Can the odds ratio change based on a multiple logistic regression model?

Question

[Bioman]

I have an assignment asking me to fit a model including age, sex, type of diagnosis and BMI. I would use the multiple logistic regression model which takes 1 nominal variable (the diagnosis type 1 or 0) and multiple variables, age, sex, and bmi. but the assignment is then asking for the odds ratio off getting diagnosis type 1 using diagnosis method 1 of 2.

~My question is, how can age, bmi or sex change the outcome of the diagnosis, if I already have that data? I have been given the data which already tells me how many people have been diagnosed and with which method. knowing the sex isnt gonna change the fact that the person is diagnosed 1 and that wont change the odds ratio.~

I think i figured it out. since the dependent variable in this case is either or, either diagnosed or not. I believe i have to fit the model using a multiple logistic regression model, and then calculate the odds ratio using the estimated value based on the model

Answer 1

[rpolicastro]

I'm not completely sure what the question is, but in general if you are building a model it's for inference and/or prediction. With inference you try to understand the relationship between variables, e.g. how does being a male effect your chances of being diagnosed compared to if you were a female. Prediction on the other hand lets you guess the diagnosis of an unknown individual given the particular variables you included when building the original model. Since it's asking for a log odds ratio of one of your predictor variables (diagnosis method), the question is likely asking for inferential statistics, or how likely the diagnosis method is to determine diagnosis 1 from 0.