How would you determine whether observed differences in healthcare outcomes are due to confounding factors?
