Thomas, you might want to review the Centers for Disease Control and Prevention, Bahavioral Risk Factor and Surveillance System database at http://www.cdc.gov/brfss/ it represents a database relate to a series of health questions targeting population health studies. Also SEER data which another set captures data related to disease prevalance, incidence and mortality rates by gender, race and ethnic status. I hope this helps. Michele
How will you compare the data from various sources? Do they match semantically? Perhaps have a look at detailed clinical models. See e.g. www.openCIMI.org
The MIT has just released a new version of their anonymised open access Intensive Care database: MIMIC-3. I have worked with version 2, and I can confirm the high quality of the data. Many incredible papers have been produced from MIMIC. The version 3 now contains nearly 60,000 ICU admissions, and the data structure has been simplified. Highly recommended!