I think the thrust of this summary post is right on the money. And that list of concepts and resource recommendations from NIMH is good to see. Co-occurrence is the rule rather than the exception, even though most of the hard-working folks invested in helping patients with substance use problems are not sufficiently trained to address non-addictive mental illnesses, and psychiatrists are not encouraged to embrace addictions in their clinical work where it appears they already have enough to do and/or their health systems define practice pathways that effectively cut patients up into pieces.
There's a great translational researcher named Andy Chambers from Indiana University who has studied and lamented this problem in psychiatry -- demonstrating that there is a woeful lack of necessary expertise at the physician level to properly take care of patients with co-occurring disorders. He also runs a basic science lab that primarily uses animal models to investigate the neuroscience of co-occurrence. I think it's a useful insight that his findings would suggest the answer to your question "what came first" at a foundational level is to say that both the chicken and the egg developed together. Clinically, it is useful to have conversations about when and under what circumstances patients first started to have difficulties... but it's equally useful to appreciate that the neuroscience would say vulnerability to both types of illness categories overlaps, as do the etiologic and perpetuating factors for them. For this reason, it's a terrible shame that our treatment programs are mostly siloed rather than integrated. For the same reason, I am grateful to have communities of like-minded colleagues like the virtual one here to share support in making things better in ways that recognize the accurate science and use it to deliver more sophisticated care.
Here are a couple of Andy's papers that might be useful -- an older concept manuscript that outlines the neuroscientific framework in an integrative way, and a newer one that shows real findings in support of that model. I find his work and thinking so compelling, because I think it has practical relevance to improving systems of care and to being properly holistic in approach to patients in the way that this original post advocates.
https://pubmed.ncbi.nlm.nih.gov/20585464/
https://pubmed.ncbi.nlm.nih.gov/30592364/
Cheers -- J.J.
