Eiko Fried, a professor in the Department of Clinical Psychology and Leiden University, has a new paper arguing for a paradigm shift in research practices in mental health. His approach to studying mental disorders as complex systems highlights how current understandings of psychiatric diagnostic categories may be convenient fictions.
“Research is often carried out as if DSM categories cut nature at her joints,” Fried writes. “Because diagnostic labels are convenient fiction – because diagnostic categories and thresholds to distinguish healthy from sick are manufactured rather than discovered – the question remains how useful, such designs can be.”
Historians of psychiatry have suggested that the diagnostic categories within the DSM are indeed manufactured, not “discovered.” But much of current psychiatric research design reflects twin approaches to mental disorders that oversimplify psychiatry’s history: diagnostic literalism and reductionism.
Diagnostic literalism assumes that ascriptions of psychiatric diagnosis are explanatory (“Why is Peter so sad? Because he has depression”). But as Fried points out, most diagnostic categories in the DSM have been shaped by historical forces and cultural attitudes, not empirical evidence. Reductionism proposes a hierarchy of factors for understanding the person, with one type – physiological forces – being primary. But reductionist models of the mind, too, have proved to be a dead-end for psychological researchers seeking meaningful change.
As Thomas Insel, former NIMH director, is quoted as concluding, “[after] 13 years at NIMH pushing on the neuroscience and genetics of mental disorders … I succeeded at getting lots of really cool papers published by cool scientists at fairly high costs – I think $20 billion – I don’t think we moved the needle in reducing suicide, reducing hospitalizations, improving recovery for the tens of millions of people who have mental illness.”
The reductionist approach in research, paired with the assumption of diagnostic literalism, has provided results of little utility – if not outright harm.
On the other hand, many robust phenomena can be explained by a systems or continuum approach to mental health disorders: for example, that inter-rater reliability for diagnosis is often only moderate; that many diagnoses feature marked symptom heterogeneity; and that considerable comorbidity between diagnostic categories. If today’s diagnoses are not best described as categories but are instead rough clinical approximations that refer to problems that lie on a dimension of severity, these robust phenomena – including especially the fuzzy boundaries between disorders – are easily explained.
The shift to a systems approach, and the rejection of both diagnostic literalism and reductionism, has important consequences for research that has for decades proceeded from the assumption that DSM categories cut nature “at its joints.” On this view, both biological mechanisms and psychological and environmental processes play complex roles in mediating disease trajectory.
“It is therefore questionable,” Fried writes, “how useful current research efforts can be that map biological (or other) markers onto specific DSM labels that are, for the most part, convenient fictions.”
Fried thus argues that a systems approach to studying mental disorders is necessary to reflect both the complexity and contingency of our current diagnostic categories in psychiatry. This involves widening our focus to biopsychosocial elements of mental health systems and moving beyond the current focus on symptoms.
If mental health phenomena are not simply physiological processes but emergent properties of complex, interrelated systems, they can no longer be decomposed into simple cause-effect relations. Thus, “understanding the macro-level behavior of mental disorders will require building more interdisciplinary bridges between biological, psychological, and social investigators.”
Fried goes on to argue for the necessity of mathematical models over the “ambiguous verbal accounts currently dominating the literature.” If the researcher’s task, in Fried’s view, involves translating verbal data into mathematical inputs, however, she must be careful not to simply naturalize or reproduce the inequitable and ineffective systems in which they operate.
Fried, E. (2021). “Studying mental disorders as systems, not syndromes.” (PsyArXiv preprint) 10.31234/osf.io/k4mhv (Link)
Editor’s Note: Eiko Fried, the author of the article covered here, removed the phrase “convenient fiction” from the preprint article three days ago. Dr. Fried’s comment is below:
Thanks for featuring the preprint. I removed the core term “convenient fiction” that you use in the header 3 days ago, because I found the term can be easily misunderstood. I updated the preprint before you featured the piece (so this is *no* criticism of your piece), but I wanted to make sure to let your readers know.
The relevant section of the manuscript now reads:
“The superimposition of categorical diagnoses on the complex landscape of mental health problems explains the robust phenomena I listed above (Hyman, 2021). As Ash put it in 1949: “it is likely that the lack of congruity between the diagnostic label and the complexities of the biodynamics of mental structure is itself at the heart of diagnostic failure” (Ash, 1949). While many clinicians and researchers are aware of these complexities, clinical research is often carried out as if DSM categories cut nature at her joints. Scientific studies often take diagnoses for more than they are, which I refer to as diagnostic literalism. Case-control studies dominate the literature, where a healthy control group is compared to a group diagnosed with one specific mental disorder, regarding some variables of interest. Because mental health categories and thresholds are manufactured rather than discovered, the question remains whether such designs are optimally positioned to inform research on psychopathology (Cai et al., 2020; Hitchcock et al., 2021). Of note, this position should not be mistaken view the view that mental disorders do not exist, do not involve suffering, or should not be treated. My point is that our categorization of mental disorders remains somewhat arbitrary.”