Gravar-mail: Methodological challenges to control for immortal time bias in addressing drug effects in type 2 diabetes