Friday, 30 October 2015

People commit crimes not their clinical labels

"Oregon Killer’s Mother Wrote of Troubled Son and Gun Rights" and "The Myth of the ‘Autistic Shooter’".

Those were just two of the headlines that I read recently about the tragedy that rocked the town of Roseburg in the United States and the inevitable attempts to understand such a senseless act.

Accepting that our thoughts and prayers should focus on the those murdered, and the long and painful journey that now faces families and loved ones (including that of the perpetrator's family), mention of the autism spectrum as 'potentially' being part and parcel of the killer's 'profile' is something that perhaps requires some science-based discussion. I appreciate that significant emotions come into such tragic stories as per previous instances and the question of 'why', but this is a blog about science and autism. I'm gonna stick to the available peer-reviewed literature specifically on the topic of autism and offending without hopefully sounding too cold nor too dispassionate.

I think it is worth going over a few things first for any newcomers straying across this post.

First, a few sentences about autism and/or the autism spectrum. Clinically, autism describes a developmental disorder that variably affects communication and social interactions (social affect) among other things. Alongside a heightened risk for various comorbidity - psychological and somatic - a diagnosis on the autism spectrum is both "profound and pervasive" in terms of impact on a person's life. For some that means a lifetime of round-the-clock care; for others, sometimes wrongly labelled as 'high-functioning', it can mean struggling with even mundane daily activities, not made any easier by societal attitudes and stereotypes and often accompanied by a lack of appropriate social and healthcare support. Although not wishing to paint too bleak a picture, the increased rates of suicide ideation (see here) and even requests for euthanasia (see here) for example, can represent the extremes of the struggles faced by people on the autism spectrum. I might also add that the 'lack of social and healthcare support' sentence previously mentioned similarly extends to quite a few families caring for people with autism too.

Next, although a diagnosis of autism does not provide immunity against offending behaviour, people with autism are far more likely to be a victim of crime over and above a perpetrator. Indeed, some of the traits associated with autism mean that many people on the spectrum are uniquely vulnerable to issues such as bullying (see here for the most recent research review), harassment or sometimes worse. Such traits can also lead to some people on the autism spectrum being drawn into criminality or committing criminal acts without fully comprehending the intentions of their accomplices and/or understanding the gravity of their actions. I hasten to add that such 'naivety' (if I can call it that) is likely multi-factorial in terms of the hows and whys; sometimes moderated by associated learning difficulties for example, and other times not.

OK. I hope that clears up a few things. The other point I want to make is that whilst the label of autism describes some of the behaviours of a person, I personally don't subscribe to the view that autism does (or should) define a person, in the same way that the labels of depression and anxiety or even schizophrenia don't define people. In this context, the important point is that 'people commit crimes not their clinical labels'. Keep that in mind as I continue.

Accepting that at the time of writing this post, we don't have all the details (or confirmation of of all the details) about whether indeed the killer "struggled with Asperger’s syndrome, an autism spectrum disorder" or not, there is some science on this topic in relation to such extreme offending behaviour.

Although making uncomfortable reading, I want to start with the paper by Clare Allely and colleagues [1] (open-access) which garnered quite a bit of media attention when it was first published back in 2014 on the basis of a suggestion that "a significant proportion of mass or serial killers may have had neurodevelopmental disorders such as autism spectrum disorder or head injury." Retrospectively looking at several accounts of mass or serial killers, the authors concluded that there was some evidence that "in at least some cases, neurodevelopmental problems such as ASD [autism spectrum disorder] or head injury may interact in a complex interplay with psychosocial factors to produce these very adverse outcomes." I can remember various reactions to this paper when it saw the light of day; quite a few rooted in the fear that sweeping generalisations would ensue and similar to the historical situation in schizophrenia, all autism would be generalised and equated with dangerousness.

As it turns out that didn't happen. Indeed, I actually thought the Allely paper made some important points in their review. They didn't, for example, say that every serial or mass killer 'had autism', indeed not even close: "we are able to say that probably more than 10% of serial/mass killers have ASD and a similar proportion have had a head injury." With the estimated rates of autism these days (1 in 46 according to some reports), one can perhaps see how that percentage might cover at least some of what would be expected in the general population anyway.

