The paper being referenced was a systematic review and meta-analysis rather than new data. Given the inherent methodological and population differences between original studies , I doubt that much could be said about causality; it’s pretty conclusively an association though. And that’s probably enough basis on which to be vigilant and target support if you’re a therapist. But, yes, I suspect it’s more complicated than A->B. It was in my own case that nearly could have been a subject of this study.