OT: keep TTT! [PK / PD]

posted by yjlee168 Homepage – Kaohsiung, Taiwan, 2013-07-23 01:42 (4350 d 12:09 ago) – Posting: # 11035
Views: 22,090

Dear Helmut,

❝ [...]

❝ Please no! Contrary to R²adj TTT is based on PK-grounds. I would not call it model-dependent. OK, it is based on a one-compartment model but the estimation of λz is within the boundaries of NCA, IMHO. For multiple distribution / elimination phases it is the responsibility of the user to select the last loglinear phase visually.


OK. I plot your simulated data as semi-log scale (log(C) vs. time) to see if the 'breakpoint' (which is occurring at 9) is easy to figure out visually. It is not. I don't know if you put too much noise or not. If we can visually figure out the 'breakpoint' easily, then we probably can change current TTT algo as a semi-automatic way. That is to pick the last 3-4 data points first (automatically) and then to show plots to allow user to add more until reaching 'breakpoint' (manually). I guess it will work only if user knows which is the 'breakpoint.' One more question: what if the 'breakpoint' is just not one of scheduled sampling times?

❝ Anyhow, I’m not overly optimistic yet – especially if at least one of the following conditions applies: Noisy data. [...]



Agree.

All the best,
-- Yung-jin Lee
bear v2.9.2:- created by Hsin-ya Lee & Yung-jin Lee
Kaohsiung, Taiwan https://www.pkpd168.com/bear
Download link (updated) -> here

Complete thread:

UA Flag
Activity
 Admin contact
23,424 posts in 4,927 threads, 1,677 registered users;
39 visitors (0 registered, 39 guests [including 4 identified bots]).
Forum time: 13:52 CEST (Europe/Vienna)

Complex, statistically improbable things are by their nature
more difficult to explain than
simple, statistically probable things.    Richard Dawkins

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