Subject replacement [Design Issues]

posted by Ohlbe – France, 2020-01-08 11:41 (1568 d 20:01 ago) – Posting: # 21068
Views: 2,889

Dear Pharma_88,

❝ ❝ In the design I have in mind, the group/cohort is completed with 3+3 patients, not just 3. And you enrol new patients in each cohort (if you only use the same patients who have a good tolerability, you have some bias).


❝ Yes. but initial cohort (means first one) is started with 3 patients only. subsequent cohort will have 3+3 patients. Right?


No. The initial cohort is started with 3 patients. If you have 0 or 1 DLT, you enrol 3 more in the same cohort. If you meet the pre-defined criteria, the DMC will allow you to move to the next cohort. There you go through the same process: start with 3, if you have 0 or 1 DLT you add 3 more, and so on until you reach the highest tolerated dose or the highest dose defined in your protocol. All cohorts are treated equally, including the first one.

❝ ❝ ❝ Further, in next cohort suppose 1 patient is withdrawn or have some AE then its compulsory to add 3 more patients to inline with multiplication of 3?

❝ ❝

❝ ❝ If a patient is withdrawn due to a DLT, see my first paragraph. The patient is not replaced. If he is withdrawn for another reason: you really have to be extra-sure it is really totally unrelated to a DLT. You may decide to replace that patient (meaning, only 1 extra-patient). Make sure this is properly defined in your protocol. I would not use 3 patients to replace just 1.


❝ So, in this case, total 4 patients are required to enroll (1 for replacement and 3 others). Correct?


Two possible situations:

Regards
Ohlbe

Complete thread:

UA Flag
Activity
 Admin contact
22,993 posts in 4,828 threads, 1,656 registered users;
108 visitors (0 registered, 108 guests [including 5 identified bots]).
Forum time: 08:42 CEST (Europe/Vienna)

Never never never never use Excel.
Not even for calculation of arithmetic means.    Martin Wolfsegger

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