Mental disorder is a cause of crime: "crime" vs. psychiatric phenotypes

The term “crime” is in no less need of a precise definition than “mental”, “disorder”, or “cause”. Leaving aside the legal definition, let us consider how crimes, generally in the form of violent, sexual or aggressive behaviours against others, are approached from the perspective of being caused by mental disorders. 

The focus on mental disorder also directs the searchlight of forensic psychiatry towards individual criminal acts or towards patterns of criminal behaviors occurring in individuals rather than to crime as a societal or group phenomenon. For a number of questions, this is too narrow a perspective.

A crime takes place in a situation, between people, and the vast majority of crimes are clearly influenced by the situations in which they arise. Only rarely is a crime planned and determined by a single mind. A major shortcoming of the psychiatric approach is the emphasis on the individual and the relative down-tuning of the role of the interaction between people, including co-perpetrators and victims. 

The capacity to empathize and act compassionately shows not only a constitutional inter-individual variation but also an intra-individual variation in partially state-dependent actual functioning (cf. Constantino & Todd, 2003; Gabbard, 2004). Each and every one of us may stop forming meta-representations of the other’s mind, the ordinary household quarrel being just as good an example as more dramatic scenes of conflict. 

A person who commits a heinous crime on his own is more likely to differ from the normal variation on at least some mental features than someone taking part in a similar crime as part of a group of offenders. Even small groups may release dynamics that deprive their members of inhibitory forces. A mathematical hypothesis to predict an individual’s actual capacity for empathy (E) would assume that his or her natural capacity for empathy (e) should be divided by the square root of the number of people (n) involved and interacting in the actual act.

Another situational factor that plays a major role in the background to many violent crimes is the influence of drugs. These effects are not easily defined in relation to other mental factors or to situations. Alcohol, for example, may trigger aggression and reduce inhibitory faculties but can also diminish reactivity and reduce anxiety, thus acting as a susceptibility factor or as a protective factor depending on the situation, the degree of influence, and the subject’s other psychological and psychiatric problems. When faced with the task of explaining the background to a particular criminal act, aspects of reduced or changed mental abilities have to be considered in the context of situational, social factors, each of which may constitute an INUS condition.

Perhaps due to this empirical dilemma, psychiatric research has instead attempted a shortcut to explain crimes by diagnosing patterns of crimes as mental disorders. Here, the lack of definitional clarity has become abysmal. Diagnoses such as kleptomania, intermittent explosive disorder, paedophilia, or psychopathy, have been defined on the basis of criminal behaviour patterns and mainly researched among convicted offenders. In order to have them constitute mental disorders, heterogeneous aspects of inner phenomena or cognitions have been assembled into diagnostic designations. By their circular reasoning and limited empirical support from general population studies, these diagnoses have continued to fuel heated controversies about which aspects should be counted as “belonging” to the respective syndromes. It came as no surprise when a recent large-scale meta-analysis of the predictive value of the different “facets” of psychopathy for crimes showed that the strongest predictor was – criminal behaviours (Walters, 2008).

This blog post is largely excerpted from the paper "Mental disorder is a cause of crime" I co-authored with Susanna Radovic, Christer Svennerlind, Pontus Höglund and Filip Radovic in 2009. 

Persistent aggressive antisocial behaviour: a promising phenotype

The main ambition for this blog is to provide a decent, science-related and thought-provoking text on forensic psychiatry once a week ("Forensic Friday"). Let's start by the quite basic question: what is the object of forensic psychiatry? What are we working with? During my first decade or so in this peculiar speciality, I often wondered where the focus of my research and clinical work lay. In psychoses, disorders of empathy, psychopathy, personality disorders, autism spectrum disorders, ADHD or even learning disorders? I have always been fascinated by psychiatry as science and loved working with patients. Frankly, I came to forensic psychiatry because there were good research opportunities and plenty of time to spend with each patient, not because I was more interested in criminology than any other occasional reader of Agatha Christie. But somehow, I had to integrate criminal behaviours in the models I worked with. And some years ago, it dawned on me that the real object of our work lies there, hidden in plain sight. Persistent aggressive antisocial behaviour per se is the core phenotype that forensic psychiatry should put center stage, diagnose, understand from the perspectives of etiology, comorbidity and longitudinal development, and learn to prevent and treat.

