psychosis, violence, and the fear of madness

The psychological community often attempts to combat the stigma suffered by the mentally ill by challenging the view that people suffering schizophrenia and other psychotic conditions are physically dangerous. We're often told, for example, that schizophrenic subjects are more likely to suffer than to commit violence. In what follows I'll unpick some examples of this destigmatising message from section 4.4 of the professionally representative 2017 BPS report Understanding Psychosis and Schizophrenia.

To spill the beans, I'll be arguing that this section of the (generally good, if excessively normalising) report underplays the dangerousness that schizophrenic subjects can present when unwell. However I'll also argue that the widespread fear of actively psychotic people is only partly due to a fear of the (real or imagined) dangerousness of their actions.

what the BPS report says

In contrast to media stereotypes, in reality few people who experience paranoia or hear distressing voices ever hurt anyone else. It is very slightly more common for people with psychiatric diagnoses to commit violent crimes than for those without such diagnoses. However, the difference in rates is extremely small: far less, for example, than the increased risk associated with any one of: being male, being young, having consumed alcohol or used street drugs, or having been violent in the past.28 

The reference (28) given to support this assertion leads us to a paper on the 'measurement and reporting of the duration of untreated psychosis' which has nothing at all to say about violence - presumably it's just a mistake. That no valid reference at all is given for this central claim is, I think, rather striking. The above quote is also potentially misleading: 'people with psychiatric diagnoses' is a very large category, and what we're surely interested in is whether people who have such mental illnesses as involve psychosis - in particular: schizophrenic illnesses - are significantly more violent than those without. (If you're anti-psychiatric by inclination, and disposed to read talk of 'mental illnesses' as somehow all by itself implying a biomedical approach, then help yourself out at this point by inserting 'diagnoses of ' before 'such mental illnesses'.) After all it could be that people who, say, suffer from depression are too apathetic to be violent, and so bring the total violence of the mentally ill down to the population average. In fact even limiting our interest to schizophrenic individuals may be misleading, since 8 out of 10 such individuals may be significantly less likely to commit violence than non-schizophrenic individuals (perhaps violence is significantly negatively associated with negative symptoms or the use of major tranquillisers, for example), whilst 2 out of 10 (perhaps the acutely unwell ones - i.e. the ones that people actually fear) may be significantly more likely to commit violence than those without schizophrenia. This potentially vitiates the suggestion a little later in section 4.4 that 

specific diagnoses like schizophrenia do not predict dangerousness.31

The reference (31) given here is to a paper on the link between violence and mental health (and drug and alcohol use) which found that 'if a person has severe mental illness without substance abuse and history of violence, he or she has the same chances of being violent during the next 3 years as any other person in the general population.' Limitations of this paper, however, are that it relies on self-reports of diagnosis and of violence towards others (might not people under-report their own mental illnesses and might some people who have been psychotic and recovered not recall their violence?), and that it includes both medicated and non-medicated individuals in the sample. (When we're thinking about the relationship between mental illness and violence, what we're surely wanting to know about is the relation between untranquillised currently psychotic individuals and violence. That, after all, is what we intuitively understand the alleged stigma to relate to.) Once again, perhaps young tranquillised schizophrenic men are less likely than an average young man to commit violence. That, however, doesn't touch on the worry of the uninformed member of public which is that actively psychotic individuals are dangerous. In fact the BPS report rather backs up this worry:

Most violence is committed by people who have never been in contact with mental health services and the overwhelming majority of mental health service users have never been violent.

For all we know so far (I turn later to the actual data) this could be because untreated individuals who are having a first episode of schizophrenic psychosis truly are likely to be violent.

