Numerous studies have attempted to relate brain morphology to treatment outcome in schizophrenia in an attempt to define subtypes of schizophrenia that may respond well to antipsychotic treatment (Brown & Herz, 1989). This has potential clinical significance since, if we can predict outcome based on brain structure, it would then influence treatment.

Despite considerable advances in the treatment of schizophrenia, there remains a sizeable subgroup of individuals with this illness who are considered difficult to treat. There are a number of factors which may contribute to this phenomenon and given increasing pressure to reduce health care expenditures, the challenge of managing this patient population is becoming increasingly pressing.

There is a general consensus that 5–25% of schizophrenic patients are partially or totally unresponsive to antipsychotic drug therapy (Brenner et al, 1990). Agreement is needed on how to define the varying degrees of response and resistance in individual patients on conventional antipsychotic medication. Apparent “treatment resistance” may be due to noncompliance; excessive protein binding with inadequate bio-availability of the neuroleptic; an inadequate or excessive dose resulting in poor efficacy or side-effects outweighing therapeutic effects (Osser, 1989); or incorrect diagnosis.

There is no longer much doubt that there is a small but real association between psychosis and violence directed at others, as well as between psychosis and self-directed violence, including suicide. Schizophrenia and the affective psychoses appear to have a similar order of association with suicide (Caldwell & Gottesman, 1990), but schizophrenia is more likely to be associated with serious other-directed violence. The evidence for the effect of schizophrenia comes from three main directions. There are two substantial cross-sectional USA community studies (Swanson et al, 1990; Link et al, 1992), respectively showing a significant quantitative association between schizophrenia and violence. Comparative studies of illness and offending careers (Lindqvist & Allebeck, 1990; Hodgins, 1992; Coid et al, 1993; Taylor, 1993; Wessely et al, 1994), all show different patterns of violent offending by people with schizophrenia compared with those without a psychotic illness, the 1993 studies confirming that the onset of violence is almost invariably after the onset of illness. The third type of evidence is from phenomenology. Taylor (1985) and Link & Stueve (1994) have shown a strong association between delusions and serious violence, the former demonstrating a specific effect of acting on delusions. (For a more extended discussion see Taylor, 1995.)

A recent challenge in schizophrenia has been the management of patients who have failed to respond not only to standard therapeutic regimes but also to trials of atypical neuroleptics such as clozapine and risperidone. This article focuses on the further psychological and pharmacological management of such patients.

When we consider what defines a person with schizophrenia as ‘difficult to treat’, a number of questions are raised. These concern some fundamental assumptions about patients, therapy and the sick role in society (Parsons, 1951).