d_labes
★★★

Berlin, Germany,
2008-11-21 13:03
(5800 d 19:26 ago)

Posting: # 2765
Views: 9,865
 

 3-treatment-3-period designs [Design Issues]

Dear All,

from occasion of a recent post I would like to discuss again the question of designs for 3-treatments 3-periods.
Here in the forum it was repeatedly advocated that only a design with the full six sequences (let's letter!)
  ABC  Latin square 1
  BCA
  CAB
  -----
  ACB  Latin square 2
  BAC
  CBA

should be used.

Of course this design is balanced with respect to the conditions
  1. Each formulation occurs only once with each subject.
  2. Each formulation occurs the same number of times in each period.
  3. The number of subjects who receive formulation i in some period followed by formulation j in the next period is the same for all i # j
as Helmut has stated here.

But I cannot understand why this is desirable, especially the last condition.

As far as I have understood until now: The balance under 3. is only necessary if one is willing to adopt the view of first-order (simple) carry-over and incorporate this effect also in the analysis model. Only in that case the full 6-sequence design has its merits.

Following the view of SENN1) that the model of simple carry-over is obsolet and not testing for it, especially in case of BE studies with appropriate wash-out (which view has found its way into the new EMEA DRAFT, one of the rare things I appreciate), what are then the benefits of using balance with respect to condition 1.-3., the so called Williams designs?

And why should regulators insist on it?

Has anybody factual experiences that studies were rejected if only balanced designs with respect to the conditions 1. and 2. (Latin square designs) were use?

BTW which guidance state the necessity of the full 3x3x6 (treatment/period/sequence) design?

1) S. Senn, Cross-over Trials in Clinical Research
John Wiley & Sons, Chichester, chapter 10

Regards,

Detlew
Helmut
★★★
avatar
Homepage
Vienna, Austria,
2008-11-21 16:19
(5800 d 16:10 ago)

@ d_labes
Posting: # 2767
Views: 9,137
 

 Efficiency of higher-order designs

Dear DLabes!

❝ As far as I have understood until now: The balance under 3. is only necessary if one is willing to adopt the view of first-order (simple) carry-over and incorporate this effect also in the analysis model. Only in that case the full 6-sequence design has its merits.


Not only that. To quote Senn1:

Perfect balance with respect to sequences, is not, however, an absolute requirement for an analysis which allows for period effects as well as treatments: it is simply that other things being equal such designs are more efficient.

(my emphasis)

Jones and Kenward2 give a comparison of different four-treatment designs (no carry-over in the model).
The full set of latin squares (sequences = treatments! = 1×2×3×4 = 24):
 1 A B C D    13 A D C B
 2 B C D A    14 D C B A
 3 C D A B    15 C B A D
 4 D A B C    16 B A D C
 5 A C B D    17 A C D B
 6 C B D A    18 C D B A
 7 B D A C    19 D B A C
 8 D A C B    20 B A C D
 9 A B D C    21 A D B C
10 B D C A    22 D B C A
11 D C A B    23 B C A D
12 C A B D    24 C A D B   Efficiency 100%


Orthogonal latin squares (balanced):
 1 A B C D     7 C B D A
 2 B A D C     8 D A C B
 3 C D A B     9 A C D B
 4 D C B A    10 B D C A
 5 A D C B    11 C A B D
 6 B C A D    12 D B A C   Efficiency 90.91%


Williams' design (balanced):
 1 A D B C
 2 B A C D
 3 C B D A
 4 D C A B   Efficiency 90.91%


One latin square:
 1 A B C D
 2 B C D A
 3 C D A B
 4 D A B C   Efficiency 18.18%


The Williams’ design shows the same efficiency as the orthogonal set (90.91%) - with only 4 as compared to 12 sequences. The simple latin square clearly is inferior to all others.

For the 6×3 Williams’ design they give an efficiency of 80% (unfortunately no value for a 3×3). There's also the routine3 XOEFFICIENCY for GenStat.

❝ Following the view of SENN1) that the model of simple carry-over is obsolet and not testing for it, especially in case of BE studies with appropriate wash-out (which view has found its way into the new EMEA DRAFT, one of the rare things I appreciate),…


True!

❝ … what are then the benefits of using balance with respect to condition 1.-3., the so called Williams designs?


See the quote of Senn above, and also this one:

It has been my habit in designing cross-over trials to use all six sequences for a three-period three-treatment design and to use a single Latin square for a four-treatment design.


❝ And why should regulators insist on it?


Do they?

❝ Has anybody factual experiences that studies were rejected if only balanced designs with respect to the conditions 1. and 2. (Latin square designs) were use?


No, but my experience is limited to 3×3 studies I’ve seen as a consultant (in my studies I never used this design). Any Williams’ design has the advantage that pairwise comparisons may be extracted (also recommended by Byron Jones in a personal communication at the BioInternational 2003) - which are also balanced (needed for nonparametric comparisons, which seems to be of historical interest in the EU…). I’ve seen small studies where due to dropouts the extracted 2×2 sets were extremely imbalanced - or even worse, didn't work any more at all.

❝ BTW which guidance state the necessity of the full 3x3x6 (treatment/period/sequence) design?


