I have created a prognostic model using logistic regression. It predicts diabetes based on 6 independent variables (age, sex BMI, etc). I understand that Exp(Xi) = Odds Ratio of each (risk) factor in the model. I also understand that the probability of diabetes = exp(bo+ bi Xi) /((exp(bo + bi Xi) + 1) where Xi reflect relevant values for each predictor for an individual, and bo is the model intercept.
I would like to create a simple summary score based on the resulting (rounded?) beta coefficients. Can I simply sum the relevant ORs to make a score?
Thanks for your feedback..:yup:
I would like to create a simple summary score based on the resulting (rounded?) beta coefficients. Can I simply sum the relevant ORs to make a score?
Thanks for your feedback..:yup: