Effect of Education with Telephone Follow-Up on Diabetes-Related Wound Reoccurrence in Adult Patients

#1
I am doing a Retrospective Chart review with an Intervention.


I am investigating the relationship between the use of education with mobile phone bi-weekly follow-ups in the reduction of diabetes-related wound reoccurrence.

What test would you use for this?
I would be comparing the rate of days before the patient wound recurred in the retrospective vs post-intervention.
I also did a knowledge quiz pre and post-test for the intervention group.

PLEASE HELP IM STUCK
 

hlsmith

Less is more. Stay pure. Stay poor.
#2
A regression model, likely logisitic regression. You need to use a regression type model since you will likely need to control for differences in patient characteristics between time periods within the model.
 
#3
if its a pre-post intervention time to event, I think 'frailty' model is what gets used for the correlated time to event? Apparently sas supports thee now but I have never had occasion to use them. But if you can dichotomize the outcome somehow logistict regression or mcnemars test is going to preserve your sanity more.
 

hlsmith

Less is more. Stay pure. Stay poor.
#4
@fed2 - good catch. I just saw the reoccurrence and neglected to address time. Survival analysis could be used in order to censor those that did not have an event. I would still control for background difference via covariates or weights. Also, do you have complete capture of the outcome, meaning if they went to another facility for treatment would you know, or do people get lost-to-follow-up?