The report also tells us that 

as a result of people’s fear and prejudice, mental health service users are much more likely than others to be victims of violence.36 

The supporting reference (36) given here is to a report written by Victim Support on 'the criminal victimisation of people with mental health problems'. This report, which looked at the experiences of a total of just 81 individuals with varying diagnoses, found (inter alia) that people with severe mental illnesses were five times more likely to be a victim of assault than people in the general population, and up to four times more likely to be victimised by their relatives or acquaintances. What this report doesn't show, however, is that being the victim of such violence is (to quote the BPS report) actually 'as a result of people's fear and prejudice'. (It does detail however that 25% of those interviewed - i.e. 20 people - felt that they became victims of violence specifically because of their mental health status.) Thus perhaps the reason why people with severe mental illness suffer more violence than those without is in part because they're more likely to be living with relatives or peers who are themselves mentally ill, or in part because they're visiting or detained in psychiatric facilities and are there more likely to be victims of violence from mentally ill subjects (9 individuals reported suffering violence on the wards), or perhaps because they're more likely to act in such a way as would, regardless of who was so acting, provoke violence. It's worth noting, too, that only 15 of these 81 individuals suffered schizophrenia; 40 suffered depression, 8 PTSD, 10 personality disorder, etc.

The final piece of the BPS report which I want to touch on notes that

The reason that people associate a diagnosis with violence is most likely a result of negative and stereotyped media reporting about mental health.32 A survey found that homicide and crime were the most frequent themes in media coverage of mental health.33 Films and television dramas also often depict people with mental health problems as violent and unpredictable.34, 35 

The supporting reference for the claim of the first sentence here is to a piece on the media's depictions of mental illness hosted on John Grohol's 'PsychCentral' website. Yet whilst this information piece does indeed aptly survey the negative and stereotyped media reporting about mental health, it doesn't provide evidence that people's association of diagnoses and violence is likely due to that reporting. In fact it doesn't even consider any other reasons. (Perhaps the reporting and the stigma are effects of another common cause?) The piece also makes the above-mentioned claim that "Studies have found that dangerousness/crime is the most common theme of stories on mental illness ... But ... research suggests that mentally ill people are more likely to be victims than perpetrators of violence."

 Here I want to re-emphasise that the 'But' is misleading - it tempts us to infer from the putative fact that mentally ill people are more likely to be victims than perpetrators of violence to a conclusion about how violent mentally ill people are likely to be compared to those who are not mentally ill. (Compare: "Racing car drivers are more likely to be victims than perpetrators of car crashes on the track". Yes, true... because it's more often that one person causes several others to crash than several others causing one person to crash. But that tells us nothing about how frequent or infrequent it is that one of these drivers causes a crash.) I rather suspect that there's something about this widely repeated claim that somehow stops us thinking clearly about the issue at hand.  

the literature on psychosis and violence  

I now want to consider a little of the data we have on the rates of homicide and violent assault by people with schizophrenic and other psychotic conditions.

Here's one data source relevant to the UK psychiatric scene. Every year the University of Manchester produces what is now called The National Confidential Inquiry into Suicide and Safety in Mental Health. This considers (inter alia) the homicide rates by 'mental health patients' over the preceding decade.

The 2019 report tells us that between 2007 and 2017 there were 732 mental health patients convicted of a homicide offence, and that this represents 11% of people convicted of homicide (i.e. there was a total of 6597 homicides - an average of 599/year - during this time). (Some good news: during this period the number of homicide convictions per year for mental health patients reduced from 74 to 38, and this was a more or less steady decline.) Recent reports no longer provide much detail about the psychiatric profile of these homicides; the 2015 report however tells us more. Thus between 2003-2013 there were 630 homicides by mental health patients (an average of 57 per year). This was out of a total of 5,835 homicide convictions (an average of 530 per year) and 6,141 victims (an average of 558 per year). In short, 9.3% of these homicides were committed by the mentally unwell. 55% of the homicides by the mentally unwell were committed by people with a history of schizophrenia (and other delusional disorders). Of these, 81% had symptoms of psychosis (delusions and/or hallucinations) at the time of the offence. 203 (59%) of those with schizophrenia were extant patients.

