Unit of Analysis


Trying to determine which analysis is appropriate. Let me explain:

-We have access to a community hospital records on the patient level.
-We are measuring some attribute of patients, say blood pressure. All the blood pressure measurements are under an implicit condition at this time, such as hospital policy.

-In 2 months, we collect more blood pressure readings, now from a different set of patients than we did initially, but from the same hospital, and under another condition, such as the hospital instituting a revised policy.

-Our goal is to show the hospital the change in average blood pressures from the policy change. We are interested in performance measurement pretty much.

-Just want to verify this is an independent groups design despite being in the same hospital (though how might we still account for dependencies and unobserved confounders?)