Helmut
★★★
avatar
Homepage
Vienna, Austria,
2015-07-20 19:34
(3173 d 21:23 ago)

Posting: # 15110
Views: 20,774
 

 Certara (Pharsight): Phoenix/WinNonlin licensing policy [Software]

Dear all,

Certara (the company behind Pharsight) changed their pricing policy for CROs (only!). Additionally to the annual single-user license fee (2,444 USD) CROs will be charged a fee based on the number of studies/projects performed per year… Part of the business-lingo I received:

As services providers use the Phoenix software on behalf of many sponsors and their compounds, they generally receive far greater value from these tools than companies using the software of their own com­pounds.

The premium is based on the number of preclinical or clinical studies the software is used in.

A Study/Project will be considered a discrete trial for a discrete indication for a discrete molecule. For cost purposes, multiple trial “options” or “iterations” or “multiple deliverables” against a discrete trial / indication / molecule would still be considered a single “study” or “project”.
[image]

 studies   fee USD    USD/study
  1 –  10    1,500   394 – 3,944
 11 –  50    7,500   199 –   904
 51 – 150   22,500   166 –   489
151 – 300   45,000   158 –   314


45 days before the end of the contract period, you will be asked to submit a certification document to reconcile the # of actual studies completed to those you purchased.

If you have gone over the Tier paid for, Certara would ask for the difference at that time.


For a small CRO (say ≤50 studies/year) that’s an increase in cost of 307%. Wow!

Another trick: Until recently the single user license was issued for a particular machine. Now it is issued for a particular user. If in the past two users ran the software (though not simultaneously) on the same machine, no problem. No way with the new “licensing model”. An easy way to double Certara’s revenue squeezed out from small CROs.

Dif-tor heh smusma 🖖🏼 Довге життя Україна! [image]
Helmut Schütz
[image]

The quality of responses received is directly proportional to the quality of the question asked. 🚮
Science Quotes
Lucas
★    

Brazil,
2015-07-20 21:32
(3173 d 19:25 ago)

@ Helmut
Posting: # 15111
Views: 19,029
 

 Certara (Pharsight): Phoenix/WinNonlin licensing policy

Hello guys.

Yep, I also receive this news recently, when I was speaking with Certara's agent in South America. They've increased the price considerably with this new policy, and they may lose market due to that. I do think Phoenix is a great software for BA/BE and use it now for many years, but IMHO this is a very risky strategy for them. We use here the floating licenses, that could be used in any computer connected to our server, one user per license but could be any user... Small CROs will be indeed squeezed. The license fee will be cheaper than this extra fee. And also, don't know how they'll account for the studies conducted, but CROs sometimes uses the software with investigational non commercial purposes... I wonder if those will count.
May be time to improve my R "almost non-existing" skills! :-D

Lucas
Helmut
★★★
avatar
Homepage
Vienna, Austria,
2015-07-21 03:42
(3173 d 13:15 ago)

@ Lucas
Posting: # 15112
Views: 18,845
 

 Considering to jump ship

Hi Lucas,

❝ […] they may lose market due to that. I do think Phoenix is a great software for BA/BE and use it now for many years, but IMHO this is a very risky strategy for them.


[image]Agree. I use it since the DOS-version (PCNONLIN 1) back in 1986 and WinNonlin 1 (1998). The license-fee increased expo­nentially (I have no earlier data in my files: from USD 949 in 2004 to 2,444 this year).

Even if I take the US’ inflation into account that’s +7.2% / year. Nice business model. Maybe I should seriously reconsider my consultancy fees…

❝ […] And also, don't know how they'll account for the studies conducted, …


I guess small CROs – which will not jump ship – will opt for the lowest fee (still an amazing cost increase of 75% from the 2014 fee of 2,254 USD). Then what?

[…] you will be asked to submit a certification document to reconcile the # of actual studies completed to those you purchased.

A “certification document”? Oh yes, I performed 100, but state 5. Will Certara hire a private dick sneaking into my office and check? C’mon! Or is there a hidden counter in the software which phones home every time I execute a workflow?

