Mixlogit: group() vs. id()

Dear Statalist,
I am working with insurance choice dataset, where individuals choose insurance plans once a year. I have panel data, which means that I observe the same individuals choosing plans in multiple years. I am trying to estimate a mixlogit model of the effect of plan specific characteristics on choice. I have a variable id which identifies the individuals across years. Since the choice is made each year, I generated idyear which creates unique identifiers for each individual in each year. When I run mixlogit, I use group(idyear). When I try to specify group(idyear) and id(id) together, I get an error message.
Do I need to specify id(id) if I set the group(idyear) variable to be person/year specific? I have read through the Stata journal article on mixlogit multiple times, and the section on group vs. id does not clarify my issue.
I would appreciate any advice/discussion about the role of group() and id() in the panel context, and whether I am misspecifying my model with this current set-up.

Thank you,
Shooshan