opc
Lab Rat
Posts: 30
|
Post by opc on Aug 19, 2013 12:23:40 GMT -5
Opc The difference between the groups (C-T) was 0.37-0.13=0.24. The CI from the poster was 0.08 to 0.40, which 0.24+/-0.16. The margin was 0.40. 0.40=0.40. I believe it needed to be less than the margin, not less than or equal to the margin. Read the guidance. The fact that 0.24 falls between 0.08 to 0.40 is irrelevant. The mean always falls between its CI bounds J. HTH jn Ok, this is getting old. So let's back up. You previously stated the poster did not state "non-inferiority." In fact it did in the 'RESULTS' section: "TI was noninferior." Then you said the CI from the poster was .08 to .40. In the 'RESULTS' section and I believe in an earlier post of yours, it states -.11 to -.38. I see you are getting the results from the tables only. Perhaps reading more would help as the tables are rounding results. I can't explain why the same document is showing different numbers in different places, but it doesn't matter because the differences are small/immaterial. What matters is that the difference between the two groups was either -.24 (poster table/company PR) or -.28 ((8.41-8.20) vs (8.48-7.99)) depending upon where you get your info. Most importantly, you are citing an incorrect margin. It was .06 according to the poster table; NOT .16. Regardless of how one arrives at the differences (exclusive of your margin mistake), BOTH are within the CI range. |
|
|
|
Post by johnnicholas14 on Aug 19, 2013 13:12:31 GMT -5
I am quoting from the first table in the poster "End of Study Efficacy, etc"
The Delta (change) in HbA1c % MNKD=0.13 +/- 0.06 Comparator 0.37 +/-0.06
The mean difference between the groups 0.37-0.13 = 0.24. The 95% CI for this difference was 0.08 to 0.40 (Yes, that's 0.24 +/-0.16) using the analysis of covariance (ANCOVA) to express the CI. ANCOVA is the stat model of record and would be what the FDA would recognize for non-inferiority. The press release cited these data, not the 0.11 to 0.38 data.
The results section that you cite, that the difference would lie between 0.11 and 0.38 uses a different model, the mixed-model method (MMRM), however, I do not believe the FDA would accept that model. Or let me put it to you another way: Would the FDA responded with the CRL had non-inferiority actually been met? There are no safety issues with Afrezza, but the FDA had 'clinical utility' concerns, and mgmt had repeatedly stated in conferences and quaterly CCs that the FDA had trouble with Study 009. And they quickly quit using "non inferiority, etc." To describe the data after the first CRL.
|
|
|
|
Post by mdcenter61 on Aug 19, 2013 13:41:48 GMT -5
Well, John, (yawn), looks like you certainly have your particular axe to grind or point to make, which is that you do not feel that the FDA will accept these results due to clinical utility. Could you please clarify your investment stance on this Company? Personally, I don't believe that Deerfield makes capricious multi-million dollar investments nor does Greenhill take on "losers", and since they actually are in the business of making money on these deals and not lurking on message boards I believe I will stand pat with my investment. Thanks for your input anyway.
|
|
opc
Lab Rat
Posts: 30
|
Post by opc on Aug 19, 2013 13:52:03 GMT -5
well aware of where both of us are getting data from. point is that whether you used one or the other, the data still evidenced non-inferiority imo.
Thank you for pointing out the two different models utilized. i've learned something new. can you explain why you assume FDA wouldn't consider the alternative model?
Still, i have no idea where you are getting +/-.16 from...
as to clinical utility, i have seen nothing that connects the FDA's clinical utility concerns w/ 009 data other than a lot of third party speculators commingling mgmt discussions. Peter Richardson stated at the time, "To remind you, there were three substantive areas for which the agency raised questions; clinical utility, or where does the project and the product fit into the therapeutic armamentarium with patients with diabetes; the assay methodology used in our study to demonstrate bio equivalency of the inhaler used in our pivotal clinical studies and that to me marketed; and finally, technical aspects of the requirements of the device, such as how often is it changed and how we mark the amount of drug and identify [ph] each cartridge. As we indicated to you in our previous updates, we are confident that we will be able to address these on the basis of data we already had available, or in the process of generating as part of our previously agreed development of our next generation device, formerly known internally as Dreamboat. We were able to have a very helpful and productive discussion in the meeting with the agency on June 9, during which we were able to establish that the data would be sufficient for them to resume the review, based upon the additional clinical data from Study 117."
If you are suggesting afrezza was inferior and mgmt lied, then i suggest bringing the concerns to the SEC. Last, this data was presented to the FDA a second time and was not the reason for the 2nd crl. clinical utility was not an issue in the second crl. i completely disagree w/ your assumptions.
|
|
|
|
Post by StevieRay on Aug 19, 2013 13:53:50 GMT -5
In all fairness I appreciate open and frank discussions even if they do or especially if they do raise doubt. I believe this is just one of MANY data points the FDA will consider in making their decision. Like a jury weighs ALL the data points before making their decision. How much weight this particular aspect of the 009 study carries I'm not sure but I know the most recent studies that used different titrations methods clearly showed improvements across the board. I believe the FDA will review everything in TOTAL and not attempt to draw conclusions based on some narrow marginal data point.