❝ … but CROs sometimes uses the software with investigational non commercial purposes... I wonder if those will count.


Yep. I regularly recalculate stuff posted here (~10/year). Apart from Pharsight’s employees I’m the top-poster in Certara’s Phoenix-Forum. Similar over there. Shall I pay for sumfink I do in my spare time at no costs? Gimme a break!

❝ May be time to improve my R "almost non-existing" skills! :-D


Why not? Many companies use PHX/WNL for NCA only (and run the stats in SAS). Package bear is an alternative for NCA in R.
Have your heard about RapidNCA? Seems that only the bloody linear trapezoidal method is implemented.

Dif-tor heh smusma 🖖🏼 Довге життя Україна! [image]
Helmut Schütz
[image]

The quality of responses received is directly proportional to the quality of the question asked. 🚮
Science Quotes
nobody
nothing

2015-07-21 11:19
(3173 d 05:39 ago)

@ Helmut
Posting: # 15114
Views: 18,715
 

 Considering to jump ship

❝ Or is there a hidden counter in the software which phones home every time I execute a workflow?


...and then there was the story of this S-plus CD I could not install some years later as the antivirus/firewall on my Microtrash machine complained about some malware...

Just saying...

Kindest regards, nobody
mittyri
★★  

Russia,
2015-07-21 11:46
(3173 d 05:11 ago)

@ Helmut
Posting: # 15115
Views: 18,710
 

 WNL+SAS=double price?

Hi Helmut!

❝ Many companies use PHX/WNL for NCA only (and run the stats in SAS).


I heard that some CRO's do that, could you explain it? Don't they trust the BEQ module in Phoenix/WinNonlin?
As you wrote here, there are some discrepancies between WNL partial tests and SAS Type III, but is it so important thing for final results? I don't think so...
Why are they paying the double price (SAS+WNL)?

Kind regards,
Mittyri
Helmut
★★★
avatar
Homepage
Vienna, Austria,
2015-07-21 14:29
(3173 d 02:29 ago)

@ mittyri
Posting: # 15117
Views: 19,043
 

 Double price = ½ the fun

Hi Mittyri,

❝ ❝ Many companies use PHX/WNL for NCA only (and run the stats in SAS).


❝ I heard that some CRO's do that, could you explain it?

  1. I see it in many publications and the majority (!) of reports moving around my desk.
    When I search the forum with the keyword “SAS” I get 1,000+ hits. SAS+bioequivalence gives 49,000 [image]-hits.
  2. I will try at the end (educated guesses / crystal ball gazing / tassology).

❝ Don't they trust the BEQ module in Phoenix/WinNonlin?


Duno. Given the posts in the forum asking for solutions in SAS maybe those members could answer this question?

❝ As you wrote here, there are some discrepancies between WNL partial tests and SAS Type III, but is it so important thing for final results? I don't think so...


Correct; doesn’t matter at all. The residual variance / means – and therefore, the PE and CI – are independent from this stuff.

❝ Why are they paying the double price (SAS+WNL)?


I guess it is a combination of tradition, using what they have already, etc. SAS is part of the bio­sta­tis­tical curriculum (though R is catching up). I know some (young) statisticians which are familiar with SAS and had no idea about NCA when they start their career in the industry. Instead of pro­gramming macros themselves and/or trusting what’s on the net1,2 they tell their boss “It will take weeks to code/validate that. Let’s pay some bucks for a Kodak3… Sooner or later one will not only need NCA, but more sophisticated features like Nonparametric Superposition, basic modeling in order to optimize sampling schedules, etc. I would not start that from scratch in SAS.
On the other hand agencies published SAS-code for some BE-methods (FDA: RSABE for HVDs and NTIDs; EMA: ABEL). It is certainly easier to copy-paste code compared to setting up / validating it in PHX/WNL (took me some weeks). Up to now nobody succeeded to code this stuff in R. :crying:


