melogit PASS ib(1).MODEL || id:, or allbaselevels
estimates store base
melogit PASS ib(1).MODEL ib(1).E || id:, or allbaselevels
estimates store base_E
melogit PASS ib(1).MODEL ib(1).E ib(1).SEX || id:, or allbaselevels
estimates store base_E_S
melogit PASS ib(1).MODEL ib(1).E ib(1).SEX ib(1).AGE || id:, or allbaselevels
estimates store base_E_S_A
melogit PASS ib(1).MODEL ib(1).E ib(1).SEX ib(1).AGE ib(1).BMI || id:, or allbaselevels
estimates store base_E_S_A_B
estimates stats base base_E base_E_S base_E_S_A base_E_S_A_B
* no need to adjust for BMI, all other variables are relevant to adjust for
* final model 
melogit PASS ib(1).MODEL ib(1).E ib(1).SEX ib(1).AGE || id:, or allbaselevels
* checking for intertaction
melogit PASS ib(1).MODEL ib(1).E ib(1).SEX ib(1).AGE b(1).MODEL#b(1).E || id:, or allbaselevels
estimates store model_E
melogit PASS ib(1).MODEL ib(1).E ib(1).SEX ib(1).AGE b(1).MODEL#b(1).SEX || id:, or allbaselevels
estimates store model_S
melogit PASS ib(1).MODEL ib(1).E ib(1).SEX ib(1).AGE  b(1).MODEL#b(1).AGE || id:, or allbaselevels
estimates store model_A
estimates stats base_E_S_A model_E model_S model_A
* none of the interaction significant 
* final model
melogit PASS ib(1).MODEL ib(1).E ib(1).SEX ib(1).AGE || id:, or allbaselevels
*
* 3M 1873V and 3M 8833 signmificantly better than reference, factor 1.5
* Cardinal significnatly worse than reference, factor 2
* others no difference to reference
*

* the fit is worse for black, asian and mixed in comparison to white
*
* the fit is better for men than women
*
* the fit is similar for all age groups except for the very young where it is worse
*