{"kind":"user","text":"We are estimating the average treatment effect (ATE) of the new drug on the amount of days the patient stays in the hospital, and we already controlled for the severity. Should we also control for the 'hospital' variable? Please answer with 'yes' or 'no'."}