ignorability.

A causal-inference assumption that treatment assignment is as good as random after conditioning on measured covariates.

Use it for

On the page

The estimate depends on ignorability after adjustment for age, baseline severity, and clinic.