We just published a new paper in JoPACS on the importance of socio-environmental causes of psychopathology (PDF), with the brilliant Merlijn Olthof and Anna Lichtwarck-Aschoff. In this brief blog, I will summarize the paper and add some context.

Mental health and socio-environmental risk

I’ve been thinking a lot about mental health problems as emerging from systems of biological, psychological, and social variables in the last few years. Two papers in 2022 were a bit of an eye opener regarding social or environmental factors in particular, or externalism, a framework emphasizing “the role of the relation between the person and the external environment in the constitution and maintenance of psychiatric illness” (reference).

First, in a short 2022 paper on externalist arguments against assisted suicide, Hane Maung engages in a thought experiment highlighting ethical challenges. Assume assisted suicide can only be carried out once a group of clinicians has decided that they have tried all possible ways to help a person. If we accept that causes for stress can also lie in the environment rather than just within the person, can we ever truly say we have exhausted all treatment options by providing e.g. cognitive behavior therapy and antidepressants to a person with depression? In his own words, “externalism underscores how social prejudices and structural barriers that contribute to psychiatric illness constrain the affordances available to people and result in them seeking medical assistance in dying when they otherwise might not have had under better social conditions”.

Second, the 2022 paper by Chater & Loewenstein powerfully differentiates Individual (I) and System (S) level explanations and interventions. In my own words and based on the keynote Lowenstein gave at APS this year about the paper, the carbon footprint is a good example: it was invented by a UK petroleum company trying to outsource responsibility from the S-level at which problems could be tackled (laws, taxes, regulations) to the I-level. Solutions for the obesity crisis have simultaneously been outsourced to the I-level (‘just eat less and exercise more’), rather than, at least in the US, having any sensible S-level legislation of what can be in food, how food can be advertised, what needs to be declared, and so on. The paper mentions addictions briefly, but I think there is a much larger point to be made here about mental health problems and therapy as well, on which I want to write a future paper. Our focus on resolving these issues is nearly entirely based on the I-level, individual therapy, when helping people getting jobs and getting them housed would likely be a much more impactful strategy at the population level.

And third, together with Don Robinaugh and a large group of interdisciplinary colleagues, we’ve been working on an extension or update or revision of Engel’s 1977 biopsychosocial model, for which I’ve done a lot of research into social and environmental determinants of mental health. More on this in a few month.

Externalist causes for the p-factor

This was the place I was at when Merlijn Olthof approached me a few months ago about writing something on external causes for the p-factor, a paper that came out just a few days ago (PDF) in the Journal of Psychopathology and Clinical Science. The paper is entitled “Reification of the P Factor Draws Attention Away From External Causes of Psychopathology”, and Anna Lichtwarck-Aschoff helped us write it up. It’s a short piece I won’t reiterate in full, but the gist is this.

1. Higher order factors have a long history in psychopathology research (Aschenbach, 1966), such as externalizing or internalizing disorders, and, more recently, the p-factor of psychopathology.
2. The p-factor summarizes covariance of a large number of transdiagnostic symptoms into a single score, which is often defined as a person’s vulnerability or liability to mental disorder.
3. If we now look at the theories that try to explain why some people have many symptoms / a high score on the p-factor, virtually all of these are personal-internal: intellectual functioning, disordered thought, negative emotionality, emotion dysregulation, and unspecified (individual) mechanisms such as genetic variation (we summarized these theories in a paper led by Ashley Watts, 2023).
4. That is, we are in a situation in the literature where we talk about causes for the statistical construct p-factor—which is basically just a sum of all symptoms in a dataset (reference)—as being purely person-internal.
5. The first problem then is reification, which we briefly tackle. This is a common problem in language use and thinking in mental health science, e.g. when we say a person’s attention-deficit hyperactivity disorder makes them hyperactive, instead of saying that their hyperactivity is classified as attention-deficit hyperactivity disorder.
6. The second problem is the internalization of mental health causes, and here the two papers by Maung and Chater & Loewenstein above fit in very nicely. However, causes for transdiagnostic symptom load need not reside within individuals. Symptoms can also hang together because of environmental causes, such as poverty, structural discrimination, and victimization. We cite some work strongly supporting this line of reasoning, e.g., parents of families living in intergenerational poverty will score high on symptom criteria such as worry (about their finances), anxiety (about their children’s future), feelings of reproach (about the chances they can offer their children), and so on, all primarily caused by their circumstances.
7. We conclude that communicating the p-factor as a person-internal vulnerability can be stigmatizing and harmful, as understanding oneself as a chronically vulnerable person, irrespective of context, may induce feelings of hopelessness. In the worst case, this within-person reification of risk may hide issues of social injustice from sight in both research and practice. And this goes back to Chater & Loewenstein in terms of clinical relevance, given that structural adverse life circumstances beg for societal interventions and not (only) medical or psychological ones.

Due to the journal’s very tight word limit, some of the work I mention above have shaped my views on these and relateed topics, but didn’t make it into the paper, so I wanted to add them here in some more detail and explain how they have shaped my view. Thanks to Merlijn and Anna for the collaboration, and Ashley, Hane, as well as Chater & Loewenstein for the inspiration.