• Allows systematic comparison and meaningful inferences between groups

• Great way to learn in areas of controversy

• Participation in RCT often is of benefit to the patient’s care

Caveats:

• Primary outcome, sample size/power and analytic issues can pose challenges

• Inferences may be limited by generalizability (eligibility criteria, outcomes)

Inferences must be made based on observational studies and subject to numerous biases:

• Confounding (residual, by severity)

• Misclassification bias

• Immortal time

For a nice overview of the issues as illustrated by COVID-19 see Tleyjeh et. al, OFID 8(8) 2021