Bioequivalence and Bioavailability Forum • Partial replicate designs

The Outlaw Torn
★

Europe,
2013-09-25 10:21
(4249 d 07:14 ago)

Posting: # 11556
Views: 6,202

Partial replicate designs [RSABE / ABEL]

Good morning Europe (and hello to the rest of you),

I'd like to get you folks' take on something and some informtion on the other.

Thing one (not-so-subtle Dr. Seuss reference! and yes, you guessed it, I have young children). :sleeping:

Okay, what is the advantage of conducting a full replicate BE study on an suspected HVDP if the ability to widen CIs is based solely on the reference product variability (basically, it's cheaper to conduct a partial with RTR/RRT, right?)?

The second thing is the following. What is the sample size estimate for a drug (not Schützomycin) with a intra-subject variability of 37.5%? And how would those subjects be distributed per group in said partial replicate?

Just in case ya'll are wondering. I'm actually not asking ya'll to do my work for me, I'm trying to wrap my head around a concept I don't seem to be able to wrap my head around when I should.

Thank you,
Outlaw Torn

Helmut
★★★

Vienna, Austria,
2013-09-25 11:12
(4249 d 06:23 ago)

@ The Outlaw Torn
Posting: # 11557
Views: 5,403

Partial replicate designs

Post reply

Good morning from Amman,

❝ […] what is the advantage of conducting a full replicate BE study on an suspected HVDP if the ability to widen CIs is based solely on the reference product variability

You learn something about your formulation.

❝ (basically, it's cheaper to conduct a partial with RTR/RRT, right?)?

First, the partial replicate is TRR|RTR|RRT – which is a lousy design. I would opt for a full replicate with three periods (RTR|TRT). Similar sample size/costs, more information.

❝ The second thing is the following. What is the sample size estimate for a drug (not Schützomycin) with a intra-subject variability of 37.5%? And how would those subjects be distributed per group in said partial replicate?

For EMA’s widening, 80% power, and a ratio of 90% (I would not suggest to use 95% for HVDs/HVDPs)

library(PowerTOST)

sampleN.scABEL(targetpower=0.8, theta0=0.9, CV=0.375, design="2x3x3")

+++++++++++ scaled (widened) ABEL +++++++++++

            Sample size estimation

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

Study design:  2x3x3 (partial replicate) 

log-transformed data (multiplicative model)

1e+05 studies simulated.



alpha  = 0.05, target power = 0.8

CVw(T) = 0.375; CVw(R) = 0.375

Null (true) ratio = 0.9

ABE limits / PE constraints = 0.8 ... 1.25 

Regulatory settings: EMA 

- CVswitch =  0.3, cap on ABEL if CVw(R) > 0.5

- Regulatory constant = 0.76 



Sample size search

 n     power

42   0.7793 

45   0.8037



sampleN.scABEL(targetpower=0.8, theta0=0.9, CV=0.375, design="2x2x3")

+++++++++++ scaled (widened) ABEL +++++++++++

            Sample size estimation

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

Study design:  2x2x3 (TRT|RTR) 

log-transformed data (multiplicative model)

1e+05 studies simulated.



alpha  = 0.05, target power = 0.8

CVw(T) = 0.375; CVw(R) = 0.375

Null (true) ratio = 0.9

ABE limits / PE constraints = 0.8 ... 1.25 

Regulatory settings: EMA 

- CVswitch =  0.3, cap on ABEL if CVw(R) > 0.5

- Regulatory constant = 0.76 



Sample size search

 n     power

42   0.7612 

44   0.7777 

46   0.7919 

48   0.8049

—
Dif-tor heh smusma 🖖🏼 Довге життя Україна!
Helmut Schütz

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

The Outlaw Torn
★

Europe,
2013-09-25 13:10
(4249 d 04:26 ago)

@ Helmut
Posting: # 11558
Views: 5,338

Partial replicate designs

Post reply

Hello again,

❝ First, the partial replicate is TRR|RTR|RRT – which is a lousy design. I would opt for a full replicate with three periods (RTR|TRT). Similar sample size/costs, more information.

Don't mean to interupt your "vacation" in beautiful Jordan, but if you have "nothing better to do--it's not like they have good falafels or kebabs there" :hungry:

mind explaining why it's a lousy design?

❝ ❝ The second thing is the following. What is the sample size estimate for a drug (not Schützomycin) with a intra-subject variability of 37.5%? And how would those subjects be distributed per group in said partial replicate?

So, only 15 or 16 subjects per groups, eh? Hmmmm. Would have expected more. I'll have to do a bit more head wrapping and may have a follow-up.

Thanks for everything. And enjoy Amman.

Take care,
Outlaw Torn

Helmut
★★★

Vienna, Austria,
2013-09-25 18:35
(4248 d 23:00 ago)

@ The Outlaw Torn
Posting: # 11559
Views: 5,332

Partial replicate designs

Post reply

Hi Outlaw,

❝ […] it's not like they have good falafels or kebabs there"

Well, I’m already in famous for Μουσακάς and Τζατζίκι. :-D

❝ mind explaining why it's a lousy design?

Little bit short in time. In a partial replicate the test is not repeated, so the model is overspecified (speaking in mixed effects terms). Only with EMA’s crippled model you would have no problems (assuming identical variances of T and R anyhow). There are a lot of posts already.

❝ So, only 15 or 16 subjects per groups, eh?

If you insist in the partial replicate, yes.

❝ And enjoy Amman.

Already enjoying Athens.

—
Dif-tor heh smusma 🖖🏼 Довге життя Україна!
Helmut Schütz

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

The Outlaw Torn ★ Europe, 2013-09-26 10:34 (4248 d 07:01 ago) @ Helmut Posting: # 11564 Views: 5,302	Partial replicate designs Post reply
	Hello Helmut, I hear the lamb ribs are worth a try and grilled octopus is a great escort to ouzo. Thanks for the info (it's best to stay away from patials, I believe) and enjoy your world tour (are you selling t-shirts?). Outlaw.

Helmut
★★★

Vienna, Austria,
2013-09-26 10:56
(4248 d 06:39 ago)

@ The Outlaw Torn
Posting: # 11565
Views: 5,465

OT

Post reply

Hi Outlaw,

❝ I hear the lamb ribs are worth a try and grilled octopus is a great escort to ouzo.

I prefer the latter, though octopus comes also great boiled in olive oil together with onions and garlic. Ouzo is out of discussion. :-D

❝ […] are you selling t-shirts?

How did you know? Do you prefer pay-pal? Also consider the limited editions »I love microdosing« and »I love BDDCS« with a personal dedication from Les Benet.

—
Dif-tor heh smusma 🖖🏼 Довге життя Україна!
Helmut Schütz

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