  1. Matos-Pita AS, de Miguel Lillo B. Noncompartmental Pharmacokinetics and Bioequivalence Analysis. PharmaSUG (Pharmaceutical Industry SAS® Users Group), May 22–25, 2005, Phoenix, AZ, USA.
       “The performance and validity of the program was tested against WinNonlin®, one of the most commonly used programs for pharmacokinetic analysis in the Pharmaceutical Industry. The results of twenty bioequivalence clinical trials were evaluated using both WinNonlin and SAS. PROC COMPARE of SAS was used to test for differences. There was a 100% agreement in all 20 studies.”
  2. He J. SAS Programming to Calculate AUC in Pharmacokinetic Studies—Comparison of Four Methods in Concentration Data. PharmaSUG, June 1-4, 2008, Atlanta, GA, USA.
  3. “You Press the Button, We Do the Rest”

Dif-tor heh smusma 🖖🏼 Довге життя Україна! [image]
Helmut Schütz
[image]

The quality of responses received is directly proportional to the quality of the question asked. 🚮
Science Quotes
d_labes
★★★

Berlin, Germany,
2015-07-21 15:04
(3173 d 01:53 ago)

@ Helmut
Posting: # 15119
Views: 18,628
 

 Call for duty

Dear Helmut,

❝ On the other hand agencies published SAS-code for some BE-methods (FDA: RSABE for HVDs and NTIDs; EMA: ABEL). It is certainly easier to copy-paste code compared to setting up / validating it in PHX/WNL (took me some weeks). Up to now nobody succeeded to code this stuff in R. :crying:


If you mean the statistics of such methods: Which one do you like first? I could imagine some ambitious R-coders out there which could be able to do the job.
Except of course implementing the SAS Proc mixed code for replicate designs :PCchaos:.

BTW: The reason for using SAS for the stats is very, very easy: "SAS is validated."

Regards,

Detlew
Helmut
★★★
avatar
Homepage
Vienna, Austria,
2015-07-21 15:24
(3173 d 01:33 ago)

@ d_labes
Posting: # 15121
Views: 18,646
 

 Call for duty

Dear Detlew,

❝ ❝ Up to now nobody succeeded to code this stuff in R. :crying:


❝ I could imagine some ambitious R-coders out there which could be able to do the job.


Sure. Not complicated at all.

❝ Except of course implementing the SAS Proc mixed code for replicate designs :PCchaos:.


This is exactly what I had in mind.

❝ BTW: The reason for using SAS for the stats is very, very easy: "SAS is validated."


True, of course. :-D

Dif-tor heh smusma 🖖🏼 Довге життя Україна! [image]
Helmut Schütz
[image]

The quality of responses received is directly proportional to the quality of the question asked. 🚮
Science Quotes
luvblooms
★★  

India,
2015-07-21 11:48
(3173 d 05:09 ago)

@ Helmut
Posting: # 15116
Views: 18,787
 

 Other Options

Hi HS,

❝ ❝ […] And also, don't know how they'll account for the studies conducted, …

❝ Oh yes, I performed 100, but state 5. Will Certara hire a private dick sneaking into my office and check? C’mon! Or is there a hidden counter in the software which phones home every time I execute a workflow?


Yes. Couple of days back, I had a word with a "scientist" who is with Certara and he informed me about the proposed fee model (the one that we got in mail) and also that there will be a counter/Log sort of ting in the tool which will record the total run studies.

❝ Yep. I regularly recalculate stuff posted here (~10/year). Apart from Pharsight’s employees I’m the top-poster in Certara’s Phoenix-Forum. Similar over there. Shall I pay for sumfink I do in my spare time at no costs? Gimme a break!


Time to increase your consultancy fee accordingly :-D

❝ Why not? Many companies use PHX/WNL for NCA only (and run the stats in SAS). Package bear is an alternative for NCA in R.


There is one more tool (actually it is a set of software under name Rx Express) called as PK express. Have checked it a couple of time when it was in beta phase and flow looks quite easy. Though there were some flaws in it initially but company is working hard to improve it. Also heard that FDA had evaluated their software and have said ok with some minor suggestions.
AFAIK, their proposed fee structure is very lucrative.

Wonder if anyone else on the forum evaluated that. If yes do share the thoughts on the quality.

❝ Have your heard about RapidNCA? Seems that only the bloody linear trapezoidal method is implemented.