What Allely et al did observe is: (a) that "serial and mass killings are rare" and (b) that: "The gaps in our understanding about the actual mechanisms of development toward these most negative of outcomes are enormous." Further: "the great majority of those with ASD or head injury had also experienced psychosocial risk factors such as parental divorce, physical or sexual abuse, and major surgery during childhood." That last point might tie in with some of the details coming out of the Roseburg tragedy, although with the important provisos that (i) correlation is not necessarily the same as causation and that (ii) sweeping generalisation is usually the mother of all mistakes.

Continuing the theme of other factors/variables occurring alongside autism as also being potentially important to instances of offending behaviour are the findings reported by Newman & Ghaziuddin [2]. Surveying some of the scientific literature on the topic of violent crime specifically in relation to Asperger syndrome, the authors concluded that some 30% of cases were accompanied by "a definite psychiatric disorder" and a further 50% had a "probable psychiatric disorder at the time of committing the crime."

This research reiterates the idea that autism, some autism, offers little in the way of protection when it comes to risk of other psychopathology occurring alongside. Screening for such comorbidity should be much more of a priority than it currently is. Without hopefully shifting blame between labels - remember people commit crimes not their labels - there is a body of research emerging suggesting that issues such as psychosis for example, may show a complicated relationship with some autism. I've covered this topic a few times on this blog (see here and see here) particularly where the manifestation(s) of psychosis has led to a subsequent diagnosis on the autism spectrum. Psychosis by the way, is characterised by disrupted perception and/or interpretation of the world around. Although by no means a universal relationship, there is some evidence that particularly in first-episode psychosis, homicide rates may be heightened [3]. That all being said, I will also draw your attention to the systematic review from King & Murphy [4] on offending profiles with autism in mind which reported: "poor evidence of the presence of comorbid psychiatric diagnoses (except in mental health settings) amongst offenders with ASD."

Moving on, and the paper by Helverschou and colleagues [5] provides some further potentially important details about such offending profiles in individuals diagnosed with an ASD. So; "Unlike most others who commit criminal acts, the majority of the individuals with autism spectrum disorder in this study showed no evidence of substance abuse, had a close relationship to their victims and were willing to confess to the accused crime." Further: "in most cases, autism spectrum disorder characteristics, such as idiosyncratic comprehensions and obsessions appeared to be related to the motive for the offence."

Focusing specifically on substance abuse, this is something of a common thread in many discussions about offending profiles and psychiatry. In a recent post, I talked about some quite large-scale population research that suggested there may an intricate relationship between something like attention-deficit hyperactivity disorder (ADHD) and future risk of psychosis and/or schizophrenia. Specifically, discussions turned to how substance abuse may be one or several factors 'priming' ADHD for later psychopathology and onwards where substance abuse might fit in relation to violent crime [6] in this group. It is a complicated relationship and difficult to summarise in a few words, but the idea that substance abuse may intersect with psychiatry and violent behaviour is the key tenet.

Alongside substance abuse, I would also draw your attention to some more general science literature on various prescription medicines that have been associated with violent behaviour [7]. Moore et al surveyed the US FDA Adverse Event Reporting System (AERS) with a view to "any case report indicating homicide, homicidal ideation, physical assault, physical abuse or violence related symptoms." They found a few possible 'associations' which were reported in the mainstream media (see here) overlapping with some pharmacotherapy that might also be indicated for some aspects of autism. At the time of writing, we don't know the specific medical history of the Roseburg perpetrator nor whether illicit or prescription medication was a part of his recent clinical picture. We also don't know whether specific medicines were being withheld or any associated circumstances around compliance.

Finally, I'm going to carefully introduce the paper by O'Nions and colleagues [8] into the conversation, and some potentially pertinent discussions about a label called Pathological Demand Avoidance (PDA). PDA is an interesting diagnostic concept insofar as being described as a feature of the autism spectrum but at the time of writing, not actually being formally included in any of the standardised diagnostic texts. O'Nions et al report how most of their group with PDA met criteria for ASD yet demonstrated some important differences from more classical descriptions of autism: "this high scoring group was characterised by lack of co-operation, use of apparently manipulative behaviour, socially shocking behaviour, difficulties with other people, anxiety and sudden behavioural changes from loving to aggression." Without any further shifting of blame between labels intended (people, not labels) or indeed casting aspersions, further investigations are required on any longer-term 'correlates' of a diagnosis of PDA specifically with offending in mind. That such a diagnosis may also border on other psychopathology [9] including "anti-social traits approaching those seen in the conduct problems and callous-unemotional traits group" taps into the continuing theme of comorbidity covered a few paragraphs back.

