Leukocyte count-ischemic stroke. Regression type?

#1
Hello,

I have a study which includes 293 patients and I would like to see if the leukocyte count on admission has an impact on predicting the mRS(a score which quatifies the severity of the outcome of an ischemic stroke). I transformed the mRS in a binary variable (0=negative outcome, 1=positive outcome). I also have other risk factors for stroke such as hypertension, smoking, etc. which might influence the mRS. What type of regression should I perform? I used multiple logistic regression in GraphPad Prism 8, but all the risk factors seem to have OR <1. I don't really know how to interpret these results and I don't know where I made a mistake. I attached a picture of the data I used and the results. I appreciate your support!
 

Attachments

hlsmith

Less is more. Stay pure. Stay poor.
#2
What does the mRS score look like (format, range, etc.)? It is typically not a good idea to collapse variables because you lose information.
You also have a fairly oversaturated model (a lot of predictors per event [6.3]) and it looks like only 2 variables were associated with the outcome. If you feel like all of the estimate were the wrong direction the model may be using the wrong group as the reference. Not an SPSS users, so you will have to look at their documentation.
 
#3
What does the mRS score look like (format, range, etc.)? It is typically not a good idea to collapse variables because you lose information.
You also have a fairly oversaturated model (a lot of predictors per event [6.3]) and it looks like only 2 variables were associated with the outcome. If you feel like all of the estimate were the wrong direction the model may be using the wrong group as the reference. Not an SPSS users, so you will have to look at their documentation.
Thank you for your response! The mRS is a categorical variable ranging from 0 to 6. I read the documantation and it says that when the odds ratio is under 1, the odds of a positive outcome decrease as the independent variable increases by 1. I guess the reason why odds ratio of leukocyte count is 0.999 is because the numbers are too big comparing to the other variables and it makes little difference on the outcome.