Persistent aggressive antisocial behaviour indeed seems a highly promising phenotype for research. In its prototypical form, it has an early onset (it may be recognized sometimes between 3 and 5 years of age according to the life-course persistent overt type described by Moffitt), clear genetic liability (about 65% of variance, Burt et al, 2009), high stability into adulthood (a majority still have aggressive antisocial behaviours as adults) and is observable and thus relatively easy to measure. There are interesting etiological hypotheses, such as associations with sex steroids and monoaminergic neurotransmission (e.g. the cerebrospinal fluid ratio between dopamine and serotonin metabolites), hypofunctioning in prefrontal cortex or limbic dysregulation, impulsivity and reactions to drugs. An adolescent-onset subtype is more related to psychosocial factors and has better prognosis.

Its epidemiology is actually well studied. The small subgroup of the population who have the early-onset form of persistent aggressive antisocial behaviour is responsible for the majority of violent crimes (the prevalence will vary across societies, but in modern Sweden, it is as low as 1%, Falk et al 2013). The male:female sex ratio is at least 9:1, maybe even 98:2 for the most aggressive types, and other known co-variates are substance abuse (type-2 alcoholism with early-onset, polysubstance abuse), ADHD and personality disorder. The group of aggressive children also grow up to account for considerable proportions of all psychiatric patients, being over-represented in every diagnostic category in adult psychiatry (Kim-Cohen et al., 2003, one of the most important papers ever published in psychiatry!). Märta Wallinius defended her PhD thesis on aggressive antisocial behaviour in 2012, and it is accessible here.

The Life History of Aggression (LHA) scale was developed by Brown and Goodwin (1982) for research in veterans displaying pathological aggression and self-harm, and has been further validated by Coccaro and co-workers (1997). It contains 11 items describing aggressive, antisocial and self-harming behaviours, each measured by frequency and yielding 0-5 points, thus giving a maximum score of 55, with three subscales for the different types of behaviours included. Scores have a left-skewed (thankfully!) but overall normal distribution, with clear pathology at about +2 standard deviations. The LHA may be used as an expert assessment or as a self-rating. We are currently investigating whether it can also be rated for a specific time-frame, thus measuring also changes in aggressive behaviour.

There are also numerous scales measuring traits related to aggression, such as aggressiveness, anger, impulsivity and so on, but I think that it is important to keep the phenotype definition behavioural, as this will be much more reliable and clinically relevant. In forensic psychiatry, we really cannot be too preoccupied with people's feelings of anger, it is the behaviours that have to change.

Some individuals with persistent aggressive antisocial behaviour may be happy with their behaviour style as it is. In these cases, psychiatry will have little to contribute and it seems best that the penal system is left to do its job undisturbed by medical opinion. But in my experience as a doctor in correctional facilities of different kinds, the vast majority of antisocially aggressive persons are profoundly unhappy about how they constantly screw up their relationships, life plans and dreams. They would be more than happy to test various forms of treatments, be they pharmacological, physical, psychological or educational. And if an intervention would help some tens of percents of affected persons to refrain from half of their future violent crimes, a very significant reduction in societal violence at large might be achieved. With enormous positive health effects at large. 

Why  did it take so long for me too relaize this stunningly simplicit model? And why is this promising phenotype obscured in ratings of controversial diagnoses like psychopathy, conduct disorder or antisocial personality disorder? I think that the answer is simple and a bit humiliating for forensic psychiatry. Persistent aggressive antisocial behaviour is too simple, too obvious, not enough psy-whatever and is difficult to make a big fuss about. There is little money or prestige involved for the expert assigning this diagnosis or from creating diagnostic instruments and specialist training to recognize it. So therefore we have chased illusional diagnoses instead of finding a way to help people stop being aggressive with each other outside of those situations when physical aggression can be pro-social: in sports with clear rules and in tasks contributing to societety's legally regulated monopoly of violence. There is an urgent need for controlled treatment studies with reduced aggressive behaviour as primary end-point, and this should be the primary goal for forensic psychiatric research.