There is other peer-reviewed literature on this topic but I'd like to think that the selected studies provide the best evidence that we have so far when it comes to what is known or suspected about offending behaviour overlapping with a label on the autism spectrum. Hopefully what you can see from the collected literature is that offending behaviour is complicated in instances where autism is mentioned; indeed, as complicated as it is when autism is not mentioned.

No-one will ever know exactly why the Roseburg killer did what he did and what were the precise circumstances around this heinous crime. It is likely however that lots of variables coincided including the ideas of notoriety and possibly a sort of 'contagion' combining with seemingly easy access to weapons. The research evidence so far on this topic tells us that any role played by a label on or off the autism spectrum is likely to be a tangled one and certainly not one working in any sort of isolation [10]. Subsequent sweeping generalisations therefore about all autism and 'dangerousness' are probably inaccurate and most certainly offer little in the way of usefulness or comfort for anyone: victims, their families or the wider autism community. Indeed perhaps only serving to wrongly stigmatise an already heavily stigmatised group as per other examples where clinical labels have been mentioned alongside murder.

What such a tragic event does however highlight is that there is a continued need for science to investigate the precipitating factors around their occurrence - biological, medical, familial, social, political - and where possible, offer evidence-based ways and means of intervening and potentially averting such extreme acts. Although of little comfort to those families and communities that have lost loved ones in such a manner, forensic analysis of the perpetrators (including those who were stopped) remains a primary tool in discerning clinical profiles and circumstances in such cases; mindful however of how sweeping generalisations can often do more harm than good [11] and also being careful not to feed any publicity that sometimes accompanies such cases.

To close, I leave you with a ray of light from the tragedy, and the story of Chris Mintz: "a father to a young boy with autism" hailed as hero.


[1] Allely CS. et al. Neurodevelopmental and psychosocial risk factors in serial killers and mass murderers. Aggression and Violent Behavior. 2014; 19: 288-301.

[2] Newman SS. & Ghaziuddin M. Violent crime in Asperger syndrome: the role of psychiatric comorbidity. J Autism Dev Disord. 2008 Nov;38(10):1848-52.

[3] Nielssen O. & Large M. Rates of homicide during the first episode of psychosis and after treatment: a systematic review and meta-analysis. Schizophr Bull. 2010 Jul;36(4):702-12.

[4] King C. & Murphy GH. A systematic review of people with autism spectrum disorder and the criminal justice system. J Autism Dev Disord. 2014 Nov;44(11):2717-33.

[5] Helverschou SB. et al. Offending profiles of individuals with autism spectrum disorder: A study of all individuals with autism spectrum disorder examined by the forensic psychiatric service in Norway between 2000 and 2010. Autism. 2015 Oct;19(7):850-8.

[6] Fazel S. et al. Schizophrenia, substance abuse, and violent crime. JAMA. 2009 May 20;301(19):2016-23.

[7] Moore TJ. et al. Prescription Drugs Associated with Reports of Violence Towards Others. PLoS ONE 2010; 5(12): e15337.

[8] O'Nions E. et al. Identifying features of 'pathological demand avoidance' using the Diagnostic Interview for Social and Communication Disorders (DISCO). Eur Child Adolesc Psychiatry. 2015 Jul 30.

[9] O'Nions E. et al. Pathological demand avoidance: exploring the behavioural profile. Autism. 2014 Jul;18(5):538-44.

[10] Søndenaa E. et al. Violence and sexual offending behavior in people with autism spectrum disorder who have undergone a psychiatric forensic examination. Psychol Rep. 2014 Aug;115(1):32-43.

[11] Metzl JM. & MacLeish KT. Mental Illness, Mass Shootings, and the Politics of American Firearms. American Journal of Public Health. 2015; 105: 240-249.

---------- Newman, S., & Ghaziuddin, M. (2008). Violent Crime in Asperger Syndrome: The Role of Psychiatric Comorbidity Journal of Autism and Developmental Disorders, 38 (10), 1848-1852 DOI: 10.1007/s10803-008-0580-8 King C, & Murphy GH (2014). A systematic review of people with autism spectrum disorder and the criminal justice system. Journal of autism and developmental disorders, 44 (11), 2717-33 PMID: 24577785