Treatment to reduce risk of psychosis: The need to consider the potential harms as well as the benefits

To the Editor

In a recent Perspective on DSM-5, Gin Malhi (2013) dismissed the psychosis risk concept, claiming that there are no data showing accurate prediction or treatments that are effective in lowering risk.

In a letter responding to this view, Patrick McGorry (2013) argued that relevant data do exist. He cites a meta-analysis showing that 36% of people defined as fitting a psychosis risk syndrome progress to a psychotic illness within 3 years, which he describes as ‘a high predictive value’. He also cites meta-analyses of randomized controlled trials of treatments for this at-risk group. He describes these treatments as ‘very effective in reducing risk with the number needed to treat (NNT) between four and nine’.

I would argue that whether the predictive value is ‘high’ or the treatments ‘very effective’ must be seen in the context of the potential harms, as well as the benefits, of these treatments. Although both Professors Malhi (2013) and McGorry (2013) agree that ‘simply extrapolating treatments from established disorders to at-risk populations is not appropriate’, this is unfortunately what some researchers in this area have attempted to do. Professor McGorry and colleagues have published two randomized trials where antipsychotic medication was used to treat at-risk young people (McGorry et al., 2013; Phillips et al., 2007), and he has defended the ethics of a third trial of antipsychotics that was abandoned due to recruitment problems (McGorry, 2012). I have argued previously that antipsychotic medications involve potential cardiometabolic, neurological and psychiatric harms, which make their use on at-risk young people ethically debatable (Jorm, 2012). According to Professor McGorry’s (2013) figures, 64% of at-risk young people are predicted not to develop a psychotic illness and, for every person who benefits from treatment, there would be between three and eight who would not. This means that many would be exposed to potential harm unnecessarily.

More reassuring is Professor McGorry’s (2013) statement that ‘cognitive behavioural treatment (CBT) is the most appropriate treatment’ for at-risk individuals, which is supported by a recent meta-analysis of trials. If this turns out to be confirmed by further research, it raises another issue. Most people who meet the definition of a psychosis risk syndrome also meet criteria for a mood or anxiety disorder. Since CBT is a first-line treatment for both depression and anxiety disorders, should it not be provided to these individuals anyway, regardless of their increased psychosis risk? If the answer is ‘yes’, then the psychosis risk concept would become unnecessary for guiding treatment, and its use would be confined to selection for clinical trials where an enriched sample is needed to investigate transition to psychosis.