How do I compare three treatments doing delta analyze?

[aska_ujita]

Hello there, hope you are great and healthy.

I have three treatments ()A, b and C) and tested variables (VAR1 , VAR2 and VAR3) with categorical answers (scores 1 to 5) in different days (day 1, 2 and 3).

They suggested to make an analyze with the delta. But seems like that is possible to do it just for two treatments.

I would like to confirm that, and if is possible with three treatments, how can I do that??

Thank you!

Best regards and take care, Aska.

[ChrisNZ]

Moved to stats forum.

[SteveDenham]

Probably the easiest way would be to calculate the delta as day2 - day1 and day3-day2 for each of the variables, and then do a multivariate ANOVA using PROC GLM.

A better way that helps to avoid regression to the mean would be to use generalized estimating equation approach (PROC GEE) or a conditional repeated measures approach (PROC GLIMMIX), where the response variables have a multinomial distribution with a cumulative logit link (either procedure).  Selection of a covariance structure in the two approaches will deal with the dependency/correlation in time rather than the delta method.

SteveDenham

[aska_ujita]

Hello, thank you for your hlp.

Is there a way to do this day2 -day1 and day3-day2 automatically?

I have a column called "DAY". Because there is a lot of variables to test, would be great if has a way to do it faster.

Thank you!!

Kind regards, Aska.

[SteveDenham]

I haven't done this analysis since mixed modeling became the standard for repeated measures analysis, but PROC GLM has a PROFILE option that will do exactly this within the REPEATED statement.  It will take some data shaping as you will have to get your response values into a wide format.  Here is an example:

data dogs;
   input Drug $12. Depleted $ Histamine0 Histamine1
         Histamine3 Histamine5;
   LogHistamine0=log(Histamine0);
   LogHistamine1=log(Histamine1);
   LogHistamine3=log(Histamine3);
   LogHistamine5=log(Histamine5);
   datalines;
Morphine      N  .04  .20  .10  .08
Morphine      N  .02  .06  .02  .02
Morphine      N  .07 1.40  .48  .24
Morphine      N  .17  .57  .35  .24
Morphine      Y  .10  .09  .13  .14
Morphine      Y  .12  .11  .10   .
Morphine      Y  .07  .07  .06  .07
Morphine      Y  .05  .07  .06  .07
Trimethaphan  N  .03  .62  .31  .22
Trimethaphan  N  .03 1.05  .73  .60
Trimethaphan  N  .07  .83 1.07  .80
Trimethaphan  N  .09 3.13 2.06 1.23
Trimethaphan  Y  .10  .09  .09  .08
Trimethaphan  Y  .08  .09  .09  .10
Trimethaphan  Y  .13  .10  .12  .12
Trimethaphan  Y  .06  .05  .05  .05
;

proc glm;
   class Drug Depleted;
   model LogHistamine0--LogHistamine5 =
         Drug Depleted Drug*Depleted / nouni;
   repeated Time 4 (0 1 3 5) profile / summary printe;
run;

But I would advise against doing this, and approach the analysis as a repeated measures analysis using one of the mixed model procedures.

SteveDenham