The problems associated with the diagnosis and treatment of schizophrenia and related disorders in combination with substance misuse and dependence are among the most intractable in psychiatry. This comorbidity, often referred to misleadingly as ‘dual diagnosis’, is particularly difficult to manage because the disorders individually are treated in completely different ways: schizophrenia is best treated by careful adherence to a regimen of prescribed drugs that are, by and large, unpleasant in their immediate effects and beneficial in the long term, whereas substance misuse treatment concentrates on avoiding adherence to (non-prescribed) drugs that are very pleasant in their immediate effects and extremely damaging in the long term. The treatment of substance misuse is most successful if the patient is highly motivated and actively engaged in treatment; the successful treatment of schizophrenia involves treatment adherence in a more passive capacity. Gentle persuasion may be the preferred approach for both treatments, but whereas it is the stereotypic mantle of the drug worker, in schizophrenia mental health workers often have to deliver compulsory treatment in emergencies.

Why are these conditions so commonly found together, why are they perceived to create such a challenge to psychiatric services (Reference Gournay, Sandford and JohnsonGournay et al, 1997), and why is our evidence base for intervention so shallow (Reference Weaver, Renton and StimsonWeaver et al, 1999)? It is useful to review the size and significance of the problem before examining the scope of interventions.

Extent of comorbidity and its impact

The latest evidence of the size of the problem comes from the Comorbidity of Substance Misuse and Mental Illness Collaborative (COSMIC) study (Reference Weaver, Madden and CharlesWeaver et al, 2003a); this found that more than 40% of patients managed by community mental health teams reported problem drug use and/or harmful alcohol use in the past year. This is not a chance co-occurrence, and even allowing for Berkson's bias (the tendency for people with multiple diagnoses to seek and receive treatment, and therefore be overrepresented in study populations drawn from treatment sources), is much greater than could be expected by random association. The general perception is that comorbidity has a much greater impact on services than its single components with increased psychiatric admission (Reference Hunt, Bergen and BashirHunt et al, 2002), violence (Reference Scott, Johnson and MenezesScott et al, 1998) and poor treatment outcome (Reference Hunt, Bergen and BashirHunt et al, 2002). However, these are not universal findings, and other studies show no increase in service use and minimal influence on other clinical variables (e.g. Reference CantwellCantwell, 2003), or cheaper costs overall (Reference Laugharne, Byford and BarberLaugharne et al, 2002). In this context, it is perhaps worth emphasising that most studies of the societal and service impacts of comorbidity have been done in the USA, and it should not be assumed that their findings apply to the UK context.

Management and treatment of comorbidity

The gains made in the treatment of comorbidity are modest at best. In the words of a recent systematic review, ‘there is no clear evidence supporting an advantage of any type of substance misuse programme for those with severe mental illness over standard care. No one programme is clearly superior to any other’ (Reference Ley, Jeffery and McLarenLey et al, 2001). In particular, there is no evidence for the superiority of what is often called the ‘integrated programme’ – an approach to management in which substance misuse and severe mental illness are treated together by a dedicated team (Reference RiesRies, 1993), over serial programmes (the consecutive treatment of substance misuse and severe mental illness) or parallel programmes (their simultaneous treatment by teams working separately). This may surprise some readers, since the integrated approach has received a very favourable press from its product champions (Reference Drake, Mercer-McFadden and MueserDrake et al, 1998). The statement that ‘after 20 years of development and research, dual diagnosis services for clients with severe mental illness are emerging as an evidence-based practice’ (Reference Drake, Essock and ShanerDrake et al, 2001) certainly seems to be premature, and there may be important, but subtle, differences between US and UK services that could prevent generalising from the US experience (Reference Fiander, Burns and GregoryFiander et al, 2003).

There is better evidence of efficacy from some individual treatment interventions. Satisfactory randomised controlled trials are remarkably few, but it must be acknowledged that this is a difficult area to research, and evidence of successful interventions in the whole field is very limited. Everyone involved in the treatment of drug dependence – politicians, public health specialists, sociologists and psychiatrists – has had to eat humble pie as each brave initiative to prevent or treat drug dependence has foundered in spite of initial optimism. It is therefore not surprising that the combination of drug dependence with schizophrenia, itself prone to chronicity and failure to respond to treatment, is proving difficult to treat.

However, there are some encouraging signs of progress in the treatment of comorbidity. In a controlled trial of patients living with family members, Barrowclough et al (Reference Barrowclough, Haddock and Tarrier2001) demonstrated the success of a combination of cognitive therapy, motivational interviewing and family intervention in producing more abstinence and better general functioning, both immediately and in the longer term. This intervention is now being extended in a bigger trial supported by the Medical Research Council.

Further developments

The development of effective treatments for comorbidity would be greatly enhanced by a good theoretical model of the association between psychosis and substance misuse. We do not know the extent to which this association is generated by social vulnerabilities, or by biological vulnerabilities such as a need for greater internal stimulation in those with schizophrenia or some novel effect of substance misuse on disordered brains. However, there is an attraction between these apparent opposites, and the extent of its co-occurrence, which even extends to tobacco with all its dangers (Reference McCreadie and KellyMcCreadie & Kelly, 2000), is not a coincidence. Atypical neuroleptic drugs reduce many of the unwanted effects of psychiatric treatment, but there is no evidence that they reduce craving for other substances and we urgently need new drugs that do. This is not an impossible task; elucidation of the biochemical basis of addiction is getting closer by the day and already there are suggestions that the cortical and hippocampal dysfunctions in schizophrenia could also be responsible for the greater reinforcing properties of drugs of misuse (Reference Chambers, Krystal and SelfChambers et al, 2001).

We also need to reconsider the approach we take towards the case management of patients with comorbidity. Current thinking is that assertive outreach is the preferred approach to management because it optimises engagement. However, our own work suggests that although assertive outreach can be useful in some populations (Reference Hassiotis, Ukoumunne and ByfordHassiotis et al, 2001), greater gains can be achieved generally by developing an approach we described as ‘sensitive anticipatory action’ (Reference Weaver, Tyrer and RitchieWeaver et al, 2003b). This focuses on the prevention of relapse (e.g. by thorough medication review) and the avoidance of crises (by social care planning primarily aimed at securing accommodation), coupled with advance planning (with the patient) for relapse and crisis management. There is increasing evidence that even moderate consumption of drugs can have a significantly negative effect on outcomes and render patients impervious to the benefit of treatments with established efficacy (Reference Hunt, Bergen and BashirHunt et al, 2002). The case for preventive, rather than responsive, action in the context of comorbidity is therefore strong, and needs to be explored further.