Bioequivalence and Bioavailability Forum • announce the release of bear v2.3.0

yjlee168
★★★

Kaohsiung, Taiwan,
2009-03-09 22:32
(5893 d 17:23 ago)

Posting: # 3348
Views: 8,802

announce the release of bear v2.3.0 [🇷 for BE/BA]

Dear all colleagues,

We like to announce the release of bear v2.3.0 and share it with all colleagues of this Forum. The new release has the following updates and new functions:

add sample estimation and lme (mixed model) data analysis for the parallel BE study
add References into bear's output (such as sample estimation, outliers detection, etc.)
label outliers' subjects number for boxplot only if there is any (see crossover demo)
add input data summary of BA measurement (class level information, means, etc.)
add interpretations for some statistical tests (such as Hotelling T² test)
add add Two One-Sided Tests (TOST) and Anderson-Hauck's tests (for educational purposes ONLY)
allow users to change BE acceptance range now (not fixed on 80%-125% any more!); different countries have different regulatory guideline.
show MSResidual and MSSubject(seq) values when calculating inter- and intra-subject CVs
change Hotelling T² test layout

Please use "Update packages" function to upgrade your bear if you have already installed any previous version of bear. For new installation, please refer R website for "How to install R add-on packages" and also browse bear website. Previous version of bear announced in the Forum are listed as follows: bear v2.2.0, v2.1.0, v2.0.1, v1.1.4, v1.0.0.

Here are some results generated from bear v2.3.0.
<<Sample Size Estimation>>




 Upper acceptance limit = 125 %

 Lower acceptance limit = 80 %

     Expected ratio T/R = 95.00 %

           Target power = 80.00 %

Inter- or Intra-subj CV = 20.0 %



 study    2x2x1         2x2x2        2x2x3        2x2x4

design   parallel     crossover   replicated   replicated

------- ----------- ------------ ------------ ------------

   N       36            20           16           10



Estimated power = 80.99398 % (for parallel study)

Estimated power = 83.46802 % (for crossover/replicate study)

where 2x2x2 means 2-treatment, 2-sequence, and 2-period crossover study.

Data Summary of BA measurement


(...)             

--------------------------------------------------------------------------

   subj drug seq prd Cmax    AUC0t   AUC0INF   lnCmax   lnAUC0t lnAUC0INF

1     1    1   2   2 1739 14445.27 14925.656 7.461066  9.578123  9.610837

2     2    1   1   1 1481 12516.14 13134.553 7.300473  9.434774  9.483002

3     3    1   2   2 1780 15371.40 15999.744 7.484369  9.640264  9.680328

4     4    1   1   1 1374 11063.10 11647.240 7.225481  9.311371  9.362824

5     5    1   2   2 1555 13971.45 14556.963 7.349231  9.544771  9.585825

6     6    1   1   1 1756 15376.35 15950.591 7.470794  9.640586  9.677251

(...)





  Class Level Information                  

--------------------------------------------------------------------------

  Class       Levels      Values

  SUBJECT      14          1 2 3 4 5 6 7 8 9 10 11 12 13 14 

  DRUG          2          1 2 

  SEQUENCE      2          1 2 

  PERIOD        2          1 2 

-----------

DRUG 1: the Ref. product; DRUG 2: the Test product

SEQUENCE 1:  Ref. -> Test, and SEQUENCE 2: Test -> Ref.

-------------------------------------------------------



 Means                  

--------------------------------------------------------------------------

  SEQUENCE     Cmax    AUC0t  AUC0INF  lnCmax  lnAUC0t lnAUC0INF

1        1 1602.071 14109.25 14713.39 7.37045 9.538976  9.582570

2        2 1589.786 13357.85 13970.66 7.36052 9.487205  9.532541





   SUBJECT SEQUENCE   Cmax    AUC0t  AUC0INF   lnCmax  lnAUC0t lnAUC0INF

1        1        2 1686.0 13369.59 13950.10 7.429620 9.497491  9.540791

2        2        1 1659.0 13907.39 14645.04 7.408181 9.535147  9.586510

3        3        2 1926.5 15277.70 15838.53 7.560560 9.634131  9.670149

4        4        1 1501.5 12522.38 13133.29 7.310602 9.428436  9.476462

5        5        2 1470.0 12911.54 13523.30 7.291343 9.462496  9.509240

6        6        1 1639.0 14607.04 15146.74 7.399287 9.587870  9.624130

(...)



  PERIOD     Cmax    AUC0t  AUC0INF   lnCmax lnAUC0t lnAUC0INF

1      1 1632.714 13855.25 14531.50 7.390958 9.52988  9.577774

2      2 1559.143 13611.85 14152.54 7.340011 9.49630  9.537337





  DRUG     Cmax    AUC0t  AUC0INF  lnCmax  lnAUC0t lnAUC0INF

1    1 1539.643 13171.14 13781.86 7.32951 9.474286  9.519987

2    2 1652.214 14295.97 14902.19 7.40146 9.551895  9.595125

(...)    BE Summary Report                            

--------------------------------------------------------------------------

  Dependent Variable: lnCmax                                               

--------------------------------------------------------------------------

        n1(R=>T) = 7 

        n2(T=>R) = 7 

        N(n1+n2) = 14 

  Lower criteria = 80 %

  Upper criteria = 125 %

        MEAN-ref = 7.32951 

       MEAN-test = 7.40146 

             MSE = 0.01780339 

              SE = 0.05043155 

Diff. (test-ref) = 0.07195028 



**************** Classical (Shortest) 90% C.I. for lnCmax ****************



  CI90_lower Point_estimated CI90_upper

1     98.223         107.460    117.566



---------------------- Two One-Sided Tests (TOST) -------------------------



     TOST   T value   P value

1 T_lower   -5.8514    0.0000

2 T_upper   -2.9980    0.0056



**Interpretation:

Ho: Theta =<  0.8  or  Theta >=  1.25 

Ha:  0.8  < Theta <  1.25 

Theta = Mean_test/Mean_ref  

Because all P values are less than 0.05, we will reject the null hypothesis (Ho). 

