# Which analysis to use?

Hi

Last edited:

#### GretaGarbo

##### Human
Well, did the drug C have an effect on A?

#### Karabiner

##### TS Contributor
How large were the groups? Was allocation of patients to groups at random? How were doses of A, B, and C determined? Wereo dose levels for A and for B similar between groups? Is the relationship between B dose and its effect on concentration of A linear?

With kind regards

Karabiner

#### Karabiner

##### TS Contributor
What about a regression model with concentration of A as the dependent variable,
the adjusted doses of A and B as interval scaled predictors, and C as dummy-coded
predictor? Also the interactions should be included.

You could perform separate analyses for the different time points.If you want one
model which Includes all time points, then you will probably need multilevel
modeling.

With kind regards

Karabiner

#### GretaGarbo

##### Human
Wow, so this is real data! And on organ transplantations!

(I guessed that this was a student homework with hypothetical data.)

So there seems to be a time effect. A patient is measured 2 or 3 times. Maybe a standard statistical model like a repeated measurement model can be used.

Some are treated with drug C and some are not.
But there is a difficulty in that:
Doses of A was initially a standard dose, then adjusted by blood concentration levels.
B is initially a standard dose dependent on body weight, then adjusted by blood concentration levels.
You want to check the effect of C on A, but A has been adjusted. Does that mean that the effect of on A is how much A has been adjusted?

#### Karabiner

##### TS Contributor
Wow, so this is real data! And on organ transplantations!

(I guessed that this was a student homework with hypothetical data.)
Yes, strange as it seems.

Maybe the correct to the such a question
would be that the op should contact the ethics
committee for the study, since letting amateurs
analyse real patient data might violate the
ethical approval of the study.

Regards

Karabiner