|
|
|
|
Post by StevieRay on Aug 19, 2013 13:57:17 GMT -5
|
|
opc
Lab Rat
Posts: 30
|
Post by opc on Aug 19, 2013 14:07:55 GMT -5
www.news.mannkindcorp.com/phoenix.zhtml?c=147953&p=irol-newsArticle&ID=1402294&highlight="The Complete Response letter related to the AFREZZA application requested several items, including information and currently available clinical data that support the clinical utility of AFREZZA and information about the comparability of the commercial version of the MedTone inhaler to the earlier version of this device that was used in pivotal clinical trials. " You are making my point. From Q1'10 Richardson states, "In addition to the technical questions, you will no doubt recall that we have been asked to discuss the so-called clinical utility of the product on the basis of currently available data. Because of the rather unusual wording of the request, we saw clarification in the agency and replied with the following, and I quote from the email.”We are basically asking you to clarify where AFREZZA fits in the treatment armamentarium to diabetes," language very similar to that which I use at our update last month." The data is a separate issue from the clinical utility question IMO. |
|
opc
Lab Rat
Posts: 30
|
Post by opc on Aug 19, 2013 14:11:50 GMT -5
Yes. This is a point I made earlier. |
|
|
|
Post by StevieRay on Aug 19, 2013 14:14:18 GMT -5
Opc, thanks for your patience and working through this. The discussion is very helpful.
|
|
|
|
Post by johnnicholas14 on Aug 19, 2013 14:18:14 GMT -5
mdcenter If you choose to believe me: I am long MNKD. Also, I have a BS in Chemistry and an MBA. I have been in biotech for 20+ years. I have about six semesters worth of stats in my edu background, and have a co-worker with a stats PhD who is familiar with submissions to consult if I have any questions  . I think MKC met their primary end-point for Study-171, but just barely, and the data leave a lot to be desired. The study was designed for a ‘home run,’ but ended up being a ‘suicide squeeze.’ I agree with your point about Deerfield. Also, I would add it is very difficult for Afrezza, with its PK profile to lower HbA1c on its own with Type Ones; But that’s what it has to do for approval I do not think that the end-point was reached with Study 009. I happen to think ‘clinical utility’ means positive outcomes with HbA1c. what else could it mean? It was a huge red flag. Study 009 did not make that case very well. Opc: You can’t tell where I got 0.16 from? 0.40 to 0.08, is the confidence interval for the true difference between test arms in Study 009. The mean difference between test arms was 0.37 - 0.13= 0.24. The confidence interval for 0.24 is plus, or minus, 0.16. That’s how you end up with 0.08 and 0.40 for the CI. SR: Thanks for the comment. Study 009 raised doubt where it should have erased it! It’s only relevant today because when I saw the data for Study 171, it was so similar to Study 009. Where Study 009 just missed its non-inferior end point, Study 171 just barely met it. HTH jn |
|
|
|
Post by mdcenter61 on Aug 19, 2013 14:27:21 GMT -5
John,
Will take you at your word; I do appreciate your input and I recognize that all data, good and bad, need to be discussed and considered. I'm definitely a little jaded after the fiasco of the last few days and the abuse/manipulation through message boards and media.
|
|
opc
Lab Rat
Posts: 30
|
Post by opc on Aug 19, 2013 14:35:45 GMT -5
mdcenter If you choose to believe me: I am long MNKD. Also, I have a BS in Chemistry and an MBA. I have been in biotech for 20+ years. I have about six semesters worth of stats in my edu background, and have a co-worker with a stats PhD who is familiar with submissions to consult if I have any questions . I think MKC met their primary end-point for Study-171, but just barely, and the data leave a lot to be desired. The study was designed for a ‘home run,’ but ended up being a ‘suicide squeeze.’ I agree with your point about Deerfield. Also, I would add it is very difficult for Afrezza, with its PK profile to lower HbA1c on its own with Type Ones; But that’s what it has to do for approval I do not think that the end-point was reached with Study 009. I happen to think ‘clinical utility’ means positive outcomes with HbA1c. what else could it mean? It was a huge red flag. Study 009 did not make that case very well. Opc: You can’t tell where I got 0.16 from? 0.40 to 0.08, is the confidence interval for the true difference between test arms in Study 009. The mean difference between test arms was 0.37 - 0.13= 0.24. The confidence interval for 0.24 is plus, or minus, 0.16. That’s how you end up with 0.08 and 0.40 for the CI. SR: Thanks for the comment. Study 009 raised doubt where it should have erased it! It’s only relevant today because when I saw the data for Study 171, it was so similar to Study 009. Where Study 009 just missed its non-inferior end point, Study 171 just barely met it. HTH jn non-inferior means EQUAL TO or within an acceptable range. The difference between the two groups was .24. That's within the range of .08 to .40. Period. The range around it is .16 to establish the CI. And I'll repeat it again, since the quotes I've provided along with everything else hasn't quite set in. The same data was submitted a second time along with additional data. Clinical utility was not raised in the second CRL. ....I think we've already learned educated/smart people are still subject to making mistakes. Al has demonstrated this a couple times. |
|
|
|
Post by spiro on Aug 19, 2013 14:36:18 GMT -5
John,
If you believe your interpretation of studies 009 and 171 are correct, it makes no sense for you to be long MNKD. Not that it matters to you, but I personally doubt your sincerity in this discussion. It seems to me your spin on the results has a hidden agenda. I also do not think you have been successful in your attack on the overwhelming supportive data that should result Afrezza's FDA approval.
|
|
|
|
Post by johnnicholas14 on Aug 19, 2013 15:08:52 GMT -5
Opc: please read the FDA Guidance on Non Inferiority. 0.24 is not the data point of interest, it's 0.40, the upper confidence bound. That number equals the margin the FDA suggested for the study. Consult the guidance.
Spiro: I feel the company just barely met its end point for Study 171, and this result may be sufficient for approval. It helps that Deerfield signed onto the results.
Douge: Wow.
Jn
|
|
|
|
Post by liane on Aug 19, 2013 15:19:05 GMT -5
Just a friendly reminder to everyone on this thread to keep the discussion respectful. Thank you!
|
|
.