BE acceptance range is set at 80 % -  125 % .



------------------------ Anderson-Hauck Test ------------------------------



          P value = 0.005515 



**Interpretation:

Ho: Theta =<  0.8  or  Theta >=  1.25 

Ha:  0.8  < Theta <  1.25 

Theta = Mean_test/Mean_ref  

Because all P values are less than 0.05, we will reject the null hypothesis (Ho). 

BE acceptance range is set at 80 % -  125 % .



---------------------------------------------------------------------------

Ref.:

1. Chow SC and Liu JP. Design and Analysis of Bioavailability-       

   Bioequivalence Studies. 3rd ed., Chapman & Hall/CRC, New York (2009).

2. Schuirmann DJ. On hypothesis testing to determine if the mean of a  

   normal distribution is continued in a known interval. Biometrics, 37, 

   617(1981).                                                           

3. Schuirmann DJ. A comparison of the two one-sided tests procedure and the 

   power approach for assessing the equivalence of average bioavailability.

   Journal of Pharmacokinetics and Biopharmaceutics, 15, 657-680 (1987). 

4. Anderson S and Hauck WW.  A new procedure for testing equivalence in 

   comparative bioavailability and other clinical trials. Communications 

   in Statistics-Theory and Methods, 12, 2663-2692 (1983).     

-------------------------------------------------------------------------- 

(...)

Please let us know if you have any question about bear. Thank you all.

All the best,
Hsin-ya Lee, Yung-jin Lee
College of Pharmacy,
Kaohsiung Medical University,
Kaohsiung, Taiwan

d_labes
★★★

Berlin, Germany,
2009-03-13 10:16
(5890 d 05:39 ago)

@ yjlee168
Posting: # 3359
Views: 7,207

Sample size of bears

Post reply

Dear Bears,

❝ Here are some results generated from bear v2.3.0.

 <<Sample Size Estimation>>
❝ 

❝  Upper acceptance limit = 125 %

❝  Lower acceptance limit = 80 %

❝      Expected ratio T/R = 95.00 %

❝            Target power = 80.00 %
❝ Inter- or Intra-subj CV = 20.0 %
❝ 

❝  study    2x2x1         2x2x2        2x2x3        2x2x4
❝ design   parallel     crossover   replicated   replicated
❝ ------- ----------- ------------ ------------ ------------

❝    N       36            20           16           10

From the methodological and educational point of view I would suggest not to mention parallel group design side by side with the cross-over designs.

These a so different pairs of shoes because of the different variabilities involved. The "naive" user may be mislead in choosing his right design.

To illustrate what I mean let's argue "naive":
The cost for one subject in a BE study is mainly determined by observing the concentration time course. For the parallel group design we have 36 concentration time course to observe. For the 2x2x2 crossover we have 20 in period one and 20 in period 2. So concerning our budget (too small for the study, as usual :-P

) we choose of course the parallel group design.

Of course this argument is not for the informed users who knows what he does. But ...
It needs only a short look into the sample size category of this forum ... :wink:

—
Regards,

Detlew

yjlee168
★★★

Kaohsiung, Taiwan,
2009-03-16 09:04
(5887 d 06:51 ago)

@ d_labes
Posting: # 3369
Views: 6,885

Sample size of bears

Post reply

Dear D. Labes,

Thank you for your insightful comments. Indeed, when we were coding this part (sample size estimation), Hsin-ya and I already had arguments about this. The argument was whether we should put all these side by side. However, we did not discuss the naive users to use bear at that moment. Hsin-ya was against to put all these together. The reason was they might confuse users. It doesn't make any sense to put all these different things together. However, I didn't think it might confuse users at that moment. Of course, I didn't take naive users into consideration since the user would be asked to enter intra-/inter-subject CVs before showing all estimated results on screen. Then the user must know what he/she is trying to do. To do so is that I just try to make bear as simple as possible. Therefore, we will separate the sample size estimation of the parallel study from the crossovers (and replicates) in the next version. Hsin-ya is very happy about this.

❝ the study, as usual :-P ) we choose of course the parallel group design.

I estimated the cost based on your proposed study too. I found the cost for sample analysis be less with the parallel (36's data sets) when it was compared to the 2x2x2 crossover (40's data sets). However, the parallel will need to recruit 18 more subjects than the crossover. In Taiwan, the parallel study may not cost less than the 2x2x2 crossover based on your proposed study. This can one of reasons misleading me to code bear in this way, I guess.

—
All the best,
-- Yung-jin Lee
bear v2.9.2:- created by Hsin-ya Lee & Yung-jin Lee
Kaohsiung, Taiwan https://www.pkpd168.com/bear
Download link (updated) -> here