Hearing this for first time. Let me check it out.

~A happy Soul~
Helmut
★★★
avatar
Homepage
Vienna, Austria,
2015-07-21 15:03
(3173 d 01:54 ago)

@ luvblooms
Posting: # 15118
Views: 18,691
 

 Other Options

Hi Luv,

❝ ❝ Or is there a hidden counter in the software which phones home every time I execute a workflow?


❝ Yes. Couple of days back, I had a word with a "scientist" who is with Certara and he informed me about the proposed fee model (the one that we got in mail) and also that there will be a counter/Log sort of ting in the tool which will record the total run studies.


That’s bizarre. I regularly re-evaluate already completed projects (exploring new PK metrics, impact of other sampling schedules, Nonparametric Superposition, …) in order to design new studies. Will that increase the counter? Oh dear!
Furthermore, how will Certara access the counter? I have both a hardware firewall on my router and a software firewall on my machine. Trojan, backdoor, rootkit in the “Pharsight Licensing Wizard”? Any­body out there already installed the new license? What is stated in the EULA we all click away that fast?

❝ ❝ Shall I pay for sumfink I do in my spare time at no costs? Gimme a break!


❝ Time to increase your consultancy fee accordingly :-D


I try hard. “Clever” clients try to issue contracts where my fee is fixed for some years…

❝ There is one more tool […] PK express.


Looks promising. I just requested a trial-version. At the bottom of the webpage I found

 [image] Validation Kit 

Let’s see what that means (e.g., surviving our reference data sets)…

❝ AFAIK, their proposed fee structure is very lucrative.


Couldn’t find anything on the website. Can you give me an idea?

❝ ❝ Have your heard about RapidNCA?


❝ Hearing this for first time. Let me check it out.


Brand-new. Version 1 was released on March 1, 2015.


PS: Seems to be a “popular” topic. ~15 views / hour so far.

Dif-tor heh smusma 🖖🏼 Довге життя Україна! [image]
Helmut Schütz
[image]

The quality of responses received is directly proportional to the quality of the question asked. 🚮
Science Quotes
Lucas
★    

Brazil,
2015-07-21 15:31
(3173 d 01:26 ago)

@ Helmut
Posting: # 15122
Views: 18,779
 

 Considering to jump ship

Hi Helmut!

❝ I guess small CROs – which will not jump ship – will opt for the lowest fee (still an amazing cost increase of 75% from the 2014 fee of 2,254 USD). Then what?

[…] you will be asked to submit a certification document to reconcile the # of actual studies completed to those you purchased.

A “certification document”? Oh yes, I performed 100, but state 5. Will Certara hire a private dick sneaking into my office and check? C’mon! Or is there a hidden counter in the software which phones home every time I execute a workflow?


I asked about this, they said that the studies we conduct are confidential and they'll not be able to control that. So they would not know how many studies we did in fact. This log that Luvblooms was talking seems bizarre, since the number of times I use Phoenix is not the same number of studies I conduct, as established above... And we haven't even talked about the studies with two or three analytes, that sometimes we have to run separates workflows (different concentration units, sampling schedule, etc.).

❝ ❝ May be time to improve my R "almost non-existing" skills! :-D


❝ Why not? Many companies use PHX/WNL for NCA only (and run the stats in SAS). Package bear is an alternative for NCA in R.


Time goes by in the rush of our daily routines, and unfortunately little time is left for developing new skills... But I gotta find some room for that in my day, maybe I could squeeze that in 00-06am. :-D
I use R now and then, but still a bit painful to use it comparing with Phoenix.

❝ Have your heard about RapidNCA? Seems that only the bloody linear trapezoidal method is implemented.


Well, this is the first time I'm hearing about this also... Really seems nice, have you checked it out personally? I couldn't find a demo version to download.

Regards

Lucas
Helmut
★★★
avatar
Homepage
Vienna, Austria,
2015-07-21 15:40
(3173 d 01:18 ago)

@ Lucas
Posting: # 15123
Views: 18,642
 

 Considering to jump ship

Hi Lucas,

❝ This log that Luvblooms was talking seems bizarre, since the number of times I use Phoenix is not the same number of studies I conduct, as established above... And we haven't even talked about the studies with two or three analytes, that sometimes we have to run separates workflows (different concentration units, sampling schedule, etc.).


