In May 2019 the advocacy group Family First produced a press release expressing concern about the upcoming cannabis referendum in New Zealand in 2020. The primary focus of the press release was to argue for possible links between the use of cannabis and violent behaviour, apparently mediated via increases in psychotic symptomatology/psychotic illness following cannabis use. Finally, the press release cited a number of studies and sets of statistics as evidence for their position. In our view, the attempt to link cannabis use to violence via psychotic illness is tenuous at best, for several reasons.

The first major reason is the nature of the evidence linking cannabis use to increases in psychosis. While the link has been well-established for a number of years,1 the extent to which cannabis leads to an increase in psychotic illness, rather than merely increasing symptoms, is less well understood, primarily due to heterogeneity in study outcomes. For example, the Dunedin Multidisciplinary Health and Development Study reported an increase in schizophrenia symptoms and schizophreniform disorder associated with early cannabis use (prior to age 15),2 while our own study (the Christchurch Health and Development Study) reported an increase in psychotic symptomatology as measured by the SCL-90 in more frequent users of cannabis to age 25.3 It is important to note that psychotic symptomatology, while relatively uncommon, comprises a variety of symptoms ranging from relatively mild (“believing others can hear your thoughts”) to severe (“believing that a… force could control your movements or thoughts against your will”), with mild symptoms being more commonly reported.4 The heterogeneity between studies makes it difficult to ascertain precisely which ‘psychosis’ is being increased as a result of cannabis exposure.

The second major reason for the tenuousness of the Family First argument is related to the nature of the studies linking psychosis with violence. Most of the research linking increased risk of violence to psychosis have examined patients with a diagnosis of a psychotic disorder, rather than those with reported symptoms.5 The magnitude of the elevated risk among those diagnosed with a psychotic disorder is relatively small,5 but greater than among those who report symptoms but do not meet criteria for disorder,4 and it is relatively rare for psychosis to precede violent behaviour.6 In addition, many of these studies examined violence among psychiatric inpatients who have been involuntarily committed to hospital due to risk of harm to self or others, suggesting that sample selection may play a strong role in the observation of violent behaviour among these individuals.5 Furthermore, earlier studies that have examined the links between psychotic symptomatology and violence have found that there were a specific subset of symptoms, related to perceived threat and internal control-override that were related to violent behaviour, rather than psychotic symptomatology more generally.7 Therefore, while the risk of violence is elevated among individuals diagnosed with psychosis, the nature of the links between specific features of psychosis and violence is not well understood.

A further concern with the press release by Family First is that several studies have been cited, but none of them directly link cannabis exposure, psychosis and violence. In addition, many of the studies have been undertaken with selective samples (eg, men convicted of intimate partner violence), from which conclusions about these linkages in the general population cannot be drawn. It should also be noted that some studies conflate “substance use disorder” with alcohol use disorder, which has been shown to have an unequivocal link to increased risk of violent behaviour.8,9 Further, the press release fails to note methodological weaknesses in the cited studies (eg, failing to control for anti-social personality disorder), as well as cautious interpretations made by the original authors. Finally, however, the press release also quotes statistics from various jurisdictions in which the influence of cannabis is inferred by the authors of the release, but clearly cannot be shown to be causal.

We agree with Family First on one important point; more research is needed on the possible linkages between cannabis exposure and violence. However, the use of tendentious arguments, and failing to properly report on the strengths and weaknesses of the research literature is not the way to move forward in our discussions concerning the best way to regulate the consumption of cannabis. Furthermore, such commentary serves to perpetuate the stigmatisation of persons with serious mental illness.5 The New Zealand public deserves a good-faith approach by all involved in the debates.

‍

In May 2019 the advocacy group Family First produced a press release expressing concern about the upcoming cannabis referendum in New Zealand in 2020. The primary focus of the press release was to argue for possible links between the use of cannabis and violent behaviour, apparently mediated via increases in psychotic symptomatology/psychotic illness following cannabis use. Finally, the press release cited a number of studies and sets of statistics as evidence for their position. In our view, the attempt to link cannabis use to violence via psychotic illness is tenuous at best, for several reasons.

The first major reason is the nature of the evidence linking cannabis use to increases in psychosis. While the link has been well-established for a number of years,1 the extent to which cannabis leads to an increase in psychotic illness, rather than merely increasing symptoms, is less well understood, primarily due to heterogeneity in study outcomes. For example, the Dunedin Multidisciplinary Health and Development Study reported an increase in schizophrenia symptoms and schizophreniform disorder associated with early cannabis use (prior to age 15),2 while our own study (the Christchurch Health and Development Study) reported an increase in psychotic symptomatology as measured by the SCL-90 in more frequent users of cannabis to age 25.3 It is important to note that psychotic symptomatology, while relatively uncommon, comprises a variety of symptoms ranging from relatively mild (“believing others can hear your thoughts”) to severe (“believing that a… force could control your movements or thoughts against your will”), with mild symptoms being more commonly reported.4 The heterogeneity between studies makes it difficult to ascertain precisely which ‘psychosis’ is being increased as a result of cannabis exposure.

The second major reason for the tenuousness of the Family First argument is related to the nature of the studies linking psychosis with violence. Most of the research linking increased risk of violence to psychosis have examined patients with a diagnosis of a psychotic disorder, rather than those with reported symptoms.5 The magnitude of the elevated risk among those diagnosed with a psychotic disorder is relatively small,5 but greater than among those who report symptoms but do not meet criteria for disorder,4 and it is relatively rare for psychosis to precede violent behaviour.6 In addition, many of these studies examined violence among psychiatric inpatients who have been involuntarily committed to hospital due to risk of harm to self or others, suggesting that sample selection may play a strong role in the observation of violent behaviour among these individuals.5 Furthermore, earlier studies that have examined the links between psychotic symptomatology and violence have found that there were a specific subset of symptoms, related to perceived threat and internal control-override that were related to violent behaviour, rather than psychotic symptomatology more generally.7 Therefore, while the risk of violence is elevated among individuals diagnosed with psychosis, the nature of the links between specific features of psychosis and violence is not well understood.

A further concern with the press release by Family First is that several studies have been cited, but none of them directly link cannabis exposure, psychosis and violence. In addition, many of the studies have been undertaken with selective samples (eg, men convicted of intimate partner violence), from which conclusions about these linkages in the general population cannot be drawn. It should also be noted that some studies conflate “substance use disorder” with alcohol use disorder, which has been shown to have an unequivocal link to increased risk of violent behaviour.8,9 Further, the press release fails to note methodological weaknesses in the cited studies (eg, failing to control for anti-social personality disorder), as well as cautious interpretations made by the original authors. Finally, however, the press release also quotes statistics from various jurisdictions in which the influence of cannabis is inferred by the authors of the release, but clearly cannot be shown to be causal.

We agree with Family First on one important point; more research is needed on the possible linkages between cannabis exposure and violence. However, the use of tendentious arguments, and failing to properly report on the strengths and weaknesses of the research literature is not the way to move forward in our discussions concerning the best way to regulate the consumption of cannabis. Furthermore, such commentary serves to perpetuate the stigmatisation of persons with serious mental illness.5 The New Zealand public deserves a good-faith approach by all involved in the debates.

‍