"Many people with autism could be homeless": conflating a diagnosis of autism with significant proxy-reported autistic traits

Here I go again. Taking issue with a media headline seemingly designed to grab attention and clicks: "Many people with autism could be homeless." And how said headline doesn't quite match up with the findings of a study upon which it is supposed to be based. It's not the first time this has happened (see here) and I very much doubt that it will be the last...

The study in question was the one published by Alasdair Churchard and colleagues [1] who attempted to fill "a gap in knowledge" concerning anecdotal evidence suggesting that "autistic people experience an elevated risk of homelessness." Researchers report results based on the "entire caseload ( N = 106) of a UK homeless outreach team" where data about autistic signs and symptoms were gathered from second-hand accounts (outreach workers). Authors reported that around 1 in 10 homeless people who underwent such proxy reporting *could* have fulfilled the diagnostic criteria for autism (one homeless person had a previously recorded diagnosis of autism apparently). Around a further 9% received a 'marginal' report insofar as autistic traits reported as being potentially present, but not necessarily to the extent of reaching clinical cut-off points. Researchers also summarised their results for the lay audience too (see here).

I'm pretty sure that you can see some of the 'incongruence' between the 'many people with autism could be homeless' headline and what the study actual did and reported on. I should also mention that although the DSM-5 was the diagnostic criteria relied on during the study, the authors highlight how this was administered via their own 'creation' - "a DSM-5 Autistic Traits in the Homeless Interview, which we call the DATHI." I don't want to poo-poo such an instrument and it's usefulness in the context of autism and homelessness, but I would like to see a lot more work on its reliability and validity in future studies before any big judgements (or even bigger headlines) are made. Even better would be talking to and directly screening homeless people for autism as a next research stage; to help for example, with important clinical decisions such as whether 'symptoms cause clinically significant impairment in social, occupational, or other important areas of current functioning' to them as per the DSM-5 guidance on receipt of an autism diagnosis. Obviously such direct questioning might not be easy. One also needs to consider whether the status of homelessness might itself contribute to such 'significant impairment in social and occupational' functioning? Also throw in the idea that autistic traits are not necessarily just indicative of autism (see here for one example) and you have a recipe for some significant misunderstanding around the risk of homelessness and a diagnosis of autism which could further stigmatise. Sweeping generalisations about autism have already done more than enough damage down the years, often built on fairly flimsy evidence.

I'm not saying that there is no possible connection between autism (or autistic traits) and the status of homelessness. On the contrary, when one looks as the some of the myriad of reasons why a person becomes homeless (see here), there's more than a pinch of overlap with issues faced by those with autism. I will particularly highlight vulnerability variables like unemployment, poverty, poor physical and mental health and a lack of social support as being some of the most obvious commonalities. There are probably others. And where autism is identified in the homeless, one would expect there to be some added incentive to ensure they are well and given the advice and help they want/need to take care of themselves. Bearing in mind that is, that we aren't really in any position to dictate anyone's living arrangements to them...

And there's one final point to consider: various studies for example, have already talked about 'homelessness and the mental health scandal' (see here) where, in some cases, 4 out of 5 homeless people have been identified as having a mental health issue. Symptoms and diagnoses such as schizophrenia/psychotic disorder, mood and anxiety disorders and intellectual disability have all been talked about in the context of homelessness [2] alongside issues like substance abuse. Some of these labels/diagnoses have even been identified as having a direct influence on rough sleeping behaviour by some people (see here). You could well say that many of those behavioural and psychiatric labels/diagnoses have also featured in autism (see here and see here and see here for examples) and that within the context of 'autism rarely appearing in a diagnostic vacuum', so autism might then show a link to homelessness. But I'd be very careful to just singling out autism in this context. Very, very, very careful indeed...

----------

[1] Churchard A. et al. The prevalence of autistic traits in a homeless population. Autism. 2018 Apr 1:1362361318768484.

[2] Nishio A. et al. Prevalence of Mental Illness, Cognitive Disability, and Their Overlap among the Homeless in Nagoya, Japan. PLoS One. 2015 Sep 17;10(9):e0138052.

----------

Various childhood psychiatric disorders may be risk factors for later substance abuse

"Childhood ADHD [attention-deficit hyperactivity disorder], ODD [oppositional defiant disorder], CD [conduct disorder], and depression increase the risk of developing substance-related disorders."

So concluded the results of the meta-analysis published by Annabeth Groenman and colleagues [1] surveying the peer-reviewed research literature between 1986 and 2016. Drawing on data from nearly 40 studies covering over three quarters of a million people that "included 22,029 participants with ADHD, 434 participants with ODD or CD, 1,433 participants with anxiety disorder, and 2,451 participants with depression" researchers looked at the risk in relation to "later alcohol-, nicotine-, or drug-related disorders or substance use disorders (SUDs)."

The level of risk identified concerning childhood psychiatric disorders and later substance abuse were not by any means borderline. Risk for SUD, substance use disorder, was particularly marked for all diagnoses and confirms the authors' conclusions about their findings that "emphasize the need for early detection and intervention to prevent debilitating substance-related disorders in later life." I say that, bearing in mind that one other diagnosis or set of diagnoses - anxiety disorder - did not seem to be related to later risk of substance-related disorder ("although the findings are highly heterogeneous").

As part of a growing pattern of research exploring the risk of future adversity associated with a diagnosis of something like ADHD or CD (see here) I share the author's sentiments that 'early detection' and 'intervention' are worthwhile ventures when it comes to such diagnoses. If one also assumes that ADHD and/or CD in particular, might also increase the risk for future psychopathology (see here), the case becomes even stronger to try and intervene early and improve future quality of life for both the individual and also on a more societal level. Guidance on this topic already exists [2].

But there are other factors to consider with such research. SUD reflects a complicated set of conditions in terms of how-and-why people arrive at such a diagnosis. Factors such as the role of peers and social issues such as poverty and homelessness [3] can all influence risk of SUD as can a variety of other variables that need to be taken on board. Whilst zooming in on individuals is a big part of the strategy to minimise any excess adverse risk of SUD in relation to ADHD, CD, ODD or depression, it should not be the only focus.

Finally, allied to the Groenman results, I might also draw your attention to those published by Clarissa Bauer-Staeb and colleagues [4] talking about how "substance misuse history conveys the greatest risk in all BBV [blood-borne viruses]" in relation to those diagnosed with a severe mental illness as a further undesirable outcome potentially stemming from certain substance abuses. And with it, yet more action is required...

----------

[1] Groenman AP. et al. Childhood Psychiatric Disorders as Risk Factor for Subsequent Substance Abuse: A Meta-Analysis. J Am Acad Child Adolesc Psychiatry. 2017 Jul;56(7):556-569.

[2] Harstad E. et al. Attention-Deficit/Hyperactivity Disorder and Substance Abuse. Pediatrics. 2014; 134:

[3] Tompsett CJ. et al. Peer Substance Use and Homelessness Predicting Substance Abuse from Adolescence Through Early Adulthood. American journal of community psychology. 2013;51(0):520-529.

[4] Bauer-Staeb C. et al. Prevalence and risk factors for HIV, hepatitis B, and hepatitis C in people with severe mental illness: a total population study of Sweden. Lancet Psychiatry. 2017 Jul 4. pii: S2215-0366(17)30253-5.

----------