I didn’t check all national guidelines; the first one coming into my mind was ANVISA’s (2003).
  1. S Senn
    Cross-over Trials in Clinical Research
    John Wiley & Sons, Chichester, pp162-163 (2nd ed. 2002)
  2. B Jones and MG Kenward
    Design and Analysis of Cross-over Trials
    Chapman & Hall/CRC, Boca Raton, Chapter 4 (2nd ed. 2003)
    Parts of this chapter as a preview in GoogleBooks.
  3. B Jones and PW Lane
    XOEFFICIENCY
    In RW Payne (ed.) GenStat - Release 6.1. New Features
    VSN International, Oxford (2002)

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,
2008-11-24 13:49
(5797 d 18:41 ago)

@ Helmut
Posting: # 2785
Views: 8,494
 

 Efficiency of higher-order designs

Dear Helmut!

Sorry, but I cannot follow your reasoning, mostly.

First: Senns quote's

❝ [...] To quote Senn1:

Perfect balance with respect to sequences, is not, however, an absolute requirement for an analysis which allows for period effects as well as treatments: it is simply that other things being equal such designs are more efficient.

(my emphasis)


I must confess that this paragraph is a little bit confusing to me (especially his example with 14 subjects in a 3x3 cross-over). But I think he is not talking about balance in the sense of the 3. condition of Williams designs (every treatment follows each other the same number) but rather in terms of balance of number of subjects to sequences (which ever chosen).

A hint that this is so cames from your second quote of Senn.
But in full context this reads:
" Frequently there is no good reason to choose any one Latin square in preference to another and the choice may then be made at random. There is often no reason either why a single Latin square should be used in a given trial (although some analyses are simpler if this is done) rather than a number of squares, apart from convenience in preparing trial material.

❝ It has been my habit in designing cross-over trials to use all six sequences for a three-period three-treatment design and to use a single Latin square for a four-treatment design.


Senn's attitude in his book is always to depict with personal sentences if any thing he is discussing is a personal view/opinion or doesn't matter.

Second: Regarding your reasoning in terms of efficiencies from Jones and Kenward

If I read my Holy bible[2] of cross-over designs again and again I cannot follow you.
Again I must confess that it is like some white or black magic to me, what they do in chapter 4.
But undoubtedly: The efficiency numbers they have given for various designs (which you cite) are not obtained with no carry-over.
"In the presence of carry-over effects we define efficiency exactly as above but with Var[taui-tauj] now denoting the difference between two treatments adjusted for carry-over effects: We label this efficiency as Ed ... For information we will always include the values of the efficiencies of a design, in the order of Ed, ..." (emphasis by me). Your cited numbers are the Ed efficiencies.

BTW: I think their 100% base is different from yours.

Thus the efficiencies show that in a model with carry-over not all designs are equal as they were without such effects. But you have always to pay for inclusion of such effects (efficiencies <100%).
That is why part of Senn's criticism of the simple carry-over model is [2]Chapter 10.3.4 The estimators based on it are inefficient and Chapter 10.3.5 The designs associated with it are not necessarily better than others.

Third:

❝ ❝ ... what are then the benefits of using balance with respect to condition 1.-3., the so called Williams designs?

❝ ❝ And why should regulators insist on it?

Do they?


This question of mine is directly emerging from the statement "The chance of regulatory acceptance of a 3×3 is close to zero." in this post by HS :-D.

Fourth:

❝ [...] Any Williams' design has the advantage that pairwise comparisons may be extracted …

❝ I've seen small studies where due to dropouts the extracted 2x2 sets were extremely imbalanced - or even worse, didn't work any more at all.

Here I can follow you, but it applies only for very imbalanced, small studies, in which we end up with many drop outs or missings.

Therefore it is my habit to choose all six sequences for a three-period three-treatment design also and to use a single Latin square for a 4-treatment design, not necessarily Williams ;-).
For sequence stratified non-parametric evaluation Senn[1] (p162-163) mentioned among others the 4-period-4-treatment Latin square
  ABCD
  BADC
  CDAB
  DCBA

which is not an Williams’s design. This and similar designs have the properties that for any pair of treatments in a given sequence there is another sequence in which the treatments are reversed in periods. This allows the evaluation stratified to these sequence pairs as described by Duchateau[3] in the context of 3-period-3-treatment studies.
This evaluation is not possible with Williams design in a 4-period-4-treatment study or with a single 3x3 Latin square in a 3-period-3-treatment study.

But non-parametrics …:angry:. You know!

[1] S Senn
    Cross-over Trials in Clinical Research
    John Wiley & Sons, Chichester (2nd ed. 2002)
[2] B Jones and MG Kenward
    Design and Analysis of Cross-over Trials
    Chapman & Hall/CRC, Boca Raton, Chapter 4 (2nd ed. 2003)
[3] Duchatau et. al
    Adjusting pairwise nonparametric equivalence hypothesis
    tests and confidence intervals for period effects
    in 3x3 crossover trials
    J. Biopharm. Stat. Vol 12(2), 149-160, 2002

Regards,

Detlew
UA Flag
Activity
 Admin contact
23,249 posts in 4,885 threads, 1,653 registered users;
75 visitors (0 registered, 75 guests [including 9 identified bots]).
Forum time: 09:30 CEST (Europe/Vienna)

I have never in my life learned anything
from any man who agreed with me.    Dudley Field Malone

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