We typically think about different mental disorders as distinct conditions, which each requires its own treatment strategies and has its own risk factors. But findings from epidemiological and genetic research challenge this view: mental disorders overlap to a large extent and we should not see them as distinct categories but emerging from a common source. This is what is sometimes referred to as the p-factor. Avshalom Caspi and Terrie Moffitt, two researchers who have done a lot of work on the p-factor, have the following quote in one of their articles:

“empirical evidence has now accrued to suggest that a single dimension is able to measure a person’s liability to mental disorder, comorbidity among disorders, persistence of disorders over time, and severity of symptoms.”

Reliability over validity in the DSM

In order to understand where this comes from, it’s helpful to take a step back and look at the development of the “bible” in the diagnosis of mental disorders, the Diagnostic and Statistical Manual (DSM) issued by the American Psychiatric Association. The DSM is refined and updated every now and then, and the criteria for certain disorders change over time. For example, in the fifth and latest edition of the DSM (from 2013), hoarding is seen as a separate disorder from obsessive-compulsive disorder. The system of using checklists of symptoms to assess whether an individual has a mental disorder, however, has been unchanged since the DSM-III from 1980.

Prior to the DSM-III we didn’t have a reliable way of diagnosing mental illness. For example, two psychiatrists might have interpreted the same clinical presentation of symptoms in completely different ways.

“One diagnostician would say, “Yes,” another, “Obviously not!” Such unreliability made research impossible, and psychiatry’s scientific aspirations laughable.” - Nesse & Stein, 2012

The solution was to use checklists of symptoms, of which a certain number had to be fulfilled, to establish a diagnosis. Psychiatric research made huge progress with the introduction of DSM-III. Researchers were able to use standardised interviews in epidemiological surveys, treatment studies used consistent inclusion criteria and outcome assessments, and neuroscientists could look for brain correlates to reliably defined conditions. The DSM-III (and later editions) improved the reliability of psychiatric diagnoses (although it’s not perfect), but made less progress in terms of validity. In other words, psychiatry was able to measure consistent phenomena across time and space, but did it measure the right things?

“It’s easy to think up reliable ways of classifying mental states, but diagnoses must also be useful. They must describe genuine impairment that can be treated by mental health professionals.” - Vaughan Bell

One problem is that there is heterogeneity within diagnostic categories: two individuals may qualify for depression without having any specific symptom in common. There is also no clear line separating individuals with and without a disorder, since the term “clinically significant distress” can sometimes be hard to interpret consistently. Finally, the boundaries between different disorders are fuzzy and comorbidity is common. Why, if mental disorders are distinct conditions, do we see this?

A hierarchy of psychopathology

It could be that so-called higher-order constructs play a role. For example, one common categorization is between internalizing, externalizing, and psychotic symptoms. Specific diagnoses can be clustered into these categories: depression and anxiety disorders go into the internalizing cluster, ADHD and substance use disorder go into the externalizing cluster. The thing is though, those higher-order constructs can be explained by a general underlying susceptibility for all mental disorders, the p-factor.

Several findings support the notion of a general p-factor of psychopathology. First, mental disorders are highly correlated. If an individual meets diagnostic criteria for one disorder, they are likely to also meet criteria for another disorder. This appears both in cross-sectional studies and longitudinal studies. Second, psychopathology is heritable but the heritability is not very specific. Children of parents with a mental disorder are more likely to develop any mental disorder, not just the one(s) that their parent(s) have been diagnosed with. Third, there is a large overlap among common mental disorders both in genetic studies as well as brain imaging studies. All of these findings point to a general, not specific, susceptibility and display of mental illness.

What are the implications?

I think there is a strong case for taking higher-order constructs (such as the p-factor) into consideration in psychiatry. I think that p-factor research will be influential in how future editions of the DSM are structured, which will have huge consequences for how clinicians and the general public view mental illness.

Should we consider treatment strategies across disorders if we accept that they represent a more general factor of psychopathology rather than distinct conditions? We already do that to some extent, for example by using exposure therapy as a psychological treatment for many different anxiety disorders. Other efforts include the unified protocol approach, where trans diagnostic treatment strategies are emphasised over diagnosis-specific ones. But trans diagnostic approaches need to be carefully evaluated before we change clinical practice, particularly in areas where there are single-disorder treatments that have been proven to be highly effective.

The implications for research, however, are more clear. I think we should make an effort to study and evaluate symptoms outside the “main diagnosis” under study, in order to check whether there are other patterns that might explain why an individual responds to a certain treatment, for example. I look forward to seeing where the idea of a p-factor will lead, and the ways in which it might change our view of mental illness.