Consider now that the point prevalence of schizophrenia in the UK is between 2 and 5.9 per 1000 - let's call it 4. That's to say, that at any one point in time, 0.4% of the population suffers schizophrenia. (The rate of acute schizophrenic psychosis - which is what we're really interested in, I think, when thinking of stigma etc. - is considerably less than this. But I don't know what it is, so will ignore it for now.) Turning back now to the 2007-2017 data, we can reasonably estimate that between 2007 and 2017 there were (732*0.55=) 402 homicides committed by schizophrenic individuals. If I've got my maths right, this means that 6% of homicides are committed by people with schizophrenia despite the fact that only 0.4% of the population is schizophrenic.

To try and zoom in on the issue of acute psychosis, let's now consider the 2008 meta-analysis by Nielssen & Large of rates of homicide during the first episode of psychosis and after treatment. The paper notes that although 'the prevalence of schizophrenic disorders is usually estimated to be below 1% of the population, patients with schizophrenia comprise between 5% and 20% of all homicide offenders.' This tallies with the findings of the University of Manchester reports. To summarise, what Nielssen & Large find is that 4 in 10 of the homicides committed by people with a psychotic illness occur before treatment, that 1 in 700 people with psychosis commit a homicide before treatment, and that 1 in 10,000 patients with psychosis who receive treatment will commit a homicide each year, so that the rate of homicide in psychosis before treatment is 15 times higher than the annual rate after treatment. (It's not germane to the argument, but I think it telling to compare this 1/700 figure with the fact that as many as 1/10 of schizophrenic subjects die at their own hand within 10 years of diagnosis.)  

In the above I have focussed only on homicidal violence. But what of aggression that doesn't result in homicide? For now I'll consider just Milton et al's paper on aggressive incidents in first-episode psychosis. These authors considered 168 consecutive patients 􏰀􏰀(aged 16-64 years) with a psychotic illness making first contact with psychiatric services in Nottingham 􏰀􏰀UK between 1 June 1992 and 31 May 1994. 9.6% of these subjects demonstrated at least one act of serious aggression (weapon use, sexual assault or victim injury) during at least one psychotic episode, and a further 23.5% demonstrated lesser acts of aggression. (It's worth noting, I think, that􏰀, unemployment and comorbid substance misuse had independent effects on risk of aggression.) In short, 1/3 of the patients in this sample suffering a first episode psychosis were aggressive.

What of the rates of mental illness in the population of individuals who commit sexual assault? According to Sorentino et al's 2018 review paper on sex offenders, the evidence on the prevalence of psychotic disorders in sex offenders is mixed. Some studies found low rate of psychotic illnesses in this population. One small study found no evidence of a psychotic spectrum disorder in 113 men convicted of sexual offences. Other studies found psychotic illnesses to be associated with an increased risk of sexual reoffending. One found psychosis to increase the risk for sexual recidivism in sex offenders with mental illness; another found that sex offenders were 4.8 times more likely to receive a diagnosis of schizophrenia and 3.4 times more likely to have bipolar affective disorder. The 2017 review by Lewis and Dwyer cites studies finding that 5-10% of sex offenders have psychotic disorders, and that male sex offenders were 5 times more likely to receive a diagnosis of a psychotic disorder than non-offenders. We should also consider the possibility that the disinhibiting effects of psychosis can amount to more aberrant sexual activity in the home which parents and siblings may not report for fear of their mentally unwell child or sibling being branded a sex offender.    

what does it mean?

One of the questions with which we started was whether it's rational to be frightened of being attacked by someone who has a schizophrenic illness. Given that 'only' 6% of homicides are committed by people with schizophrenic illness (11% with mental illness), then - if you're imagining going for a walk and being worried about being assaulted or killed - you should be more worried that someone without a mental illness will kill you. And, recall, there's only a 0.00088% chance of your being homicidally killed by anyone in a year - and so only a 0.00008% chance of being killed by someone who is mentally ill, or a 0.00005% chance of being killed by someone who is schizophrenic. But you might consider matters rather differently if you're living with someone suffering a first episode psychosis and you are worried not about homicide but about being violently attacked. (There's a 0.003% chance of being killed in a car crash in any given year. ... But if you're standing right in front of a car speeding towards you, there's perhaps a 100% chance of being killed.) In that case, keeping in mind that perhaps 33% of people experiencing such an episode may understandably become violent, and that 1 in 700 will kill someone, may be helpful.

why are we frightened of serious mental illness anyway?

