What kinds of things are psychiatric disorders? - Kendler, Zachar, Craver (2011) - Article

The article “What kinds of things are psychiatric disorders” written by Kendler, Zachar and Craver was published in the journal “Psychological Medicine” in 2011. It deals with four different models how we can classify psychiatric disorders, namely the essentialist kinds, the socially constructed kinds, the practical kinds and the mechanistic property (MPC) kinds. The authors present each of these models and then evaluate them according to their practicability and usefulness.

Essentialist kinds

Our first model the essentialist kinds model proposes that everything in our case each specific disorder has one underlying essence. This essence is thought to be inherent in every individual that suffers from a disorder D. All individuals that do NOT suffer from disorder D will not have this underlying essence. The essence is a single, well-defined etiological agent that stands in direct causation to each key defining feature of the disorder.

To clarify the concept we will use the eating disorder anorexia nervosa as an example. So we start with choosing an essence as the fundamental cause for the disorder. Everyone who suffers from anorexia nervosa should have this essence and people who do not should not have it. The constant urge of losing weight might be our essence. This impacts and causes the key features of anorexia, like the preoccupation of thoughts with food, losing pleasure and interest in many activities, exercising a lot and so forth. Maybe you have noticed some flaws in this concept. So can we ever be sure that we chose for the right underlying essence? And can we ever be sure that this essence is really the only causal agent in our disorder? Most times we cannot. Usually disorders are caused by several very different factors, as we have learned so far (remember “Nature and Nurture!”). Thus, the concept of essentialism simplifies our concept of psychiatric disorders too much and reaches its limits. In addition it blends out the fact that features of a disorder are interdependent and thus interact with one another. Regarding biology it would not allow for any biological explanations since biological factors as genetics are too diverse, flexible and ever-changing.

Socially constructed kinds

The second model of interest is the model of socially constructed kinds. This model suggests that kinds of psychiatric disorders are constructed by the categorizations societies and cultures introduced to them. Surely it is known by now that social factors have a large impact on the development of psychiatric disorders. Further the authors agree that cultures define and conceptualize psychiatric disorders. The idea which they reject is that the existence of a specific psychiatric disorder is merely based on a cultures conceptual distinction of it. In their opinion we should find classifications for psychiatric disorders that integrate common biological, psychological and social factors and account for the distribution of specific disorders across cultures and history. Depression would be a popular example as a disorder which has existed for hundreds of years and occurs across different cultural contexts.

Practical kinds

Third, the authors present the model of practical kinds. As its name suggests it uses an instrumentalist approach to science, thus focus on elements that are most useful in practice. Regarding psychiatric disorders this means that a good categorization of a certain kind of disorder would help us to make reliable diagnosis, prognostication and treatment selection. If the diagnosis of someone with a psychiatric disorder would help us to describe how this person will behave in detail as well as how it can be cured, we can say that this kind would be very helpful for science. This atheoretical/descriptive approach was adopted by the DSM-III, DSM-III-R and DSM-IV.

The authors regard this practical kinds approach as more useful than essentialism and social constructionism, since it embraces both of these aspects without reaching its limits so fast. Still they see its limitations as it does not give any clear advice as how classifications should be built.

Mechanistic property cluster (MPC)

To solve the problems of the models described so far the authors introduce the mechanistic property cluster (MPC) model. This is supposed to capture all the various and complex mechanisms that underlie, produce and sustain psychiatric syndromes. The mechanistic property cluster (MPC) suggests that there are indeed essence-like underlying explanatory structures which are inherent to most psychiatric disorders. Thus, they are somewhat universal. But it also cautions us that these are messy as well as interrelated and therefore hard to disentangle. Further it emphasizes that these essences are rather regarded as relatively stable complex interactions between behavior, environment and physiology. They might have arisen through development, evolution as well as an interaction with the environment.

The authors emphasize that the mechanistic property cluster (MPC) model is open to systematic differences that might arise within subpopulations. So as well as we have different hair colors and body weight or food preferences across cultures we might have hybrids in psychiatric disorders. Still, the fuzziness of boundaries that might occur because of that does not take stability from them. As soon as we have identified these underlying properties that interact within our psychiatric disorder we can use the mechanistic property cluster (MCP) to predict, explain and control these kinds. Therefore knowing the MPC kinds clusters will enable us to make inferences about the past, present and future of an item that operates within a specific kind.

The mechanisms just described will typically span several levels and thus will also include interactions among specific symptoms. For example the psychiatric disorder of phobia leads to avoidance of the feared stimulus. This avoidance will prevent the patient from habituation to that stimulus. In schizophrenia, hallucinations are often causing delusions. Therefore, the authors conclude that individual symptoms of a psychiatric disorder interact with one another to sustain the other to be characteristic of the illness. That is why illnesses are often regarded as composed of a relatively stable set of traits. This can also create self-fulfilling negative expectations that worsen the overall condition of a patient suffering from a psychiatric disorder. For instance the cognitive biases towards negative information that are found in depression compound the depressive state even further.

A cause of a psychiatric illness might be psychological as well as biological. It also might be that a cause that was psychological at first, for instance substance dependence resulted in changes in biological levels which now in turn contribute to causing the dependence.

Mechanistic property cluster vs. essentialist kind

You may have noticed the similarity between mechanistic property cluster (MPC) and our first model of the essentialist kind since both argue that kinds of psychiatric disorders are grounded in shared features of the causal structure of the world. Still they differ in various ways:

MPC recognizes that there might be numerous causes for a specific kind of a psychiatric disorder to exist including evolutionary, developmental, genetic, physiological, psychological, behavioral and social.Essentialists argue for only one underlying mechanism (essence) that causes a specific kind of psychiatric disorder.

MPC acknowledges probabilistic relationships between relevant causal factors and the occurrence of specific symptoms. Thus, as a cause may change the probability that a specific symptom or a set of symptoms arise may change.

Essentialists on the other hand emphasize deterministic causes.

MPC assumes that the same network of symptoms might be caused by different underlying mechanisms. Thus, MPC kinds are “multiply realizable” by the causes or sets of causes that produce them.

The authors predict that more information about underlying mechanisms will provide new possibilities for classification. To get them might be rather challenging due to the potential overlap.

The concept of mechanistic property cluster (MPC) was retrieved from the philosophy of biology as the authors state it. Borsboom used a similar approach but relied more on a psychometric perspective. He argues for a “causal systems perspective” as psychiatric disorders should be seen as “sets of symptoms that are connected through a system of causal relations”. Factor analytic concepts have revealed that there are indeed underlying liabilities to illness which causes their symptoms.

Conclusion

The authors conclude their article with arguing that mechanistic property cluster (MPC) model provides a prescriptive guidance for classifying the complex concept of psychiatric disorders. It helps to link psychiatric illnesses as closely as possible to our growing knowledge of the causal factors that operate in producing, sustaining and hopefully helps to prevent and treat them some day. Even though not perfect, the authors argue that the mechanistic property cluster (MPC) kinds approach offers the most promising answer to the question “what kinds of things are psychiatric disorders?” with which they started their article.