Figure 2:

Marginal Structural Model Outputs for Primary and Secondary Outcomes.

Estimates were obtained from fitting marginal structural Cox models adjusted for the following baseline covariates: sex, age>65 years, race, active smoker, BMI≥30, comorbidities on admission (coronary artery disease, congestive heart failure, hypertension, diabetes, dyslipidemia, chronic liver disease, active cancer, pulmonary disease), ACE inhibitor use, number of days since March 1st, 2020, and prior statin usage. The following time-varying covariates were adjusted for as well: ALC, WBC, AST, CRP, CK, ALT, and ICU admission status. Models were fit accounting for immortal time bias, time-varying confounding, and discharge as a competing risk.