Loosi, thanks for that. LOS crossed my mind as a possibility after I posted.
|
Of course, hospitalization-based measures will be negative (for yes/no) or zero (for continuous values like LOS) for the noninfected vaccinated comparison population and would only vary in status with the infected cases. Not ideal from an analysis standpoint. But you could do a three way comparison: uninfected vs mild infection vs severe infection. That would be nice to see if the relationship they found still holds. Right now they're reporting a pooled effect.
The comorbidities are a nice inclusion because they are the factors associated with case severity and can be assessed for all individuals. I thought it a clever solution. Maybe I'm just easily impressed.
|
Responses:
|