Now, an interesting fact about anti-stigma campaigns regarding mental illness is that they don't really work. One possible reason for this, of course, is that the media continues to spew stigmatising or otherwise exaggerating messages about the dangers posed by people with significant mental illnesses. But reflection alone reveals, I believe, that it’s the psychotic individual in his psychosis who is feared, and not simply such of his acts as would be fearful whomever commits them. We’re disquieted by his acts under the description of ‘psychotic’, not simply under the descriptions ‘violent’ (or ‘sexually perverse’ or what-have-you); we – and he when sane – are disturbed by his delusional peculiarity, by his unrelatability of expression, by his baffling admixture of comprehensible humanity and what can seem like an inhumanity which is, other than (sometimes) motivationally, as such incomprehensible. Whilst the formation of delusions and other psychotic experiences may be motivationally intelligible, this speaks not at all to their rational intelligibility – to our ability to really ‘find our feet with’ or ‘get’ them – which is what we cannot do with the truly delusional subject. It is such foot-finding failures that are intrinsically jarring and deeply disconcerting.

It’s worth recalling that we’re not inexorably disquieted by what we don’t understand or relate to, or by what we find unpredictable. We’re typically emotionally disturbed neither by quantum physics, nor by the birds of the air, nor by the normal vagaries of the weather. Instead we blanch at that which is both close yet alien, i.e. we’re phobic of such phenomena as appear to us under the aspect of the monstrous. We fear what invites our efforts at understanding but then pulls apart our minds in the attempt. Thus we tremble neither at the dead nor at the living but at the undead. Thus we fear the psychopath all the more because of the extent to which he is, despite his deathly inhumanity, largely humanly intelligible. Thus we fear the beast who is half human, half animal. Throughout history our cultures have thrown up endless icons of the monstrous, from the minotaur to Frankenstein’s monster – to help us articulate, and thereby gain some small degree of purchase on, such otherwise thought-stopping terror. 

With respect to psychosis, now, what we fear, I suggest, is that unreason which is no disturbance of inference-making or truth-telling but rather a more primordial disturbance to thought’s footing; we’re made anxious by this jarring disconnection between us; we tremble at this unmooring within the psychotic individual – an unmooring which, when we attempt connection with her, threatens to also unseat our own basic orientation in the world. We fear this reminder that it’s but a contingent fact that we too aren’t also trapped in a waking dream, that insight can’t be secured through reasoning or will, that the anchor chain of our mind does not break and the ship of reason be flung about on the dementing wave. And if we're afraid thus, we should remember what far greater terrors beset the psychotic individual herself. We might not be able to enter into them - and there's an important sense in which neither we nor the psychotic subject herself can comprehend her psychotic experience - but we can surely understand something of how terrified she is, and agitated, and fragile, and how likely to lash out she is because of that.

Why am I banging on about all this? Well, to overcome stigmatising reactions I believe we do well to first articulate our concerns, and understand their basis, so they may then be faced with understanding. This prevents our ignoring them by misarticulating them as, for example, being due to our overestimating the risks of the physical dangerousness of those with psychotic mental illness. This work of understanding inevitably involves work on oneself, and work on oneself is typically not a matter of learning further facts about the afflicted, but rather a matter of one’s own gradual moral transformation, the careful deactivating of one’s own defences, the growth in one’s own tolerance for the alien, the owning rather than the projecting of one’s own monstrous aspect, and the growth in one’s ability to hold onto that in the other to which, despite all that is awry, we can yet offer recognition. It is perhaps no real surprise that it's properly knowing someone with mental illness – not simply being acquainted with them – and certainly not just learning some general new information about mental illness, that is one of the best predictors of a greater tolerance for the mentally ill.


Popular Posts