Abstract says they adjusted for potential confounders. Presumably wealth was one of them, as that seems like an obvious thing to control for in any health study.