Hyman2021 - Psychiatric disorders, grounded in human biology but not natural kinds

  • Type:#article
  • Year read:#read2022
  • Subject: (in brackets, can also bracket keywords in text)
  • Bibtex: @hyman2021
  • Bibliography: Hyman, S. E. (2021). Psychiatric Disorders: Grounded in Human Biology but Not Natural Kinds. Perspectives in Biology and Medicine, 64(1), 6–28. https://doi.org/10.1353/pbm.2021.0002

Example citation

Diagnosis of psychiatric disorders relies on phenomenology in the current DSM and ICD systems, but the categories are likely not discrete but rather represent underlying continuous traits where thresholds are needed to distinguish ill from well [@hyman2021].

Key takeaways

  • DSM-III imagined categories of psychiatric disorders, but evidence from genetics, neurobiology and epidemiology disconfirm this view.
  • Psychiatric disorders should not be seen as distinct categories but rather quantitative deviations from health
  • Signs that psychiatric disorders are grounded in biological phenomena: the symptoms are seen across cultures and genetic risk factors are shared across the world.
    • There is also pleiotropic in genes: shared vulnerability across disorders.

Without detailed understanding of the pathophysiology, psychiatric diagnosis relies on phenomenology (as in the DSM and ICD).

Categories vs dimensions

  • To say that something is a distinct category, we would expect a bimodal distribution or other discontinuous pattern, but most psychiatric traits are continuously distributed and follow a normal distribution pattern.
    • Good category: Infectious diseases where deviations from normalcy are clear.
    • Categories and dimensions can be combined: Type of cancer (category) plus stage and grade (dimension).
  • If we accept a dimensional approach, we usually resort to having certain thresholds indicate disorder or severity levels of a disorder.

There is no evidence for discontinuities between affected and unaffected individuals on most psychaitric traits.

A drawback of the DSM-style checklist structure is that early intervention and prevention has been difficult. Children and adolescents rarely fit a specific DSM diagnosis and may have shifting symptoms over time.

The DSM-III improved upon the previous state of affairs

The psychoanalytic tradition that dominated in the first half of the 20th century meant that each patient was seen as unique and there was an absence of shared characteristics.

A descriptive psychiatry that could make a strong case for the existence of “real”, discrete biologically based disorders represented a potent refutation of psychoanalytic theory and practice.

Unfortunately, the exuberant, if arbitrary, division of psychopathology into 265 disorders created the basis in DSM-III for artifactual comorbidity without the benefit of identifying homogeneous, well delineated classes of patients.

The author argues that the categorical paradigm that has been dominant since DSM-III has “impeded progress in understanding psychiatric disorders for decades, by ensuring that the vast preponderance of disease-related research is conducted on samples selected by DSM diagnostic criteria, most often with case-control designs that exclude or ignore comorbidity”. It may have been partly driven by Confirmation bias, lack of response to evidence that refutes the paradigm. This very much relates to the Fried 2015 paper (Depression is not a consistent syndrome).

Grounded in nature but not natural kinds

HiTOP: Haslam 2012, Krueger 2018 and Waszczuk 2017 RDoC: Insel2015 - Brain disorders Precisely

Caspi2018 - All for One and One for All - Mental Disorders in One Dimension

What are psychiatric disorders?

  • Natural kinds: Categories of things existing in nature that are well bounded and have stable, cohesive causal structures (e.g., chemical elements).

Psychiatric disorders are not natural kinds because they are heterogeneous, and symptoms are distributed continuously. We should have quantitative thresholds to distinguish ill from well (like blood pressure).

Bringing this discussion full circle, classifications are cognitive structures imposed upon data to make them more intelligible and useful for specified human purposes. A classification of psychiatric disorders not only provides an explanatory structure for data but also act as a lens through which clinicians and scientists, often unwittingly, see mental illness. A classification, especially one that is as deeply inculcated into practitioners and investigators as the DSM system is today, influences perceptions and inferences for better or worse. That does not make the symptoms and impairments of psychiatric disorders or their underlying etiology or pathophysiology less real, any more than Ptolemaic astronomy made stars and planets less real. It does mean that Copernican experiments with classification are much in need.