There's a strange contradiction in the findings, even just those in the abstract. When they look across individuals and try to predict antidepressant use based on exposure to woodlands, they find higher use among those who live closer to woodlands. It's only when they look at individual trajectories that they see reduced antidepressant use lining up with periods where they lived closer to woodlands.

It's almost like the same type of people who use antidepressants also like to live near woodlands. And they get depressed when they don't live near woodlands.