Non‐specific psychopathology: a once and future concept

A popular strategy for criticizing diagnostic categories in psychiatry is to point out that two people can meet criteria for the same disorder, yet have few or even no symptoms in common. For instance, two people can meet the diagnostic criteria for major depressive disorder and share only one symptom. For post‐traumatic stress disorder (PTSD), two people can meet the diagnostic criteria and share no symptoms.

Some critics also enumerate the different ways to meet diagnostic criteria. To illustrate, there are 126 ways to combine the nine DSM depression criteria and meet the cut‐off of five to be diagnosed. Considering all combinations, there are 227 ways to meet criteria for depression using the DSM. Does this amount to 227 kinds of depression?

When introducing the concept of operational definitions, Bridg­man wrote: “If we have more than one set of operations, we have more than one concept, and strictly there should be a separate name to correspond to each different set of operations” ^{1 , p.10}. Certainly, we should attempt to understand the implications of different operational definitions of the same diagnostic concept, but some philosophers of science believe that Bridgman took it too far.

Let us look at an example from psychological testing. On the Minnesota Multiphasic Personality Inventory‐2 (MMPI‐2) depression scale, with 57 items scored 0 or 1, a person has to score 26 or greater to cross the depression threshold. Doing the math, there over 12 quadrillion ways for a person to score 26 alone. To claim that we should attempt to name over 12 quadrillion kinds of depression on the MMPI‐2 is absurd, and at least potentially makes the claiming that there are 227 kinds of depression seem somewhat silly.

One reason why two people can meet the same diagnostic criteria and share only one symptom is that operationalized diagnostic concepts typically under‐represent the symptom picture, i.e., they lack content validity. In part, this is because many nosologists adopt a convention regarding differential diagnosis which holds that, ideally, a diagnostic criterion set should indicate when a disorder is present, distinguish the disorder from non‐disorders, and distinguish the disorder from other disorders. In technical language, diagnostic criteria should be both sensitive to the presence of a disorder and specific to that disorder.

When non‐specific symptoms are de‐emphasized, two people who share only one depression symptom may nevertheless be similar on other common features of the disorder that are not included in the over‐specified criterion set. For depression, common but non‐specific symptoms include, for instance, anger, anxiety, depersonalization, gastrointestinal distress, headaches, and rumination.

In addition to being under‐representative, operational definitions are open concepts, meaning that new information and new uses for a concept can impel us revise the concept and extend it in different directions. According to the theory of open concepts, there is an inherent indeterminacy to the phenomena of psychiatry and, thus, psychiatric concepts cannot be closed off once and for all, because there are potentially further facts on the horizon that keep the process of defining and refining alive. This means that non‐specific symptoms which have been relegated to the background can be brought into the foreground, and vice versa. The historical transitions from classic hysteria to somatic symptom disorders and PTSD might be considered an example of background‐foreground shifts.

The mutable, protean nature of psychiatric disorders is not a new observation. Writing about hysteria in the 17th century, Sydenham noted that its symptoms varied so greatly and were so irregular that it was difficult to describe the disorder with any precision ² . More recently, psychopathologists have re‐recogniz­ed the relevance of non‐specific psychopathology.

One example is the pluripotential risk syndrome described by McGorry and colleagues ³ . Phenotypically broad and difficult to subtype, it is named a “syndrome” because the symptoms are associated with a decline in functioning. These symptoms include an intensification of normal traits such as worry and anger, and the appearance of novel features such as hypervigilance and compulsivity. The symptoms also ebb and flow in a “heterotypic” fashion. Heterotypic can refer to both the same risk profile having a broad range of outcomes (“multifinality”) and a single individual expressing shifting symptom pictures over time (“a divergent trajectory”) ⁴ . Symptoms in the ebb and flow may be transient and remit. Alternatively, they may develop into more specific risk syndromes for broad categories such as mood disorder or psychosis. This may be followed by a prodrome stage and eventually a specified category such as major depressive disorder, but such a linear trajectory is not the norm.

A second example is from factor analytic psychology. The general psychopathology factor “p” represents a common cause of and liability to all forms of psychopathology ⁵ . Higher scores on “p” are associated with varied and severe symptom pictures. One reason why it has been difficult to validate disorder‐specific etiologies may be because many risk factors and causes are themselves associated with the general factor (i.e., are non‐specific).

The “p” factor has been incorporated into the project to develop a hierarchical taxonomy of psychopathology and placed at the apex of the hierarchy. Underneath “p” are broad dimensions such as internalizing, externalizing, and thought disorder. Specified categories such as major depressive disorder and panic disorder are nested under the dimensions, but it is not foreordained that digging down to more specific constructs will be the most useful strategy. As an analogy, if someone is having an allergic reaction to pain medication, one might want to know if he/she took a non‐steroidal anti‐inflammatory drug, but whether it was specifically ibuprofen or aspirin is irrelevant.

Berrios ⁶ argues that the list of symptoms used to describe psychopathology was prematurely closed in the 19th century and it is unlikely to be extended unless psychiatrists attend less to diagnosing disorders and more to describing symptoms. Maj et al ⁷ argue that it would be helpful to have measures that assess the whole range of depression symptoms beyond what is contained in diagnostic criteria lists.

A potential barrier to a project of extension is that concepts such as depression have considerable face validity, due in part to their familiarity. This entrenchment may function as an a priori constraint if people assign more weight to symptoms that seem to fit with familiar concepts, and background those that do not.

One caveat to a shift toward the study of non‐specific symptoms that cut across traditional diagnostic categories is in reference to what 19th century European thinkers called “disease forms”. Parnas ⁸ and Thornton ⁹ argue that symptoms may seem non‐specific because they often refer to decontextualized, abstract concepts such as obsessions and anhedonia. In their view, symptoms can have more specificity within the gestalts represented by constructs such as schizophrenia. For instance, obsessions and compulsions can appear transdiagnostic on the surface, but be qualitatively distinct in different diagnostic contexts. To illustrate, on the psychosis spectrum, the content of obsessions and compulsions tends to be more sexual and aggressive and the symptoms have a delusional character in which, unlike for anxiety disorders, the person does not view them as irrational.

An important scientific goal should be to explain why psychiatric problems often begin with an intensification of non‐­specific symptoms that ebb and flow, in some cases being mutable or protean and in others settling into specified syndromes. The theory of open concepts also suggests that constructs for psychiatric disorders have been and will be “imperfect” not only due to a lack of knowledge, or because they are operationalized, or because they are descriptive, not etiological. They are also imperfect because of the inherent and inevitable limits to conceptualizing complex, noisy phenomena.