I tread carefully when discussing the results published by Ragini Heeramun and colleagues  on the topic of "whether autism is associated with convictions for violent crimes" and "the associated risk and protective factors." Carefully because, as I've indicated on other occasions (see here), people commit crimes not labels, and sweeping generalisations about labels, specific types of crime and/or the concept of 'dangerousness' tend to do very little to help anyone in the long term.
Still, I do think it is important that issues such as offending in the context of autism spectrum disorder (ASD) are not brushed under the carpet. I say this bearing in mind that a few high profile cases where autism has been mentioned alongside have been highlighted in the media recently here in Blighty (see here and see here). Indeed, in these days of cyber-terrorism linked to extremism for example, I'd like to see quite a bit more scientific investigation into the ways of protecting those on the autism spectrum from being lured into such activities alongside studies on the possible hows-and-whys of people entering into such behaviours.
Heeramun et al relied on data from the Stockholm Youth Cohort where some 5,700 participants had a recorded ASD diagnosis. Data was cross-referenced with that from the Swedish National Crime Register to ascertain how many people were charged with offences where violence was a factor. Comparing those with autism with those with not-autism, researchers concluded that: "Individuals with autism, particularly those without intellectual disability, initially appeared to have a higher risk of violent offending." They do temper that finding with the observation that "these associations markedly attenuated after co-occurring attention-deficit/hyperactivity disorder (ADHD) or conduct disorder were taken into account" suggesting that comorbidity might count. Indeed, this data tallies with other independent findings suggesting that a range of 'adverse' life events can be a significant feature of ADHD and conduct disorder in the long-term (see here) and for example, that autism and ADHD in the prison population should considered (see here).
Alongside, authors talked about how various other factors might also influence risk of violent offending ("parental criminal and psychiatric history and socioeconomic characteristics") and importantly, how: "Better school performance and intellectual disability appeared to be protective." That intellectual (learning) disability accompanying autism seems to be protective against conviction for a violent crime with autism in mind is probably due in part to the increased supervision given to such individuals. This is a point that has been raised in previous peer-reviewed research too (see here).
The take-away messages are once again, sweeping generalisations about all autism being *linked to* violent crime are not needed and that comorbidity might count (as it seems to on many occasions) when it comes to variables being linked to autism (see here for another example). As with such crime in the general population, there are a myriad of often quite individual factors contributing to such behaviours, but alongside: "Better understanding and management of comorbid psychopathology in autism may potentially help preventive action against offending behaviors in people with autism."
 Heeramun R. et al. Autism and Convictions for Violent Crimes: Population-Based Cohort Study in Sweden. JAACAP. 2017. April 3.
Heeramun, R., Magnusson, C., Gumpert, C., Granath, S., Lundberg, M., Dalman, C., & Rai, D. (2017). Autism and Convictions for Violent Crimes: Population-Based Cohort Study in Sweden Journal of the American Academy of Child & Adolescent Psychiatry DOI: 10.1016/j.jaac.2017.03.011