Big data- for example, from nursing homes

I know there are all kinds of privacy issues, but with, for example, weekly testing in nursing homes, is there a way to pool the data and conduct large scale multilevel and/or logistic regression models to test for effects of simple things like diet, medication, supplements, ethnicity, preexisting conditions, etc on asymptomatic vs. symptomatic cases? My dad is the administrator of a nursing home and they test every week. About 40% of residents have tested positive at one point, with the majority being asymptomatic. I’m sure it’s been done, but I’m curious what are the predictors of asymptomatic vs symptomatic cases.