In the last 10 years a consensus has developed that the symptoms of psychosis may be better understood by linking the steps between the phenomenological experiences and social, psychological and neurobiological levels of explanation. Cognitive models of psychosis are an important link in this chain. They provide a psychological description of the phenomena from which hypotheses concerning causal processes can be derived and tested; social, individual, and neurobiological factors can then be integrated via their impact on these cognitive processes. In this paper, we set out the cognitive processes that we think lead to the formation and maintenance of the positive symptoms of psychosis and we attempt to integrate into our model research in social factors. If this model proves useful, a fuller integration with the findings of biological research will be required (Frith, 1992).

A reliable and valid tool for assessing paranoid thoughts is presented. It will provide an effective way for researchers to ensure consistency in research and for clinicians to assess change with treatment.

Text 4098 words 1 INTRODUCTIONThe high incidence of suicidal behaviour in schizophrenia is well-documented. The lifetime rate of completed suicide is estimated to be around 10% (Caldwell and Gottesman, 1990). Estimates of non-fatal suicide attempts range from 18-55% (Roy et al., 1986;Radomsky et al., 1999). Further, 40-54% of individuals with schizophrenia either make an attempt or think about ending their lives (McGlashan, 1984;Radomsky et al., 1999).There is substantial evidence demonstrating a relationship between these different suicidal expressions. Completed suicide in psychosis has been linked to past suicide attempts, and also to past and recent suicidal ideation (Hawton et al., 2005).Among studies that have investigated a range of social, psychological and clinical risk factors for suicide in psychosis, past suicidal behaviour has regularly been identified as one of the most reliable predictors (Nordentoft et al., 2002;De Hert et al., 2001;Walsh et al., 2001). Joiner et al. (2005) suggest that less severe suicidal behaviours may be a necessary precondition for the act of completed suicide, as these enable individuals to become competent at self-injuring, and to habituate to the negative aspects of suicide, such as fear, pain and taboo.On this basis, suicidal ideation can be regarded as an 'early warning' for more severe suicidal behaviour, making it an important phenomena to identify in vulnerable individuals, as well as a potential target for intervention. Gaining a clear understanding of the causes of suicidal thoughts in individuals with psychosis is an important step towards this therapeutic goal.Existing studies of suicide in psychosis have focused on socio-demographic and clinical predictors. Male gender, younger age, frequent relapses and recent loss events are well-known to be associated with an increased suicide risk (see Caldwell and 2 Gottesman, 1990;Hawton et al., 2005). There has, however, been less attention paid to the cognitive and emotional processes underlying suicidal behaviour in psychosis. Only depression and hopelessness have been reliably identified as risk factors (Drake et al., 1986;Dassori et al., 1990;Addington and Addington, 1992). There is a need for consideration of other affect-related processes, such as self-esteem and schematic beliefs. These have been identified as aetiological and maintaining factors in studies of depression (Beck, 1967;Clark, Beck and Alford, 1999), and may play a similar role in depressed mood and suicidal thinking in psychosis.Sucidality might equally be driven by the positive symptoms of psychosis.There are several reasons to expect hallucinations and delusions to be specific risk factors (Siris, 2001), including a tendency for irrational thinking and behaviour when positive symptoms are pronounced, and intentional acts of suicide taken for psychotic reasons (for example to escape imagined persecution). Findings concerning the suicide risk associated with hallucinations and delusions have however been varied (Fenton et al., 1997;Grunebaum et al...