Hello,
I have created a similar dummy data to explain my question.
I am trying to analyze the effect of repeated dose of certain drug (Treatment group) given, and the tumor volume was measured after each drug was administered 3 time (TV 1-3). I found a relative change in TV since my values are not similar. I wanted to use the relative change to calculate the effect of repeated dose of group1 and group 2, group 1 and group 3, group 1 and group 4?
What type of logistic regression would be appropriate and the ideal code. Thank you in advance.
Treatment Group | TV1 | TV2 | TV3 | Relative change in TV |
1 | 574.6 | 386.8 | 305.2 | -0.47 |
1 | 211.5 | 298.7 | 277.8 | 0.31 |
1 | 606.3 | 643.1 | 503.5 | -0.17 |
2 | 591.3 | 403.4 | 457.0 | -0.23 |
2 | 279.2 | 362.7 | 429.3 | 0.54 |
2 | 637.2 | 358.1 | 318.4 | -0.50 |
2 | 520.5 | 511.0 | 447.1 | -0.14 |
2 | 243.5 | 369.2 | 243.4 | 0.00 |
2 | 280.2 | 248.6 | 405.6 | 0.45 |
3 | 774.9 | 809.2 | 708.1 | -0.09 |
3 | 446.3 | 479.2 | 418.9 | -0.06 |
3 | 418.1 | 579.9 | 419.2 | 0.00 |
3 | 629.2 | 507.7 | 359.1 | -0.43 |
4 | 329.5 | 276.6 | 321.9 | -0.02 |
4 | 867.7 | 1110.6 | 895.7 | 0.03 |
4 | 105.5 | 229.3 | 193.8 | 0.84 |
4 | 657.0 | 597.4 | 667.3 | 0.02 |
I think you should use Mixed model by PROC MIXED or PROC GLIMMIX, since your data is repeated measure data(a.k.a have different visit number).
And your data did not contain binary/binomial variable, you could not model it by LOGISTIC MODEL.
And you could use LSMEANS or LSMESTIMATE statement to get the difference between groups.
https://support.sas.com/kb/61/830.html
And @lvm @StatDave @SteveDenham could give you more advice.
What you do depends on what exactly is the research question. Nonetheless, to make this a typical logistic regression problem you need a yes/no outcome. You could form that by having the clinicians select a meaningful threshold, like TV reduced by 10%, yes/no. That makes for a logistic regression, but dichotomizing data often reduces power, and sort of muddies the interpretation. In my example, a reduction of 9.9% is a 'no' while 10.00 is a 'yes', but is there a meaningful difference in these two volumes?
Relative % change = (TV1-TV3)/TV1 = 1-TV3/TV1. Even if TV is normally distributed the ratio TV3/TV1 almost surely isn't normal. From your data example it doesn't look like TV2 is used. So you could use the Kruskal-Wallis test, available with the Wilcoxon option in proc npar1way, to compare groups and not have to be concerned with parametric assumptions. If the overall test is significant, the DSCF for the pairwise comparisons.
There is a literature, mostly based on simulations, suggesting that ANCOVA with TV1 as the covariate is the preferred analysis. Here is an easy read on the subject: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC34605/
Join us for SAS Innovate 2025, our biggest and most exciting global event of the year, in Orlando, FL, from May 6-9. Sign up by March 14 for just $795.
ANOVA, or Analysis Of Variance, is used to compare the averages or means of two or more populations to better understand how they differ. Watch this tutorial for more.
Find more tutorials on the SAS Users YouTube channel.