Bioequivalence and Bioavailability Forum

Dear Janardhan!

Please have a look at the policy first.

❝ Can any one give definitons of terms used in Randomisation (Ex:- Blocks, Sequence, etc.) with examples.

A sequence denotes the order a subject receives treatments in a cross-over design.
A block is the smallest unit where all possibles sequences are contained in.
In a 2×2 cross-over design, there are two possible sequences (S); therefore the smallest block is of size two (or in other words, the total sample size is always an even number):
    +-------+
    |Period |
+---+---+---+
| S | 1 | 2 |
+---+---+---+
| 1 | A | B |
| 2 | B | A |
+---+---+---+

In a 6×3 cross-over design (Williams' design for 3 treatments), there are six possible sequences; the smallest block is of size six (the sample size is a multiple of 6):
    +-----------+
    |   Period  |
+---+---+---+---+
| S | 1 | 2 | 3 |
+---+---+---+---+
| 1 | A | B | C |
| 2 | B | C | A |
| 3 | C | A | B |
| 4 | A | C | B |
| 5 | B | A | C |
| 6 | C | B | A |
+---+---+---+---+

In a 4×4 cross-over design (Williams' design for 4 treatments), there are four possible sequences; the smallest block is of size four (the sample size is a multiple of 4):
    +---------------+
    |     Period    |
+---+---+---+---+---+
| S | 1 | 2 | 3 | 4 |
+---+---+---+---+---+
| 1 | A | D | B | C |
| 2 | B | A | C | D |
| 3 | C | B | D | A |
| 4 | D | C | A | B |
+---+---+---+---+---+

In a parallel design there are no sequences (each subject receives only one of the possible n treatments). The sample size is a multiple of n.

The block size does not play an important role in BE studies, since they are

• single center, and
  
• open labeled.

Only if multiple groups are employed (e.g., due to logistic reasons or limited capacity of the clinical site), groups should be allocated (statistical term: stratified) in such a way, that each group contains complete blocks.
Example:
6×3 design, sample size 30, clinical capacity 18 beds.
1^st group 18 subjects (3 blocks of 6 sequences), 2^nd group 12 subjects (2 blocks of 2 sequences).

❝ Also suggest some sites which give clear definitions in understanding of cross over design, parallel design etc. with examples.

For the three basic designs see one of my lectures (slide 12pp for the 2×2 cross-over, slide 22pp for the parallel design, and slides 25pp for a higher-order cross-over design).

If you are interested not only in the total intra-subject variability (CV_intra), but also in the variability due to treatments, a replicate design must be applied.
The two most commonly used ones are:
For reference variability the 3×2 design (3 period, 2 sequence):
    +-----------+
    |   Period  |
+---+---+---+---+
| S | 1 | 2 | 3 |
+---+---+---+---+
| 1 | T | R | T |
| 2 | R | T | R |
+---+---+---+---+
For test/reference variability the 4×2 design (4 period, 2 sequence):
    +---------------+
    |     Period    |
+---+---+---+---+---+
| S | 1 | 2 | 3 | 4 |
+---+---+---+---+---+
| 1 | T | R | T | R |
| 2 | R | T | R | T |
+---+---+---+---+---+