Advice for a multilevel model

[The Burl]

Any advice on this would be help me set this set of analyses up.

I have a three-level model.The nested groups are:
Individual
Work area (e.g., "radiology")
Hospital (e.g., Johns Hopkins)

The sample is very large, 530K individual responses in 1080 hospitals that each has 15 pre-defined work groups.

The data are survey in nature with 10 dimensions and 2 outcomes, measured all in 5 point Likerts.

What is the first step I should take? Calculate ICCS? I know MLM is essentially a multiple, multiple regression analysis...

I also know that a benefit of MLM is comparing across the groups but it seems a bit impossible to compare across all 1080 hospitals.

Any nods in the right direction would be appreciated!
Jon

[The Burl]

I also add that this is likely not enough information for a specific conversation. I really just want to get some discussion started and can fill in details later.

Thanks!