We have ran a project looking to reduce number of blood tests on wards. (data is about a thousand patients each time)

Looking at all the patients on medical wards over 1 month, measuring number of blood tests, plus: mortality, number of medications on discharge and length of stay during the observation period. Measuring to look if we can have a safe (non significant) reduction in blood tests.

And then following interventions we did the cycle a year later (same wards, different patients).

As I'm comparing seperate groups would it just be an unpaired T test to compare each measure before and after, and would this be ok to compare mortailty too? Or could we compare them (mortality, medicaions, LOS) together with a Chi squared?

Thank you in advance for your help!