This research is highly debated and is by no means a sure thing. The research so far has been correlational, inconsistent, and weak. Twenge
hugely overstates the confidence we should have in the findings.
When I worked for a research lab for a university attached to a hospital, facebook usage was highly correlated with diagnosed depression. it was unclear whether depressed people used facebook more, or using facebook more caused depression.
the correlations with facebook usage and other illnesses was weak, but depression was not.
See https://medium.com/@OrbenAmy/social-media-and-suicide-a-crit... and https://twitter.com/shuhbillskee/status/955070356091555840 for some critiques of the work.