X-actly. Crazy.

❝ Time goes by in the rush of our daily routines, and unfortunately little time is left for developing new skills... But I gotta find some room for that in my day, maybe I could squeeze that in 00-06am. :-D


Sleep deprivation might give you an interesting state of mind… :lookaround:

❝ ❝ Have your heard about RapidNCA?


❝ Well, this is the first time I'm hearing about this also...


Only useful if one has another package to perform the stats.

❝ Really seems nice, have you checked it out personally? I couldn't find a demo version to download.


1. No and 2. Me not either.

Dif-tor heh smusma 🖖🏼 Довге життя Україна! [image]
Helmut Schütz
[image]

The quality of responses received is directly proportional to the quality of the question asked. 🚮
Science Quotes
jag009
★★★

NJ,
2015-07-22 18:16
(3171 d 22:41 ago)

@ Helmut
Posting: # 15126
Views: 18,629
 

 Certara (Pharsight): Phoenix/WinNonlin licensing policy

Hi Helmut!

❝ Certara (the company behind Pharsight) changed their pricing policy for CROs (only!). Additionally to the annual single-user license fee (2,444 USD) CROs will be charged a fee based on the number of studies/projects performed per year… Part of the business-lingo I received:...


That's a big load of BS. How much money do they make from this software anyway??? Maybe later on they will make the program as a web based program, who knows!

Not to the mention the software still runs like a pig. On my 8gb i5 it's not running any faster than on my old P4 w 4gb ram. Opening and closing a project take a forever (IVIVC especially).

The only reason I use it is because I am too lazy to write SAS codes for determining half-life. Heck they don't even have an option to run partial+full replicate BE analysis yet (I mean within the program, not using your worksheet).

John
Helmut
★★★
avatar
Homepage
Vienna, Austria,
2015-07-22 19:51
(3171 d 21:06 ago)

@ jag009
Posting: # 15129
Views: 18,691
 

 SAS-makros?

Hi John,

❝ The only reason I use it is because I am too lazy to write SAS codes for determining half-life.


See Ref #1 in the post above. Of course, no option to select data points during visual inspection of fits. The 100% agreement with WinNonlin the authors proudly reported clearly means only the “automatic” method based on the maximum R²adj.

❝ Heck they don't even have an option to run partial+full replicate BE analysis yet (I mean within the program, not using your worksheet).


Yep. If I wouldn’t told them I guess they wouldn’t know until today that RSABE exists and what it means. ;-) Took me ~10 years to convince them that the reported power is for the dead and buried 80/20 rule. Since v6.4 we have – additionally – power for the TOST. Wow!
BTW, if using SAS one should know what to do as well. Today I got a response from the large Indian CRO (back-story: useless post-hoc power 29% but reported by SAS with 100%). The answer of the CRO’s “statistician” was:

Further, we have checked the analysis part pertaining to power calculation once again and also veri­fied with the results from WinNonlin software Version 5.3. And, it proves that the results pro­vided earlier are ok.

2×2=5… How I love the word “verify”. Rubbish in, rubbish out. :-D

Convincing the “Far Side” to implement a workaround (!) for unequal variances in parallel designs was even tougher. Took me merely 13 years.

Dif-tor heh smusma 🖖🏼 Довге життя Україна! [image]
Helmut Schütz
[image]

The quality of responses received is directly proportional to the quality of the question asked. 🚮
Science Quotes
UA Flag
Activity
 Admin contact
22,957 posts in 4,819 threads, 1,638 registered users;
69 visitors (0 registered, 69 guests [including 9 identified bots]).
Forum time: 15:58 CET (Europe/Vienna)

Nothing shows a lack of mathematical education more
than an overly precise calculation.    Carl Friedrich Gauß

The Bioequivalence and Bioavailability Forum is hosted by
BEBAC Ing. Helmut Schütz
HTML5