2-stage design with interim sample size estimation [Two-Stage / GS Designs]

posted by d_labes  – Berlin, Germany, 2010-10-05 17:09 (4895 d 06:40 ago) – Posting: # 5997
Views: 14,737

Dear all,

I have been questioned about the following design:

The sponsor/experimenter has no sufficient information about CV and point estimate ('true' value) to be able to make an "educated" guess of the sample size for a BE study. Some distinct shift in the point estimator is expected.

To save resources (its a fixed combination drug with 3 constituents, analytical expensive) he is not willing to do a pilot study and based on this a pivotal with BE decision not using the data from the pilot.

Thus the design shall be a 2-stage design. A first group of subjects (conception is 8-12) shall be dosed, analyzed and statistical evaluated. Based on the data (CV and point estimate) a final sample size shall be derived which will be utilized for the second stage.

There is the idea of a formal extreme small alpha to spend after the first stage =0.001 (to conform EMA guideline). This will ensure with great probability that the second stage will be reached. Practical no stopping will occur by that.

The only stopping rule is a maximum sample size constraint Nmax the sponsor is willing to fund. If the estimated sample size after stage 1 is greater the second stage will not performed.

Any body out there who knows how to handle such a design?
Is there something to do to protect the overall alpha (lower nominal alpha at second stage)?
Is the sample size estimation done as usually?

Any input will be highly appreciated.



Complete thread:

UA Flag
 Admin contact
22,916 posts in 4,807 threads, 1,645 registered users;
30 visitors (0 registered, 30 guests [including 9 identified bots]).
Forum time: 22:50 CET (Europe/Vienna)

Nothing shocks me. I’m a scientist.    Harrison Ford (as Indiana Jones)

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