I'm looking to analyze data (in SAS) from a trial investigating the effect of a particular protective padding on preventing skin irritation at multiple body locations among spinal surgery patients.

(i.e. "Is Tx1 associated with lower risk of outcome-Y, relative to Tx0/no-treatment?")

There are N=50 patients, each of whom had three(3) pads applied to one knee, one hip and one shoulder (exposed/Tx side), while the opposite knee/hip/shoulder was left bare (unexposed/non-Tx side). (Right vs. Left side was randomized.) Following surgery, patients were assessed for skin damage, and categorized ordinally into "None", "Redness/Bruising", and "Skin tear." (This could also be collapsed into a dichotomous outcome.)

Summary: N=50 pts, 6 locations on each pt (3 "exposed", 3 "unexposed"), with a tri-ordinal (1,2,3) OR dichotomous (0,1) outcome assessed for each location.

...I'm not all that experienced with analyzing correlated data, and am having trouble figuring out the most appropriate method. Hoping someone here can help point me in the right direction! Any suggestions on methods or resources (books, articles, blogs, etc.) would be much appreciated.

  • $\begingroup$ This seems like a small dataset, & in particular can be represented by a contingency table. Can you post it? $\endgroup$ – gung - Reinstate Monica Sep 27 '19 at 18:32
  • $\begingroup$ @gung Unfortunately I do not have the data set yet, I'm just trying to plan ahead. But yes, it is small, particularly given that I suspect the majority (80%-90%?) will be negative for bruising/tears. I'll post once I have a cleaned dataset though... $\endgroup$ – Sgolenbo Sep 30 '19 at 15:21

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.