3. The results from the separated analyses show inequivalence for all 3 evaluations with 60, 62 and 58% power. My question: Is it reasonable? Regards, Boonchai_l |
|
ElMaestro ★★★ Denmark, 2012-05-22 01:45 (5133 d 05:13 ago) @ boonchai_l Posting: # 8596 Views: 6,636 |
|
|
Dear Boonchai, I see your point. I also wouldn't be afraid of a significant group effect. I imagine your local regulators are perhaps seeing this in the same fashion as multicenter studies (??) where one or two centers stand out. That could indicate center lapses. But with multiple groups in XO BE studies, why bother? After all, in XO studies it is all about the intra-subject comparison. I can't see how a group effect in its own right can affect a conclusion of BE. Groups, as HS said, is a between-subject factor just like a harmless sequence effect. Subjects are nested in sequences nested in groups, sort of (but screw the nesting thing; number all subjects uniquely). In your example, the three separate analyses would probably not show inequivalence but just inconclusivity re. equivalence. This would be important to you. I would not consider #4 too seriously. Aim at balancing the subject dimension across groups and that should be it. Everyone knows that balance is a nice thing, but heck we cannot even guarantee balanced sequences ever. Perhaps drugs related to treatment of disease with seasonal variations could be an exception here (pulmonary steroids for example) since groups are separated in time. There is to the best of my limited knowledge no literature covering this area. — Pass or fail! ElMaestro |
|
boonchai_l ☆ Thailand, 2012-05-21 14:05 (5133 d 16:53 ago) @ Helmut Posting: # 8593 Views: 6,689 |
|
|
Dear all, Let me continue on this issue. I found a guideline, US FDA, Statistical Approaches to Establishing Bioequivalence, Section VII.A. They stated If a crossover study is carried out in two or more groups of subjects (eg, if for logistical reasons only a limited number of subjects can be studied at one time), the statistical model should be modified to reflect the multigroup nature of the study. In particular, the model should reflect the fact that the periods for the first group are different from the periods for the second group. This applies to all of the approaches (average, population, and individual BE) described in this guidance. If the study is carried out in two or more groups and those groups are studied at different clinical sites, or at the same site but greatly separated in time (months apart, for example), questions may arise as to whether the results from the several groups should be combined in a single analysis. Such cases should be discussed with the appropriate CDER review division. In my opinion, I agree because they did not link between the ANOVA results and data combining in a single analysis. They question for the data combining in only cases of 1. those groups are studied at different clinical sites, or 2. at the same site but greatly separated in time (months apart, for example). Any reason to support the concept that the data should not be combined when the group effect is significant, please share. If anyone ever found any guideline on this issue in any view, please share. Thanks in advance. Regards, boonchai_l Edit: Guidance linked. [Helmut] |
|
sarawut ☆ Thailand, 2012-06-05 08:02 (5118 d 22:57 ago) @ sarawut Posting: # 8658 Views: 6,376 |
|
|
Sawasdee all. Thank you for all comments. I'm sorry, I think my response to my topic is static. So I'm back late. Well, according to Thai FDA document on group effect. I think it used the word "SHOULD" in the statements. I agree with the document that 1) the number of groups should be minimized 2) the number of subjects in each group should be equal 3) the washout period should be the same 4) the period time in each group should be as close as possible I think the investigator should try to follow such design, however, some deviation can be allowed according to the protocol. It should not run according to the subject's convenience or investigator's or facitility's convenience, right? What do you think if suppose that a facility can deal with a maximum of 12 subjects. They plan for 40 subjects study. A minimum washout period was set in the protocol, let's say at least 5 half-lives. If the investigator set the number of groups, number of subjects per group, washout period of each subjects according to subject's convenience. It can end up with 8 groups, some group might have 2 subjects, each subjects can have different washout time. What do you think about this study? I agree with Mr. Boonchai for group effect, it may not need for combining the result. But I would open for more comments on this group effect. Sarawut |
Ing. Helmut Schütz